Six Sigma and Beyond: Design for Six Sigma

Citation preview

SIX SIGMA AND BEYOND Design for Six Sigma

SIX SIGMA AND BEYOND A series by D.H. Stamatis Volume I

Foundations of Excellent Performance

Volume II

Problem Solving and Basic Mathematics

Volume III

Statistics and Probability

Volume IV

Statistical Process Control

Volume V

Design of Experiments

Volume VI

Design for Six Sigma

Volume VII

The Implementation Process

D. H. Stamatis

SIX SIGMA AND BEYOND Design for Six Sigma

ST. LUCIE PRES S A CRC Press Company Boca Raton London New York Washington, D.C.

SL3151 FMFrame Page 4 Friday, September 27, 2002 3:14 PM

Library of Congress Cataloging-in-Publication Data Stamatis, D. H., 1947Six sigma and beyond : design for six sigma, volume VI p. cm. -- (Six sigma and beyond series) Includes bibliographical references. ISBN 1-57444-315-1 (v. 1 : alk paper) 1. Quality control--Statistical methods. 2. Production management--Statistical methods. 3. Industrial management. I. Title. II. Series. TS156 .S73 2001 658.5′62--dc21

2001041635

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage or retrieval system, without prior permission in writing from the publisher. The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from CRC Press LLC for such copying. Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe.

Visit the CRC Press Web site at www.crcpress.com © 2003 by CRC Press LLC St. Lucie Press is an imprint of CRC Press LLC No claim to original U.S. Government works International Standard Book Number 1-57444-315-1 Library of Congress Card Number 2001041635 Printed in the United States of America 1 2 3 4 5 6 7 8 9 0 Printed on acid-free paper

SL3151 FMFrame Page 5 Friday, September 27, 2002 3:14 PM

To Christine

SL3151 FMFrame Page 6 Friday, September 27, 2002 3:14 PM

SL3151 FMFrame Page 7 Friday, September 27, 2002 3:14 PM

Preface A collage of historical facts brings us to the realization that concerns about quality are present not only in the minds of top management when things go wrong but also in the minds of customers when they buy something and it does not work. We begin the collage 20 years ago, with Wayne’s (1982) proclamation in the New York Times of “management gospel gone wrong.” Wayne quoted two Harvard professors, Hays and Abernathy, as saying, “You may have your eye on the wrong ball.” In a discussion of the cost differential between American and Japanese companies, Wayne said that American business executives argue that the Japanese advantage is largely rooted in factors unique to Japan: lower labor costs, more automated and newer factories, strong government support, and a homogeneous culture. The professors, though, argue differently, Wayne said. They claim that Japanese businesses are better because they pay attention to such basics as a clean workplace, preventive maintenance for machinery, a desire to make their production process error free, and an attitude that “thinks quality.” Other authors writing in the early 1980s made similar points. Blotnick (1982) wrote, “If it’s American, it must be bad.” The headline of an anonymous article in The Sentinel Star (1982) referred to “retailers relearning lesson of customer’s always right.” Ohmae (1982) wrote an article titled “Quality control circles: They work and don’t work.” Imai (1982) wrote that unless organizations control (eliminate) their waste, they would have problems. He identified waste as: 1. 2. 3. 4. 5. 6. 7.

The The The The The The The

waste waste waste waste waste waste waste

of of of of of of of

making too many units waiting time at the machine transporting units processing itself inventory motion making defective units

Imai pointed out some of Toyota’s advantages, specifically its autonomation system. Autonomation means that the machine is equipped with human wisdom to stop automatically whenever something goes wrong with it. Wight (1982) urged management to “learn to live with the truth.” When Honda rolled out its first American-built car, Lewin (1982) wrote, “Japanese bosses ponder mysterious U.S. workers.” Among other things, the Japanese wondered why Americans have so many lawyers, Lewin pointed out. Lohr (1983) wrote that “it’s just wishful thinking to say that Japan cannot catch up software. That is what a lot of people were saying about semiconductor industry a few years ago and the auto industry a decade ago.”

SL3151 FMFrame Page 8 Friday, September 27, 2002 3:14 PM

Holusha (1983) wrote of the “U.S. striving for efficiency.” Serrin (1983) described a study that showed that the work ethic is “alive but neglected.” Holloran (1983) wrote that an “army staff chief faults industry as producing defective materials.” Almost twenty years later, Zahary (2001) reported that Toyota strives to retain its benchmark status by continuing its focus on the Kaizen approach and genchi genbutso (go and see attitude). Winter (2001) wrote that GM is “now trying to show it understands importance of product.” McElroy (2001) wrote, “Customers don’t care how well your stock is performing. They do not care you are the lowest producer. They do not care you are the fastest to market. All they care about is the car they are buying. That is why it all comes down to product.” Morais (2001), quoting O’Connell (2000), claimed that over 100,000 focus groups were fielded in 1999, even though marketing and advertising professionals have mixed feelings about their value. Steel (1998 pp. 79 and 202–205) expressed industry’s ambivalence about focus groups. Among other things, he claimed that they are not very representative at all. The odd thing about focus groups is that we still use them to predict the sales potential of new products primarily because of their instant judgments, non-projectable conclusions, and comparatively low costs, even though we know better — that is, we know that we could do better by learning about consumers’ product needs and attitudes and understanding their lives. In the automotive industry, the evidence that something is wrong is abundantly clear, as Mayne et al. (2001) have reported. Here are some key points: 1. National Highway Traffic Safety Administration records showed more than 250 vehicle recalls as of mid-June 2001 — well on pace to exceed the previous year’s record 12-month total of 483. The 2000 total broke the previous high of 370 — set in 1999 — and shattered the next-highest mark of 328, set the year before. 2. Numbers of recalled vehicles have risen correspondingly — 23.4 million in 2000, 19.2 million the previous year, and 17.2 million in 1998. 3. The number of vehicles snared by non-compliance recalls — issued for failure to meet the Federal Motor Vehicle Safety Standards — increased to 4.5 million in 2000. This represents a 61% hike compared to 1999’s 2.8 million, and it is nearly three times the 1.6 million recorded in 1998. 4. A total of 18.9 million vehicles were recalled in 2000 because of safetyrelated defects. That is 81% of the overall recall total and a sharp increase compared to 1999 and 1998, when safety-related defects prompted recalls of 16.4 million and 15.6 million vehicles, respectively. Even more telling, perhaps, it is 9% more than the 17.3 million new light vehicles sold last year in the U.S. 5. A supplier executive, who wants to remain anonymous, bristles at the suggestion that quality problems fall at the feet of suppliers. He says quality has suffered because of the Big Three’s relentless pursuit of cost reduction. He also suggests that buyers at the Big Three are evaluated primarily on the basis of cost savings rather than on the quality of the parts they procure. In the final analysis, “Americans build to print and specification, whereas Japanese build to function.”

SL3151 FMFrame Page 9 Friday, September 27, 2002 3:14 PM

Powerful statements indeed, yet I could go on with examples involving home appliances, food, electronics, health devices, and many other types of products. However, the point is that the problems we are having are not new. The actions necessary to fix these problems are not new. What we need is a new commitment to pursue customer satisfaction and mean it. We must put quality in the design of all our products and services in such a way that the customer sees value in them. We must become like a philologist who believes that there is truth and falsehood in a comma. The pleasures of philology are such that by merely changing the placement of a comma, you can make sense out of nonsense; you can claim a small victory over ignorance and error. So, we in quality must learn to persevere and learn as much as possible about the customer. We must make strides to identify what customers need, want, and expect and then provide them with that product or service. We must do what the French philosopher Etienne Souriau observed: pour inventer il faut penser a cote. To invent, you must think aside — that is, slightly askew. Or we must follow the lead of Emily Dickinson when she wrote, “My business is circumstances,” and her readers understood the serendipity of ideas and the rewards of looking aside to see those ideas’ unlikely, or at least less than obvious, connections. This is the essence of Design for Six Sigma (DFSS). The upfront analysis and investigation of the customer is of paramount importance. So is trying to identify what is really needed (trade-off analysis) to make the difference. The DFSS approach is based on a systems overhaul and a new mindset to cure the ailments of organizations (profitability) and provide satisfaction to the customer (functionality and value). It is a proactive approach rather than a reactive approach, unlike the regular six sigma methodology. DFSS is a methodology that works for the future, rather than the present or past. DFSS is a holistic system that is based on challenging the status quo and providing a product or a service that not only is accepted by the customer but is financially rewarding for the organization. To do this, of course, managers must take risks. They must allow their engineers to design robust designs — and that means that the traditional Y = F(x) is not good enough. Now we must look for Y = F(x,n). In these equations, x is the traditional customer characteristic (cascaded to smaller and precise characteristics), but now we add the n, which is the noise. In other words, we must design our products and services in the presence of noise for maximum satisfaction. The best way to predict the future is to invent it. This suggests that the best way to know what is coming is to put yourself in charge of creating the situation you want. Be purposeful. Look at what is needed now, and set about doing it. Action works like a powerful drug to relieve feelings of fear, helplessness, anger, uncertainty, or depression. Mobilize yourself as well as the organization because you will be the primary architect of your future. One of the keys to being successful in your efforts is to anticipate. Accept the past, focus on the future, and anticipate. Consider what is coming, what needs to happen, and how you can rise to the occasion. Stay loose. Remain flexible. Be light on your feet. Instead of changing with the times, make a habit of changing a little ahead of the times. This change can happen with Designing for Six Sigma and

SL3151 FMFrame Page 10 Friday, September 27, 2002 3:14 PM

beyond. The only requirement is that we must take advantage of the future before we are ready for it. I am reminded of Flint’s (2001), Visnic’s (2001), and Mayne’s (2001) comments, respectively. American automotive companies, for example, have abandoned the car market because they do not make money on cars. They forget that the Japanese companies not only sell cars but make money from them. So what does Detroit do to sell? It focuses on price — rebates, discounts, 0% finance, lease subsidies, and so on. What does the competition do? Not only have they developed an engine/transmission with variable valve breathing, they are already using it. We are trying to perfect the five-speed, and the competition is installing six speeds; we talk about CVTs, and Audi is putting one in its new A4. We are focusing on 10 years and 150,000 miles reliability, and our competitors are pushing for 15 years and 200,000 miles reliability. In diesel technology, the Europeans and Americans are worlds apart. Even in this age of globalization, the light duty diesel markets in Europe have become more sophisticated and demanding to the point where policy makers have recognized the environmental advantages of diesel and have allowed new diesel vehicles to prove themselves as efficient, quiet, and powerful alternatives. What do we do? Our policy makers have created a regulatory structure that greatly impedes the widespread use of diesel vehicles. Consequently, Americans may be denied the performance, fuel economy, and environmental benefits of advanced diesel technology. A third example comes again from the automotive world in reference to fuel economy. One of the issues in fuel economy is the underbody design. Early on, American companies paid great attention to the design of the underbody. As time went on, the emphasis shifted to shapes that channel airflow over the bodywork, instead of what lies beneath. But while U.S. automakers were accustomed to being on top, BMW AG was redefining airflow from the ground up. Underbodies have been a priority with the Munich-based automaker since 1980. That is when BMW acquired its first wind tunnel and began development of the 1986 7-series — code named E32. Today, underbodies rank second behind rear ends, wheel housing and cooling airflow. As of right now, the initiative for BMW has gained them 2 miles per hour. When we talk about customer satisfaction we must do certain things that will help or improve the image of the organization in the perception of the customer. We are talking about prestige and reputation. Prestige and reputation differ from each other in three ways: 1. Reputation applies to individual products or services, while prestige is a characteristic of the organization as a whole. 2. Reputation can be measured on absolute scales, but prestige can only be judged in relative terms. 3. Prestige is judged relative to other organizations; reputation is not. It is prestige that we are interested from a Design for Six Sigma perspective. The reason for this is that prestige compels each organization to perform better than its competitors, thereby promoting excellence and continuously raising industry

SL3151 FMFrame Page 11 Friday, September 27, 2002 3:14 PM

standards not only for the customer but also for the competitors. To achieve prestige, we must be cognizant of some basic yet inherent items, including the following: • Be ready to engage our customers in conversation every second of the day. In the digital age, this means having an interactive medium where people can tell you what they think about your brand and your product or service whenever they have an idea, a complaint, or a compliment, or when they just want to air some ideas with somebody who knows where you are going. Easy places to start are always-open discussion boards and focus groups. When you get more sophisticated, you can try regularly scheduled special events or special meetings. The best solution? Set up an Internet communication structure that lets you have a 24 × 7 open line of communication. • Make customer relations a two-way street. Today’s customers not only want to be heard, they want to respond. They want to engage you in conversation, brainstorming, and relationship building. To facilitate this, you may want to consider two-way communication into your Web site that provides means for real-time sharing of ideas, debate, and interaction. Another way to facilitate this is through moderated chat rooms or other more organized techniques. Online events and presentations allow you to show off new ideas or development to customers, then take questions in a moderated and controlled manner, across time zones and around the world. Online meetings allow you to have customers attend “by invitation only.” Keys to success: make sure your communication is honest and credible and that the “idea flow” is going both ways. In today’s world, an organization can design digital communication systems that can provide instant information. This system can be used to brainstorm, to test concepts and features, and more importantly, to consider trade-offs. • Get your customers to help design your products and services. Most organizations ignore the best product and service designers and consultants — people who know your product or service inside out and know intimately what the market needs, more often than not. They are your customers. They can tell you a lot more than just what is right and wrong with your current products. They can tell you what they really need in future products — in functional terms. • Let your customers get to know each other. Word of mouth is a concept that no one should ever underestimate in the Internet age. The power of conversation has the lightning-quick ability to create trends, fads, and brands. People talking to each other in a moderated environment and sharing unprompted, honest opinions about your brand of product or service remains the number one way for you to get new satisfied customers. • Make your customers feel special. When you get down to it, we have been talking about delighting the customer for at least 20 years, but that is where we have stopped. We have forgotten that relationships with customers should not be any different from relationships you have with

SL3151 FMFrame Page 12 Friday, September 27, 2002 3:14 PM

close friends. You need to keep in touch. You need to be honest. You need to tell people they matter to you. To facilitate this “special attitude,” an organization may have special days for the customer, special product anniversaries and so on. However, in every special situation, representatives of the organization should be identifying new functionalities for new or pending products and shortcomings of the current products. • Never try to “understand” your customer. (This is not a contradiction of the above points. Rather, it emphasizes the notion of change in expectations.) Customers are fickle. They change. As a consequence, the organization must be vigilant in tracking the changes, the wants, and the expectations of customers. To make sure that customers are being satisfied and that they will continue to be loyal to your products and services, make sure you have a system that allows you to listen, listen, and then listen some more to what they have to say. • Shrink the globe. The world is shrinking. It has become commonplace to discuss the information revolution in terms of the creation of “global markets.” To “think global” is in vogue with the majority of large corporations. But global thinking presupposes that we also understand the “global customer.” Do we really understand? Or do we merely think we do? How do we treat all our customers as though they live right next door? One way, of course, is through a combination of modern communication technology and old-fashioned neighborliness. You need good, solid, two-way conversation with someone half a globe away that is as immediate, as powerful, and as intimate as a conversation with someone right in front of you. This obviously is difficult and demanding, and in the following chapters we are going to establish a flow of disciplines that perhaps can help us in formulating those “global customers” with their specific needs, wants, and expectations. • Design for customer satisfaction and loyalty. Some time ago I heard a saying that is quite appropriate here. The saying goes something like “everything is in a state of flux, including the status quo.” I happen to agree. Never in human history has so much change affected so many people in so many ways. The winds of change keep building, blowing harder than ever, hitting more people, reshaping all kinds of organizations. Incredible as it may sound, all these changes are happening even in organizations that think that they have understood the customer and the market. To their surprise, they have not. How else can we explain some of the latest statistics that tell us the following: 1. Business failures topped 400,00 in the first half of the 1990s and exceeded 500,000 by the end of the decade. That is double the number of the previous decade. The same trend is projected for the first decade of the new century.

SL3151 FMFrame Page 13 Friday, September 27, 2002 3:14 PM

2. Eighty-five percent of all U.S. organizations now outsource services once performed in house. 3. More than three million layoffs have occurred in the last five years. What can be done to reverse this trend? Well, some will ride the wind based only on their size, some will not make it, and some will make it. The ones that will make it must learn to operate under different conditions — conditions that delight the customer with the service or the product that an organization is offering. The organization must learn to design services or products so that the customer will see value in them and cannot stand it until it has possession of either one. As the desire of the customer increases for the service or product, the demand for quality will increase. Designing for six sigma is not a small thing, nor should it be a lighthearted undertaking. It is a very difficult road to follow, but the results are worthwhile. The structure of this volume is straightforward and follows the pattern of the model of DFSS, which is Recognize, Define, Characterize, Optimize, and Verify (RDCOV). Specifically, with each stage of the model, we will explain some of the most important tools and methodologies. Our introduction is the stage where we address the basic and fundamental characteristics of any DFSS program. It is our version of the Recognize step. Specifically, we address: 1. 2. 3. 4.

Partnering Robust teams Systems engineering Advanced quality planning

We follow with the Define stage, where we discuss customer concerns by first explaining the notion of “function” and then continuing with three very important methodologies in the pursuit of satisfying the customer. Those methodologies are: 1. Kano model 2. Quality function deployment (QFD) 3. Conjoint analysis We move into a discussion of “Best in class” by discussing benchmarking. We continue the discussion with advanced topics relating to design, specifically: 1. 2. 3. 4. 5. 6. 7. 8.

Monte Carlo Finite element analysis Excel’s solver Failure mode and effect analysis (FMEA) Reliability and R&M DOE Parameter design Tolerance design

SL3151 FMFrame Page 14 Friday, September 27, 2002 3:14 PM

We continue with relatively short discussions of manufacturing topics, specifically: 1. Design for manufacturing/assembly (DFM/DFA) 2. Mistake proofing Our discussion on miscellaneous topics is geared to enhance the overall design function and to sensitize readers to the fact that the pursuit of DFSS is a team orientation with many disciplines interwoven to produce the optimum design. Of course, we do not pretend to have exhaustively identified all methodologies and all tools, but we believe that we have identified the most critical ones. Specifically, we discuss: 1. 2. 3. 4. 5. 6. 7.

Theory of constraints Design review Trade-off analysis Cost of quality Reengineering GD&T Metrology

We follow with a chapter on innovative methodologies in pursuing DFSS such as signal process flow, axiomatic designs, and TRIZ, and then we return to classic discussions on value analysis, project management, an overview of mathematical concepts for reliability, and Taylor’s theorem and financial concepts. We conclude our discussion of Design for Six Sigma and Beyond with a formal summary in a matrix format of all the tools used, following the model of DCOV: 1. 2. 3. 4.

Define Characterize Optimize Verify

REFERENCES Anon., Retailers Relearning Lesson of Customer’s Always Right, The Sentinel Star, Jan. 17, 1982, p. 4. Blotnick, S., If It’s American, It Must Be Bad, Forbes, Feb. 1, 1982, p. 146. Flint, J., Where’s the Cars? You Can Make Money on Cars If You Really Want To, Ward’s AUTOWORLD, Sept. 2001, p .21. Halloran, R., Chief of Army Assails Industry on Arms Flaw, The New York Times, Aug. 9, 1983, p. 1. Holusha, J., Why G.M. Needs Toyota: U.S. Striving for Efficiency, The New York Times, Feb. 16, 1983, p. 1 (of business section). Imai, M., From Taylor to Ford to Toyota: Kanban System — Another Challenge from Japan, The Japan Economic Journal, Mar. 30, 1982, p. 12.

SL3151 FMFrame Page 15 Friday, September 27, 2002 3:14 PM

Lewin, T., Japanese Bosses Ponder Mysterious U.S. Workers, The New York Times, Nov. 7, 1982, p. 2 (of business section). Lohr, S., Japan’s Hard Look at Software, The New York Times, Jan. 9, 1983, p. 3 (of business section). Mayne, E., Bottoms Up! Fuel Economy Pressure Underscores Underbody Debate. Ward’s AUTOWORLD, Sept. 2001, p. 58. Mayne, E. et al., Quality Crunch, Ward’s AUTOWORLD, July 2001, p. 14. McElroy, J., Rendezvous captures consumer interest, Wards AUTOWORLD, Jan. 2001, p. 12. Morais, R., The End of Focus Groups, Quirk’s Marketing Research Review, May 2001, p. 154. O’Connell, V., advertising column, Wall Street Journal, Nov. 27, 2000, p. B21. Ohmae, K., Quality Control Circles: They Work and Don’t Work, The Wall Street Journal, Mar. 29, 1982, p. 2. Serrin, W., Study Says Work Ethic Is Alive But Neglected, The New York Times, Sept. 5, 1983, p. 4. Steel, J., Truth, Lies and Advertising, Wiley, New York, 1998. Visnic, B., Super Diesel! Anyone in the Industry Will Tell You: Forget Hybrids; Diesels Are Our One Stop Cure All, Ward’s AUTOWORLD, Sept. 2001, p. 34. Wayne, L., Management Gospel Gone Wrong, The New York Times, May 30, 1982, p. 1 (of business section). Wight, O.W., Learning To Tell the Truth, Purchasing, May 13, 1982, p. 5. Winter, D., One last speed, Wards AUTOWORLD, July 2001, p. 9. Zachary, K., Toyota Strives To Retain Its Benchmark Status, Supplement to Ward’s AUTOWORLD, Aug. 6–10, 2001, p. 11.

SL3151 FMFrame Page 16 Friday, September 27, 2002 3:14 PM

SL3151 FMFrame Page 17 Friday, September 27, 2002 3:14 PM

Acknowledgments I want to thank Dr. A. Stuart for granting me permission to use some of the material in Chapter 14. The summaries of the different distributions and reliability have added much to the volume. I am really indebted for his contribution. As with the other volumes in this series, many people have helped in many ways to make this book a reality. I am afraid that I will miss some, even though their help was invaluable. Dr. H. Hatzis, Dr. E. Panos, and Dr. E. Kelly have been indispensable in reviewing and commenting freely on previous drafts and throughout this project. I would like to thank Dr. L. Lamberson for his thoughtful comments and suggestions on reliability, G. Burke for his suggestions on R&M, and R. Kapur for his valuable comments about the flow and content of the material. I want to thank Ford Motor Company and especially Richard Rossier and David Kelley for their efforts to obtain permission for using the introductory material on “robust teams.” I want to thank Prentice Hall for granting me permission to use the material on conjoint and MANOVA analysis in Chapter 2. That material was taken from the 1998 book Multivariate Data Analysis, 5th ed., by J.F. Hair, R.E. Anderson, R.L. Tatham, and W.C. Block. I want to thank McGraw-Hill and D.R. Bothe for granting me permission to use some material on six sigma taken from the 1977 book Measuring Process Capability, by D.R. Bothe. I want to thank J. Wiley and the Buffa Foundation for granting me permission to use material on the Monte Carlo method from the 1973 book Modern Production Management, 4th ed., by E.S. Buffa. I want to thank the American Supplier Institute for granting me permission to use the L8 interaction table as well as some of their OA and linear graphs. I want to thank M.A. Anleitner, from Livonia Technical Services, for his contribution to the topic of “function” in Chapter 2, for helping me articulate some of the key points on APQP, and for serving as a sounding board on issues of value analysis. Thanks, Mike. I also want to thank J. Ondrus, from General Dynamics — Land System Division, for introducing me to Value Analysis and serving as a reviewer for earlier drafts on this topic. I want to thank T. Panson, P. Rageas, and J. Golematis, all of them certified public accountants, for their guidance and help in articulating the basics of accounting and financial concerns presented in Chapter 15. Of course, the ultimate responsibility for interpreting their guidance is solely mine. Special thanks go to the editors at CRC for putting up with me, as well as for transforming my notes and the manuscript into a user-friendly product.

SL3151 FMFrame Page 18 Friday, September 27, 2002 3:14 PM

I want to thank the participants in my seminars for their comments and recommendations. They actually piloted the material in their own organizations and saw firsthand the results of some of the techniques and methodologies discussed in this particular volume. Their comments were incorporated with much appreciation. Finally, as always, this volume would not have been completed without the support of my family and especially my navigator, chief editor, and supporter — my wife, Carla.

SL3151 FMFrame Page 19 Friday, September 27, 2002 3:14 PM

About the Author D. H. Stamatis, Ph.D., ASQC-Fellow, CQE, CMfgE, is president of Contemporary Consultants, in Southgate, Michigan. He received his B.S. and B.A. degrees in marketing from Wayne State University, his master’s degree from Central Michigan University, and his Ph.D. degree in instructional technology and business/statistics from Wayne State University. Dr. Stamatis is a certified quality engineer for the American Society of Quality Control, a certified manufacturing engineer for the Society of Manufacturing Engineers, and a graduate of BSI’s ISO 9000 lead assessor training program. He is a specialist in management consulting, organizational development, and quality science and has taught these subjects at Central Michigan University, the University of Michigan, and Florida Institute of Technology. With more than 30 years of experience in management, quality training, and consulting, Dr. Stamatis has served and consulted for numerous industries in the private and public sectors. His consulting extends across the United States, Southeast Asia, Japan, China, India, and Europe. Dr. Stamatis has written more than 60 articles and presented many speeches at national and international conferences on quality. He is a contributing author in several books and the sole author of 20 books. In addition, he has performed more than 100 automotive-related audits and 25 preassessment ISO 9000 audits, and has helped several companies attain certification. He is an active member of the Detroit Engineering Society, the American Society for Training and Development, the American Marketing Association, and the American Research Association, and a fellow of the American Society for Quality Control.

SL3151 FMFrame Page 20 Friday, September 27, 2002 3:14 PM

SL3151 FMFrame Page 21 Friday, September 27, 2002 3:14 PM

List of Figures Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Figure 2.5 Figure 2.6 Figure 2.7 Figure 2.8 Figure 2.9 Figure 2.10 Figure 2.11 Figure 2.12 Figure 2.13 Figure 2.14 Figure 2.15 Figure 2.16 Figure 2.17 Figure 3.1 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 5.7 Figure 5.8 Figure 5.9 Figure 6.1 Figure 6.2 Figure 6.3 Figure 6.4 Figure 6.5 Figure 6.6 Figure 6.7 Figure 6.8 Figure 6.9 Figure 6.10 Figure 6.11

Paper pencil assembly. Function diagram for a mechanical pencil. Ten symbols for process flow charting. Process flow for complaint handling. Kano model framework. Basic quality depicted in the Kano model. Performance quality depicted in the Kano model. Excitement quality depicted in the Kano model. Excitement quality depicted over time in the Kano model. A typical House of Quality matrix. The initial “what” of the customer. The iterative process of “what” to “how.” The relationship matrix. The conversion of “how” to “how much.” The flow of information in the process of developing the final “House of Quality.” Alternative method of calculating importance. The development of QFD. The benchmarking continuum process. Trade-off relationships between program objectives (balance design). Sequential approach. Simultaneous approach. Tomorrow’s approach … if not today’s. The product development map/guide. Manufacturing system schematic. Approaches to mistake proofing. Major inspection techniques. Function of mistake-proofing devices. Types of FMEA. Payback effort. Kano model. A Pugh matrix — shaving with a razor. Scope for DFMEA — braking system. Scope for PFMEA — printed circuit board screen printing process. Typical FMEA header. Typical FMEA body. Function tree process. Example of ballpoint pen. FMEA body.

SL3151 FMFrame Page 22 Friday, September 27, 2002 3:14 PM

Figure 6.12 Figure 6.13 Figure 6.14 Figure 6.15 Figure 6.16 Figure 6.17 Figure 6.18 Figure 6.19 Figure 6.20 Figure 6.21 Figure 6.22 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 7.6 Figure 7.7 Figure 9.1 Figure 9.2 Figure 9.3 Figure 9.4 Figure 9.5 Figure 9.6 Figure 9.7 Figure 9.8 Figure 9.9 Figure 9.10 Figure 9.11 Figure 9.12 Figure 9.13 Figure 9.14 Figure 9.15 Figure 9.16 Figure 9.17 Figure 9.18 Figure 9.19 Figure 9.20 Figure 9.21 Figure 9.22 Figure 10.1 Figure 10.2 Figure 11.1 Figure 11.2 Figure 11.3

Transferring the failure modes to the FMEA form. Transferring severity and classification to the FMEA form. Transferring causes and occurrences to the FMEA form. Transferring current controls and detection to the FMEA form. Area chart. Transferring the RPN to the FMEA form. Action plans and results analysis. Transferring action plans and action results on the FMEA form. FMEA linkages. The learning stages. Pen assembly process. Bathtub curve. A series block diagram. A parallel reliability block diagram. A complex reliability block diagram. The Weibull distribution for the example. Control factors and noise interactions. An example of a parameter design in reliability usage. An example of a partially completed fishbone diagram. An example of interaction. Example of cause-and-effect diagram. Plots of averages (higher responses are better). A linear example of a process with several factors. Contrasts shown in a graphical presentation. First round testing. Second round testing. Linear graph for L4. The orthogonal array (OA), linear graph (LG), and column interaction for L9. Three-level factors in a L8 array. Traditional approach. Nominal the best. Smaller the better Larger the better. A comparison of Cpk and loss function. Plots of averages (higher responses are better). ANOVA decomposition of multi-level factors. Factors not linear. Plots of the average standard deviation by factor level. Factor effects. Factor effects. Quality cost: The quality control system. Costs. A typical branching using signal flow graph. A simple example with signal flow graph. A hypothetical design process.

SL3151 FMFrame Page 23 Friday, September 27, 2002 3:14 PM

Figure 11.4 Figure 11.5 Figure 11.6 Figure 11.7 Figure 11.8 Figure 11.9 Figure 12.1 Figure 12.2 Figure 12.3 Figure 12.4 Figure 12.5 Figure 12.6 Figure 12.7 Figure 15.1 Figure 15.2 Figure 15.3 Figure 16.1

The graph transmission. First few terms of the probability. The effect of a self loop. Node absorption. Order of design matrix showing functional coupling between FRs and DPs. Relationship of axiomatic design framework and other tools. Relationship of savings potential to time. Project identification sheet. Cost visibility sheet. Cost function worksheet. A form that may be used to direct effort. Second step in the FAST diagram block process. A partial cost function FAST diagram. Life cycle of a typical company or product. A pictorial approach of duPont’s formula. Breakeven analysis. The DFSS model.

SL3151 FMFrame Page 24 Friday, September 27, 2002 3:14 PM

SL3151 FMFrame Page 25 Friday, September 27, 2002 3:14 PM

List of Tables Table I.1 Table 1.1 Table 1.2 Table 1.3 Table 2.1 Table 2.2 Table 2.3 Table 2.4 Table 2.5 Table 2.6 Table 4.1 Table 4.2 Table 4.3 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 6.1 Table 6.2 Table 6.3 Table 6.4 Table 6.5 Table 6.6 Table 6.7 Table 6.8 Table 6.9 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 8.1 Table 8.2 Table 8.3

Probability of a completely conforming product. Customer/supplier expanded partnering interface meetings. A typical questionnaire. A general questionnaire. Characteristic matrix for a machining process. Benefits of improved total development process. Stimuli descriptions and respondent rankings for conjoint analysis of industrial cleanser. Average ranks and deviations for respondents 1 and 2. Estimated part-worths and factor importance for respondents 1 and 2. Predicted part-worth totals and comparison of actual and estimated preference rankings. Simulated samples of 20 performance time values for operations A and B. Simulated operation of the two-station assembly line when operation A precedes operation B. Simulated operation of the two-station assembly line when operation B precedes operation A. Customer attributes for a car door. Relative importance of weights. Customer’s evaluation of competitive products. Examples of mistakes and defects. DFMEA — severity rating. PFMEA — severity rating. DFMEA — occurrence rating. PFMEA — occurrence rating. DFMEA detection table. PFMEA detection table. Special characteristics for both design and process. Manufacturing process control matrix. Machinery guidelines for severity, occurrence, and detection. Failure rates with median ranks. Median ranks. Five percent rank table. Ninety-five percent rank table. Department of Defense reliability and maintainability — standards and data items. Activities in the first three phases of the R&M process. Cost comparison of two machines. Thermal calculation values.

SL3151 FMFrame Page 26 Friday, September 27, 2002 3:14 PM

Table 8.4 Table 9.1 Table 9.2 Table 9.3 Table 9.4 Table 9.5 Table 9.6 Table 9.7 Table 9.8 Table 9.9 Table 9.10 Table 9.11 Table 9.12 Table 9.13 Table 9.14 Table 9.15 Table 9.16 Table 9.17 Table 9.18 Table 9.19 Table 9.20 Table 9.21 Table 9.22 Table 9.23 Table 9.24 Table 9.25 Table 9.26 Table 9.27 Table 9.28 Table 9.29 Table 9.30 Table 9.31 Table 9.32 Table 9.33 Table 9.34 Table 9.35 Table 9.36 Table 9.37 Table 9.38 Table 9.39 Table 9.40 Table 9.41 Table 9.42 Table 9.43 Table 9.44 Table 9.45

Guidelines for the Duane model. One factor at a time. Test numbers for comparison. The group runs using DOE configurations. Comparisons using DOE. Comparisons of the two means. The test matrix for the seven factors. Test results. An example of contrasts. L4 setup. The L8 interaction table. An L9 with a two-level column. Combination method. Modified L8 array. An L8 with an L4 outer array. Recommended factor assignment by column. Formulas for calculating S/N. Concerns with NTB S/N ratio. L8 with test results. ANOVA table. Higher order relationships. Inner OA (L8) with outer OA (L4) and test results. The STB ANOVA table. The LTB ANOVA table. The NTB ANOVA table. Raw data ANOVA table. Combination design. L9 OA with test results. ANOVA table. Second run of ANOVA. L8 with test results and S/N values. ANOVA table for data from Table 9.30. Significant figures from Table 9.31. Observed versus cumulative frequency. Attribute test setup and results. ANOVA table (for cumulative frequency). The effect of the significant factors. Rate of occurrence at the optimum settings. Door closing effort: test set up and results. ANOVA table for door closing effort. The effects of the door closing effort. Rate of occurrence at the optimum settings. OA and test setup and results. ANOVA for the raw data. ANOVA table for the NTB S/N ratios. Typical ANOVA table setup.

SL3151 FMFrame Page 27 Friday, September 27, 2002 3:14 PM

Table 9.46 Table 9.47 Table 9.48 Table 9.49 Table 9.50 Table 9.51 Table 9.52 Table 9.53 Table 9.54 Table 9.55 Table 9.56 Table 9.57 Table 9.58 Table 9.59 Table 9.60 Table 9.61 Table 9.62 Table 9.63 Table 9.64

L4 OA with test results. ANOVA table raw data. ANOVA table (S/N ratio used as raw data). Level averages — raw data. OA setup and test results for Example 2. ANOVA table (S/N ratio used as raw data). Transformed data. ANOVA table for the transformed data. Components and their levels. L8 inner OA with L8 outer OA and test results. ANOVA table (NTB) and level averages for the most significant factors. Variation runs using recommended factor target values. Calculated response variance. Cost of reducing tolerances. The impact of tightening the tolerance. Reduction of 20% in the tolerance limits of component A. Reduction of tolerance limits for component D. Reduction of tolerance limits for component C. L8 OA used for the confirmation runs with the levels set, test setup, ANOVA table and level averages. Table 9.65 Response variance. Table 10.1 Design review objectives. Table 10.2 Design review checklist. Table 10.3 Comparison between traditional and concurrent engineering. Table 10.4 Typical monthly quality cost report (values in thousands of dollars). Table 10.5 Prevention costs. Table 10.6 Appraisal costs. Table 10.7 Internal failure costs. Table 10.8 External failure costs. Table 10.9 Seven-step process redesign model. Table 10.10 GD&T characteristics and symbols. Table 12.1 Project identification checklist. Table 12.2 Idea needlers or thought stimulators. Table 12.3 The worksheet for setting the list. Table 12.4 Evolution summary. Table 12.5 Ranking and weighting. Table 12.6 Criteria affecting car purchase XXXX — pair comparison. Table 12.7 Criteria weighing. Table 12.8 Criteria comparison. Table 12.9 Criteria weight comparison — completed matrix. Table 13.1 Key integrative processes. Table 13.2 The characteristics of the DFSS implementation model using project management. Table 13.3 The process of six sigma and DFSS implementation using project management. Table 14.1 Possibilities of selecting a DFSS problem.

SL3151 FMFrame Page 28 Friday, September 27, 2002 3:14 PM

Table 15.1 Table 15.2 Table 15.3

A summary of debits and credits. Summary of normal debit/credit balances. The Z score.

SL3151 FMFrame Page 29 Friday, September 27, 2002 3:14 PM

Contents Introduction Understanding the Six Sigma Philosophy.......................................1 A Static versus a Dynamic Process ..........................................................................1 Products with Multiple Characteristics .....................................................................2 Short- and Long-Term Six Sigma Capability ...........................................................4 Design for Six Sigma and the Six Sigma Philosophy..............................................5 Design Phase.........................................................................................................5 Internal Manufacturing .........................................................................................5 External Manufacturing ........................................................................................6 References..................................................................................................................7 Chapter 1 Prerequisites to Design for Six Sigma (DFSS) ...................................9 Partnering ...................................................................................................................9 The Principles of Partnering...............................................................................11 View of Buyer/Supplier Relationship: A Paradigm Shift ..................................11 Characteristics of Expanded Partnering .............................................................12 Evaluating Suppliers and Selecting Supplier Partners.......................................14 Implementing Partnering ....................................................................................14 1. Establish Top Management Enrollment (Role of Top Management — Leadership)....................................................14 2. Establish Internal Organization.................................................................14 Option 1: Supplier Partnering Manager....................................................14 Option 2: Supplier Council/Team .............................................................15 Option 3: Commodity Management Organization ...................................15 3. Establish Supplier Involvement ................................................................15 4. Establish Responsibility for Implementation ...........................................15 5. Reevaluate the Partnering Process............................................................17 Ratings.......................................................................................................17 Terms Used in Specific Questions ............................................................19 Major Issues with Supplier Partnering Relationships........................................19 How Can We Improve?.......................................................................................20 Basic Partnering Checklist..................................................................................21 1. Leadership .................................................................................................21 2. Information and Analysis..........................................................................22 3. Strategic Quality Planning ........................................................................22 4. Human Resource Development and Management ...................................22 5. Management of Process Quality...............................................................23 6. Quality and Operational Results...............................................................23 7. Customer Focus and Satisfaction .............................................................23 Expanded Partnering Checklist ..........................................................................23 1. Leadership .................................................................................................23 2. Information and Analysis..........................................................................24

SL3151 FMFrame Page 30 Friday, September 27, 2002 3:14 PM

3. Strategic Quality Planning ........................................................................24 4. Human Resource Development and Management ...................................24 5. Management of Process Quality...............................................................25 6. Quality and Operational Results...............................................................25 7. Customer Focus and Satisfaction .............................................................25 The Robust Team: A Quality Engineering Approach .............................................25 Team Systems .....................................................................................................26 Input ...............................................................................................................27 Signal..............................................................................................................27 The System.....................................................................................................27 Output/Response ............................................................................................28 The Environment............................................................................................28 External Variation...........................................................................................28 Internal Variation............................................................................................29 The Boundary.................................................................................................29 Controlling a Team Process: Conformance in Teams........................................29 Strategies for Dealing with Variation .................................................................30 Controlling or Eliminating Variation .............................................................30 Compensating for Variation ...........................................................................30 System Feedback............................................................................................31 Minimizing the Effect of Variation................................................................31 Monitoring Team Performance...........................................................................33 System Interrelationships...............................................................................33 Systems Engineering ...............................................................................................34 “Systems“ Defined ..............................................................................................34 Implications of the Systems Concept for the Manager .....................................35 Defining Systems Engineering ...........................................................................37 Pre-Feasibility Analysis ......................................................................................38 Requirement Analysis .........................................................................................38 Design Synthesis.................................................................................................38 Verification ..........................................................................................................39 Advanced Quality Planning.....................................................................................40 When Do We Use AQP?.....................................................................................42 What Is the Difference between AQP and APQP? ............................................43 How Do We Make AQP Work?..........................................................................43 Are There Pitfalls in Planning? ..........................................................................43 Do We Really Need Another Qualitative Tool to Gauge Quality?....................44 How Do We Use the Qualitative Methodology in an AQP Setting?.................44 APQP Initiative and Relationship to DFSS .......................................................45 References................................................................................................................47 Selected Bibliography..............................................................................................47 Chapter 2 Customer Understanding....................................................................49 The Concept of Function.........................................................................................52 Understanding Customer Wants and Needs .......................................................54

SL3151 FMFrame Page 31 Friday, September 27, 2002 3:14 PM

Creating a Function Diagram .............................................................................55 The Product Flow Diagram and the Concept of Functives ...............................56 The Process Flow Diagram ................................................................................61 Using Function Concepts with Productivity and Quality Methodologies.........64 Kano Model .............................................................................................................68 Basic Quality.......................................................................................................69 Performance Quality ...........................................................................................69 Excitement Quality .............................................................................................69 Quality Function Deployment (QFD) .....................................................................71 Terms Associated with QFD...............................................................................73 Benefits of QFD..................................................................................................73 Issues with Traditional QFD...............................................................................75 Process Overview................................................................................................76 Developing a “QFD” Project Plan .....................................................................76 The Customer Axis ........................................................................................77 Technical Axis................................................................................................79 Internal Standards and Tests ..........................................................................79 The QFD Approach ............................................................................................79 QFD Methodology..............................................................................................80 QFD and Planning ..............................................................................................84 Product Development Process ............................................................................86 Conjoint Analysis ....................................................................................................88 What Is Conjoint Analysis?................................................................................88 A Hypothetical Example of Conjoint Analysis..................................................89 An Empirical Example .......................................................................................90 The Managerial Uses of Conjoint Analysis .......................................................95 References................................................................................................................95 Selected Bibliography..............................................................................................95 Chapter 3 Benchmarking ....................................................................................97 General Introduction to Benchmarking...................................................................97 A Brief History of Benchmarking......................................................................97 Potential Areas of Application of Benchmarking ..............................................97 Benchmarking and Business Strategy Development ..............................................99 Least Cost and Differentiation............................................................................99 Characteristics of a Least Cost Strategy ..........................................................100 Characteristics of a Differentiated Strategy .....................................................101 Benchmarking and Strategic Quality Management ..............................................102 Benchmarking and Six Sigma ..........................................................................105 National Quality Award Winners and Benchmarking......................................107 Example — Cadillac ....................................................................................107 A Second Example — Xerox ......................................................................108 Third Example — IBM Rochester...............................................................109 Fourth Example — Motorola.......................................................................110 Benchmarking and the Deming Management Method ....................................110

SL3151 FMFrame Page 32 Friday, September 27, 2002 3:14 PM

Benchmarking and the Shewhart Cycle or Deming Wheel.............................111 Plan...............................................................................................................111 Do .................................................................................................................111 Study — Observe the Effects ......................................................................111 Act ................................................................................................................111 Why Do People Buy? .......................................................................................111 Alternative Definitions of Quality ....................................................................112 Determining the Customer’s Perception of Quality.........................................117 Quality, Pricing and Return on Investment (ROI) — The PIMS Results .......119 Benchmarking as a Management Tool..................................................................119 What Benchmarking Is and Is Not...................................................................120 The Benchmarking Process ..............................................................................121 Types of Benchmarking....................................................................................122 Organization for Benchmarking .......................................................................123 Requirements for Success.................................................................................124 Benchmarking and Change Management .............................................................126 Structural Pressure ............................................................................................128 Aspiration for Excellence .................................................................................128 Force Field Analysis .........................................................................................128 Identification of Benchmarking Alternatives ........................................................129 Externally Identified Benchmarking Candidates..............................................129 Industry Analysis and Critical Success Factors ..........................................129 PIMS Par Report ..........................................................................................130 Financial Comparison ..................................................................................130 Competitive Evaluations ..............................................................................131 Focus Groups ...............................................................................................131 Importance/Performance Analysis ...............................................................131 Internally Identified Benchmarking Candidates — Internal Assessment Surveys..........................................................................................132 Nominal Group Process: General Areas in Greatest Need of Improvement............................................................................................132 Pareto Analysis.............................................................................................132 Statistical Process Control ...........................................................................133 Trend Charting .............................................................................................133 Product and Company Life Cycle Position.................................................133 Failure Mode and Effect Analysis ...............................................................134 Cost/Time Analysis ......................................................................................134 Need to Identify Underlying Causes ................................................................134 Problem, Causes, Solutions .........................................................................134 The Five Whys .............................................................................................134 Cause and Effect Diagram ...........................................................................134 Business Assessment — Strengths and Weaknesses.............................................135 Prioritization of Benchmarking Alternatives — Prioritization Process................139 Prioritization Matrix .........................................................................................139 Quality Function Deployment (House of Quality) ..........................................140 Importance/Feasibility Matrix ..........................................................................141

SL3151 FMFrame Page 33 Friday, September 27, 2002 3:14 PM

Paired Comparisons .....................................................................................141 Improvement Potential .................................................................................141 Prioritization Factors....................................................................................141 Are There Any Other Problems? What Is the Relative Importance of Each of These? .............................................................................................142 Identification of Benchmarking Sources...............................................................142 Types of Benchmark Sources ...........................................................................142 Internal Best Performers ..............................................................................143 Competitive Best Performers .......................................................................143 Best of Class ................................................................................................143 Selection Criteria ..............................................................................................144 Sources of Competitive Information ................................................................144 Gaining the Cooperation of the Benchmark Partner........................................148 Making the Contact ..........................................................................................149 Benchmarking — Performance and Process Analysis..........................................149 Preparation of the Benchmarking Proposal......................................................149 Activity before the Visit....................................................................................149 Understanding Your Own Operations..........................................................149 Activity Analysis..........................................................................................150 1. Define the Activity .............................................................................150 2. Determine the Triggering Event ........................................................150 3. Define the Activity .............................................................................150 4. Determine the Resource Requirements .............................................151 5. Determine the Activity Drivers ..........................................................151 6. Determine the Output of the Activity ................................................151 7. Determine the Activity Performance Measure ..................................151 Model the Activity .......................................................................................152 Examples of Modeling.................................................................................152 Flow Chart the Process ................................................................................153 Activities during the Visit.................................................................................155 Understand the Benchmark Partner’s Activities..........................................155 Identification of All of the Factors Required for Success ..........................155 Activities after the Visit ....................................................................................156 1. Functional Analysis.................................................................................156 2. Cost Analysis...........................................................................................156 3. Technology Forecasting ..........................................................................156 4. Financial Benchmarking .........................................................................157 5. Sales Promotion and Advertising ...........................................................157 6. Warehouse Operations ............................................................................157 7. PIMS Analysis.........................................................................................157 8. Purchasing Performance Benchmarks ....................................................157 Motorola Example........................................................................................158 Gap Analysis..........................................................................................................158 Definition of Gap Analysis ...............................................................................158 Current versus Future Gap ...............................................................................158

SL3151 FMFrame Page 34 Friday, September 27, 2002 3:14 PM

Goal Setting ...........................................................................................................159 Goal Definition .................................................................................................159 Goal Characteristics ..........................................................................................159 Result versus Effort Goals................................................................................159 Goal Setting Philosophy ...................................................................................159 Best of the Best versus Optimization ..........................................................159 Kaizen versus Breakthrough Strategies .......................................................160 Guiding Principle Implications.........................................................................160 Goal Structure ...................................................................................................160 Cascading Goal Structure ............................................................................160 Interdepartmental Goals...............................................................................161 Action Plan Identification and Implementation ....................................................161 A Creative Planning Process ............................................................................162 Action Plan Prioritization .................................................................................162 Action Plan Documentation..............................................................................162 Monitoring and Control ....................................................................................162 Financial Analysis of Benchmarking Alternatives................................................163 Managing Benchmarking for Performance...........................................................164 References..............................................................................................................166 Selected Bibliography............................................................................................167 Chapter 4 Simulation ........................................................................................169 What Is Simulation? ..............................................................................................169 Simulated Sampling...............................................................................................170 Finite Element Analysis (FEA) .............................................................................175 Types of Finite Elements ..................................................................................175 Types of Analyses .............................................................................................176 Procedures Involved in FEA.............................................................................178 Steps in Analysis Procedure .............................................................................178 Overview of Finite Element Analysis — Solution Procedure .........................179 Input to the Finite Element Model ...................................................................180 Outputs from the Finite Element Analysis.......................................................180 Analysis of Redesigns of Refined Model ........................................................181 Summary — Finite Element Technique: A Design Tool .................................182 Excel’s Solver ........................................................................................................182 Design Optimization..............................................................................................182 How To Do Design Optimization.....................................................................184 Understanding the Optimization Algorithm .....................................................184 Conversion to an Unconstrained Problem........................................................185 Simulation and DFSS ............................................................................................185 References..............................................................................................................186 Selected Bibliography............................................................................................186 Chapter 5 Design for Manufacturability/Assembly (DFM/DFA or DFMA) ..........................................................................................187

SL3151 FMFrame Page 35 Friday, September 27, 2002 3:14 PM

Business Expectations and the Impact from a Successful DFM/DFA.................189 The Essential Elements for Successful DFM/DFA ..............................................192 The Product Plan ..............................................................................................194 Product Design.............................................................................................194 Criteria for Decision between Crash Program and Perfect Product...........195 Case #1 — Crash Program .....................................................................195 Case #2 — Perfect Product Design ........................................................196 The Product Plan — Product Design Itself.................................................196 Define Product Performance Requirement ..................................................198 Available Tools and Methods for DFMA .............................................................198 Cookbooks for DFM/DFA................................................................................199 Use of the Human Body ..............................................................................199 Arrangement of the Work Place ..................................................................200 Design of Tools and Equipment ..................................................................200 Mitsubishi Method............................................................................................200 U-MASS Method..............................................................................................202 MIL-HDBK-727 ...............................................................................................203 Fundamental Design Guidance .............................................................................204 The Manufacturing Process...................................................................................206 Mistake Proofing....................................................................................................208 Definition ..........................................................................................................208 The Strategy ......................................................................................................208 Defects ..............................................................................................................209 Mistake Proof System Is a Technique for Avoiding Errors in the Workplace ...............................................................................................210 Types of Human Mistakes ................................................................................210 Forgetfulness ................................................................................................210 Mistakes of Misunderstanding.....................................................................210 Identification Mistakes .................................................................................210 Amateur Errors.............................................................................................211 Willful Mistakes...........................................................................................211 Inadvertent Mistakes ....................................................................................211 Slowness Mistakes .......................................................................................211 Lack of Standards Mistakes.........................................................................211 Surprise Mistakes .........................................................................................211 Intentional Mistakes .....................................................................................212 Defects and Errors ............................................................................................212 Mistake Types and Accompanying Causes ......................................................213 Signals that Alert ..............................................................................................215 Approaches to Mistake Proofing ......................................................................215 Major Inspection Techniques.......................................................................216 Mistake Proof System Devices....................................................................216 Devices Used as “Detectors of Mistakes” ..............................................217 Devices Used as “Preventers of Mistakes”.............................................217 Equation for Success ........................................................................................218 Typical Error Proofing Devices ...................................................................219

SL3151 FMFrame Page 36 Friday, September 27, 2002 3:14 PM

References..............................................................................................................219 Selected Bibliography............................................................................................219 Chapter 6 Failure Mode and Effect Analysis (FMEA) ....................................223 Definition of FMEA ..............................................................................................224 Types of FMEAs....................................................................................................224 Is FMEA Needed? .................................................................................................225 Benefits of FMEA .................................................................................................226 FMEA History .......................................................................................................226 Initiation of the FMEA..........................................................................................227 Getting Started .......................................................................................................228 1. Understand Your Customers and Their Needs ............................................228 2. Know the Function ......................................................................................230 3. Understand the Concept of Priority ............................................................230 4. Develop and Evaluate Conceptual Designs/Processes Based on Customer Needs and Business Strategy......................................................230 5. Be Committed to Continual Improvement ..................................................231 6. Create an Effective FMEA Team ................................................................231 7. Define the FMEA Project and Scope ..........................................................234 The FMEA Form ...................................................................................................235 Developing the Function...................................................................................238 Organizing Product Functions ..........................................................................239 Failure Mode Analysis......................................................................................240 Understanding Failure Mode .......................................................................240 Failure Mode Questions...............................................................................240 Determining Potential Failure Modes..........................................................242 Failure Mode Effects ........................................................................................243 Effects and Severity Rating .........................................................................244 Severity Rating (Seriousness of the Effect) ................................................245 Failure Cause and Occurrence..........................................................................246 Popular Ways (Techniques) to Determine Causes ......................................247 Occurrence Rating........................................................................................249 Current Controls and Detection Ratings .....................................................249 Detection Rating ..........................................................................................250 Understanding and Calculating Risk................................................................251 Action Plans and Results.......................................................................................253 Classification and Characteristics.....................................................................254 Product Characteristics/“Root Causes” .......................................................255 Process Parameters/“Root Causes”..............................................................255 Driving the Action Plan ....................................................................................255 Linkages among Design and Process FMEAs and Control Plan.........................258 Getting the Most from FMEA...............................................................................260 System or Concept FMEA ....................................................................................262 Design Failure Mode and Effects Analysis (DFMEA).........................................262 Objective ...........................................................................................................263 Timing ...............................................................................................................263

SL3151 FMFrame Page 37 Friday, September 27, 2002 3:14 PM

Requirements ....................................................................................................263 Discussion .........................................................................................................263 Forming the Appropriate Team....................................................................263 Describing the Function of the Design/Product..........................................264 Describing the Failure Mode Anticipated ...................................................264 Describing the Effect of the Failure ............................................................264 Describing the Cause of the Failure ............................................................264 Estimating the Frequency of Occurrence of Failure ...................................265 Estimating the Severity of the Failure.........................................................265 Identifying System and Design Controls ....................................................265 Estimating the Detection of the Failure ......................................................266 Calculating the Risk Priority Number .........................................................267 Recommending Corrective Action...............................................................267 Strategies for Lowering Risk: (System/Design) — High Severity or Occurrence ..........................................................................................267 Strategies for Lowering Risk: (System/Design) — High Detection Rating ......................................................................................................267 Process Failure Mode and Effects Analysis (FMEA)...........................................268 Objective ...........................................................................................................268 Timing ...............................................................................................................268 Requirements ....................................................................................................268 Discussion .........................................................................................................269 Forming the Team ........................................................................................269 Describing the Process Function .................................................................269 Manufacturing Process Functions...........................................................269 The PFMEA Function Questions............................................................270 Describing the Failure Mode Anticipated ...................................................270 Describing the Effect(s) of the Failure........................................................271 Describing the Cause(s) of the Failure........................................................272 Estimating the Frequency of Occurrence of Failure ...................................273 Estimating the Severity of the Failure.........................................................273 Identifying Manufacturing Process Controls...............................................273 Estimating the Detection of the Failure ......................................................274 Calculating the Risk Priority Number .........................................................275 Recommending Corrective Action...............................................................275 Strategies for Lowering Risk: (Manufacturing) — High Severity or Occurrence ..........................................................................................275 Strategies for Lowering Risk: (Manufacturing) — High Detection Rating ......................................................................................................276 Machinery FMEA (MFMEA) ...............................................................................277 Identify the Scope of the MFMEA ..................................................................277 Identify the Function ........................................................................................277 Failure Mode.....................................................................................................277 Potential Effects ................................................................................................278 Severity Rating..................................................................................................279 Classification .....................................................................................................279

SL3151 FMFrame Page 38 Friday, September 27, 2002 3:14 PM

Potential Causes................................................................................................279 Occurrence Ratings...........................................................................................282 Surrogate MFMEAs..........................................................................................282 Current Controls...........................................................................................282 Detection Rating ..........................................................................................282 Risk Priority Number (RPN)............................................................................282 Recommended Actions .....................................................................................283 Date, Responsible Party....................................................................................283 Actions Taken/Revised RPN.............................................................................283 Revised RPN.....................................................................................................284 Summary ................................................................................................................284 Selected Bibliography............................................................................................284 Chapter 7 Reliability .........................................................................................287 Probabilistic Nature of Reliability ........................................................................287 Performing the Intended Function Satisfactorily..................................................288 Specified Time Period.......................................................................................288 Specified Conditions .........................................................................................289 Environmental Conditions Profile ....................................................................289 Reliability Numbers ..........................................................................................290 Indicators Used to Quantify Product Reliability..............................................290 Reliability and Quality ..........................................................................................291 Product Defects.................................................................................................291 Customer Satisfaction .......................................................................................292 Product Life and Failure Rate ..........................................................................293 Product Design and Development Cycle ..............................................................295 Reliability in Design.........................................................................................296 Cost of Engineering Changes and Product Life Cycle....................................297 Reliability in the Technology Deployment Process.........................................298 1. Pre-Deployment Process .........................................................................298 2. Core Engineering Process.......................................................................299 3. Quality Support .......................................................................................300 Reliability Measures — Testing ............................................................................300 What Is a Reliability Test? ...............................................................................300 When Does Reliability Testing Occur?............................................................301 Reliability Testing Objectives...........................................................................301 Sudden-Death Testing ..................................................................................302 Accelerated Testing ......................................................................................305 Accelerated Test Methods .....................................................................................305 Constant-Stress Testing.....................................................................................305 Step-Stress Testing............................................................................................306 Progressive-Stress Testing ................................................................................306 Accelerated-Test Models ..................................................................................306 Inverse Power Law Model ...........................................................................307 Arrhenius Model ..........................................................................................308

SL3151 FMFrame Page 39 Friday, September 27, 2002 3:14 PM

AST/PASS..............................................................................................................310 Purpose of AST.................................................................................................310 AST Pre-Test Requirements .............................................................................311 Objective and Benefits of AST.........................................................................311 Purpose of PASS...............................................................................................311 Objective and Benefits of PASS .......................................................................312 Characteristics of a Reliability Demonstration Test .............................................312 The Operating Characteristic Curve.................................................................313 Attributes Tests .................................................................................................313 Variables Tests ..................................................................................................314 Fixed-Sample Tests ...........................................................................................314 Sequential Tests ................................................................................................314 Reliability Demonstration Test Methods...............................................................314 Small Populations — Fixed-Sample Test Using the Hypergeometric Distribution ...........................................................315 Large Population — Fixed-Sample Test Using the Binomial Distribution ......................................................................315 Large Population — Fixed-Sample Test Using the Poisson Distribution.........................................................................316 Success Testing ......................................................................................................316 Sequential Test Plan for the Binomial Distribution .........................................317 Graphical Solution ............................................................................................318 Variables Demonstration Tests ..............................................................................318 Failure-Truncated Test Plans — Fixed-Sample Test Using the Exponential Distribution ..................................................................318 Time-Truncated Test Plans — Fixed-Sample Test Using the Exponential Distribution ..................................................................319 Weibull and Normal Distributions....................................................................320 Sequential Test Plans .............................................................................................321 Exponential Distribution Sequential Test Plan.................................................321 Weibull and Normal Distributions....................................................................323 Interference (Tail) Testing ................................................................................323 Reliability Vision ..............................................................................................323 Reliability Block Diagrams ..............................................................................323 Weibull Distribution — Instructions for Plotting and Analyzing Failure Data on a Weibull Probability Chart ................................................................325 Instructions for Plotting Failure and Suspended Items Data on a Weibull Probability Chart.........................................................................331 Additional Notes on the Use of the Weibull....................................................334 Design of Experiments in Reliability Applications ..............................................335 Reliability Improvement through Parameter Design ............................................336 Department of Defense Reliability and Maintainability — Standards and Data Items.......................................................................................................337 References..............................................................................................................342 Selected Bibliography............................................................................................343

SL3151 FMFrame Page 40 Friday, September 27, 2002 3:14 PM

Chapter 8 Reliability and Maintainability ........................................................345 Why Do Reliability and Maintainability?.............................................................345 Objectives...............................................................................................................346 Making Reliability and Maintainability Work ......................................................346 Who’s Responsible? ..............................................................................................347 Tools.......................................................................................................................347 Sequence and Timing ............................................................................................348 Concept ..................................................................................................................349 Bookshelf Data .................................................................................................349 Manufacturing Process Selection .....................................................................350 R&M and Preventive Maintenance (PM) Needs Analysis ..............................350 Development and Design.......................................................................................350 R&M Planning..................................................................................................350 Process Design for R&M .................................................................................351 Machinery FMEA Development ......................................................................351 Design Review ..................................................................................................352 Build and Install ....................................................................................................352 Equipment Run-Off ..........................................................................................352 Operation of Machinery....................................................................................352 Operations and Support .........................................................................................353 Conversion/Decommission ....................................................................................353 Typical R&M Measures ........................................................................................353 R&M Matrix .....................................................................................................353 Reliability Point Measurement .........................................................................354 MTBE................................................................................................................354 MTBF................................................................................................................355 Failure Rate.......................................................................................................355 MTTR................................................................................................................355 Availability ........................................................................................................356 Overall Equipment Effectiveness (OEE)..........................................................356 Life Cycle Costing (LCC) ................................................................................356 Top 10 Problems and Resolutions....................................................................357 Thermal Analysis ..............................................................................................357 Electrical Design Margins ................................................................................359 Safety Margins (SM) ........................................................................................359 Interference .......................................................................................................360 Conversion of MTBF to Failure Rate and Vice Versa .....................................361 Reliability Growth Plots ...................................................................................361 Machinery FMEA .............................................................................................361 Key Definitions in R&M .......................................................................................362 DFSS and R&M ....................................................................................................364 References..............................................................................................................365 Selected Bibliography............................................................................................365

SL3151 FMFrame Page 41 Friday, September 27, 2002 3:14 PM

Chapter 9 Design of Experiments.....................................................................367 Setting the Stage for DOE.....................................................................................367 Why DOE (Design of Experiments) Is a Valuable Tool..................................367 Taguchi’s Approach ..........................................................................................370 Miscellaneous Thoughts ...................................................................................371 Planning the Experiment .......................................................................................372 Brainstorming....................................................................................................372 Choice of Response ..........................................................................................373 Miscellaneous Thoughts ...................................................................................379 Setting Up the Experiment ....................................................................................380 Choice of the Number of Factor Levels...........................................................380 Linear Graphs ...................................................................................................382 Degrees of Freedom..........................................................................................383 Using Orthogonal Arrays and Linear Graphs ..................................................383 Column Interaction (Triangular) Table.............................................................384 Factors with Three Levels ................................................................................385 Interactions and Hardware Test Setup..............................................................385 Choice of the Test Array...................................................................................387 Factors with Four Levels ..................................................................................389 Factors with Eight Levels .................................................................................389 Factors with Nine Levels..................................................................................390 Using Factors with Two Levels in a Three-Level Array .................................390 Dummy Treatment .......................................................................................390 Combination Method ...................................................................................390 Using Factors with Three Levels in a Two-Level Array .................................391 Other Techniques ..............................................................................................391 Nesting of Factors ........................................................................................392 Setting Up Experiments with Factors with Large Numbers of Levels.......392 Inner Arrays and Outer Arrays .........................................................................393 Randomization of the Experimental Tests .......................................................394 Miscellaneous Thoughts ...................................................................................394 Loss Function and Signal-to-Noise.......................................................................397 Loss Function and the Traditional Approach ...................................................397 Calculation of the Loss Function .....................................................................398 Comparison of the Loss Function and Cpk .....................................................402 Signal-to-Noise (S/N) .......................................................................................403 Miscellaneous Thoughts ...................................................................................404 Analysis..................................................................................................................405 Graphical Analysis ............................................................................................405 Analysis of Variance (ANOVA) .......................................................................407 Estimation at the Optimum Level ....................................................................408 Confidence Interval around the Estimation......................................................409 Interpretation and Use ......................................................................................410 ANOVA Decomposition of Multi-Level Factors .............................................410

SL3151 FMFrame Page 42 Friday, September 27, 2002 3:14 PM

S/N Calculations and Interpretations................................................................411 Smaller-the-Better (STB) .............................................................................412 Larger-the-Better (LTB) ...............................................................................413 Nominal the Best (NTB) .............................................................................413 Combination Design .........................................................................................415 Miscellaneous Thoughts ...................................................................................418 Analysis of Classified Data ...................................................................................421 Classified Responses.........................................................................................422 Classified Attribute Analysis.............................................................................422 Class 1 ..........................................................................................................425 Class 2 ..........................................................................................................426 Classified Variable Analysis..............................................................................426 Discussion of the Degrees of Freedom ............................................................428 Miscellaneous Thoughts ...................................................................................429 Dynamic Situations................................................................................................430 Definition ..........................................................................................................430 Discussion .........................................................................................................431 Conditions ....................................................................................................431 Analysis ........................................................................................................432 Miscellaneous Thoughts ...................................................................................439 For Example 1..............................................................................................440 For Example 2..............................................................................................440 Parameter Design...................................................................................................441 Discussion .........................................................................................................441 Example........................................................................................................441 Tolerance Design ...................................................................................................447 Discussion .........................................................................................................447 Example........................................................................................................448 Humidity..................................................................................................454 Testing .....................................................................................................454 DOE Checklist .......................................................................................................454 Selected Bibliography............................................................................................455 Chapter 10 Miscellaneous Topics — Methodologies .......................................457 Theory of Constraints (TOC) ................................................................................457 The Goal ...........................................................................................................457 Strategic Measures ............................................................................................458 Net Profit, Return on Investment, and Productivity.........................................459 Measurement Focus ..........................................................................................460 Throughput versus Cost World.........................................................................461 Obstacles to Moving into the Throughput World ............................................461 The Foundation Elements of TOC ...................................................................463 The Theory of Non-Constraints .......................................................................463 The Five-Step Framework of TOC...................................................................464 Selected Bibliography............................................................................................465

SL3151 FMFrame Page 43 Friday, September 27, 2002 3:14 PM

Design Review .......................................................................................................465 Failure Mode and Effect Analysis (FMEA).....................................................467 References..............................................................................................................470 Selected Bibliography............................................................................................470 Trade-Off Studies...................................................................................................470 How to Conduct a Trade-Off Study: The Process ...........................................471 1. Construct the Preliminary Matrix ...........................................................471 2. Select and Assemble the Cross-Functional Team ..................................472 3. Assign Team Members’ Roles and Responsibilities ..............................472 4. Assign Ranking Teams To Evaluate the Alternatives ............................473 Identification of Ranking Methods .........................................................473 Development of Standardized Documentation .......................................474 Timing for Report out of Selection Process...........................................474 5. Weight the Various Categories................................................................474 6. Compile the Evidence Book ...................................................................475 7. Present the Results ..................................................................................475 Glossary of Terms.............................................................................................476 Selected Bibliography............................................................................................477 Cost of Quality ......................................................................................................477 Cost Monitoring System...................................................................................478 Standard Cost ...............................................................................................478 Actual Costs .................................................................................................478 Variance ........................................................................................................480 Cost Reduction Efforts.................................................................................480 Concepts of Quality Costs................................................................................480 J. Juran .........................................................................................................480 W.E. Deming ................................................................................................480 P. Crosby ......................................................................................................481 G. Taguchi ....................................................................................................481 Definition of Quality Components ...................................................................481 Methods of Measuring Quality.........................................................................483 Complaint Indices .............................................................................................484 Processing and Resolution of Customer Complaints.......................................484 Techniques for Analyzing Data ........................................................................484 Format for Presentation of Costs......................................................................485 Laws of Cost of Quality ...................................................................................485 Data Sources .....................................................................................................487 Inspection Decisions .........................................................................................487 Prevention Costs (See Table 10.5) ...................................................................487 Appraisal Costs (See Table 10.6) .....................................................................487 Internal Failure Costs (See Table 10.7)............................................................487 External Failure Costs (See Table 10.8)...........................................................487 Diagnostic Guidelines to Identify Manufacturing Process Improvement Opportunities ..............................................................................489 Diagnostic Guidelines to Identify Administrative Process Improvement Opportunities ..............................................................................490

SL3151 FMFrame Page 44 Friday, September 27, 2002 3:14 PM

Steps for Quality Improvement — Using Cost of Quality ..............................492 Procedure......................................................................................................492 Examples ......................................................................................................492 Guideline Cost of Quality Elements by Discipline .........................................502 Cost of Quality and DFSS Relationship ..........................................................509 References..............................................................................................................511 Selected Bibliography............................................................................................511 Reengineering ........................................................................................................511 Process Redesign ..............................................................................................511 The Restructuring Approach.............................................................................512 The Conference Method ...................................................................................513 The OOAD Method ..........................................................................................515 Reengineering and DFSS..................................................................................516 References..............................................................................................................517 Selected Bibliography............................................................................................518 Geometric Dimensioning and Tolerancing (GD&T) ............................................518 References..............................................................................................................523 Selected Bibliography............................................................................................523 Metrology...............................................................................................................524 Understanding the Problem ..............................................................................524 Metrology’s Role in Industry and Quality .......................................................525 Measurement Techniques and Equipment........................................................527 Purpose of Inspection .......................................................................................528 How Do We Use Inspection and Why? ...........................................................529 Methods of Testing ...........................................................................................529 Interpreting Results of Inspection and Testing ................................................530 Technique for Wringing Gage Blocks..............................................................531 Length Combinations........................................................................................532 References..............................................................................................................533 Chapter 11 Innovation Techniques Used in Design for Six Sigma (DFSS)....535 Modeling Design Iteration Using Signal Flow Graphs as Introduced by Eppinger, Nukala and Whitney (1997) ............................................................535 Rules and Definitions of Signal Flow Graphs as Introduced by Howard (1971) and Truxal (1955) ..............................................................538 Basic Operations on Signal Flow Graphs ........................................................538 The Effect of a Self Loop.................................................................................538 Solution by Node Absorption ...........................................................................539 References..............................................................................................................539 Selected Bibliography............................................................................................540 Axiomatic Designs ................................................................................................541 So, What Is an Axiomatic Design? ..................................................................542 Axiomatic and Other Design Methodologies...................................................542 Applying Axiomatic Design to Cars ................................................................543 New Designs ................................................................................................544

SL3151 FMFrame Page 45 Friday, September 27, 2002 3:14 PM

Diagnosis of Existing Design ......................................................................544 Extensions and Engineering Changes to Existing Designs ........................544 Efficient Project Work-Flow ........................................................................545 Effective Change Management ....................................................................545 Efficient Design Function ............................................................................545 References..............................................................................................................547 Selected Bibliography............................................................................................547 TRIZ — The Theory of Inventive Problem Solving ............................................548 References..............................................................................................................551 Selected Bibliography............................................................................................551 Chapter 12 Value Analysis/Engineering ...........................................................553 Introduction to Value Control — The Environment .............................................553 History of Value Control .......................................................................................555 Value Concept........................................................................................................556 Definition ..........................................................................................................556 Planned Approach .............................................................................................556 Function ............................................................................................................557 Value..................................................................................................................557 Develop Alternatives.........................................................................................558 Evaluation, Planning, Reporting, and Implementation ....................................559 The Job Plan .....................................................................................................559 Application.............................................................................................................560 Value Control — The Job Plan .............................................................................561 Value Control — Techniques versus Job Plan ......................................................562 Techniques.........................................................................................................562 Information Phase..................................................................................................563 Define the Problem ...........................................................................................563 Information Development ............................................................................564 Information Collection ............................................................................564 Cost Visibility..........................................................................................564 Project Scope...........................................................................................565 Function Determination ...............................................................................567 Function Analysis and Evaluation ...............................................................567 Cost Visibility ...................................................................................................568 Definitions ....................................................................................................568 Sources of Cost Information........................................................................570 Cost Visibility Techniques ...........................................................................570 Technique 1 — Determine Manufacturing Cost.....................................571 Technique 2 — Determine Cost Element ...............................................571 Technique 3 — Determine Component or Process Costs ......................571 Technique 4 — Determine Quantitative Costs .......................................572 Technique 5 — Determine Functional Area Costs.................................573 Function Determination ....................................................................................573 What Is Function?........................................................................................574

SL3151 FMFrame Page 46 Friday, September 27, 2002 3:14 PM

Basic and Secondary Functions...................................................................574 Basic Functions .......................................................................................574 Secondary Functions ...............................................................................575 Function Analysis and Evaluation ....................................................................575 Technique 1 — Identify and Evaluate Function..........................................575 Technique 2 — Evaluate Principle of Operation ........................................576 Technique 3 — Evaluate Basic Function ....................................................576 Technique 4 — Theoretical Evaluation of Function ...................................576 Technique 5 — Input Output Method .........................................................577 Technique 6 — Function Analysis System Technique................................577 Cost Function Relationship..........................................................................580 Evaluate the Function ..................................................................................580 Creative Phase........................................................................................................582 Phase 1. Blast....................................................................................................584 Phase 2. Create .................................................................................................584 Phase 3. Refine .................................................................................................584 Evaluation Phase....................................................................................................585 Selection and Screening Techniques ................................................................585 Pareto Voting ................................................................................................585 Paired Comparisons .....................................................................................586 Evaluation Summary....................................................................................587 Matrix Analysis ............................................................................................587 Example...................................................................................................589 Rank and Weigh Criteria....................................................................589 Evaluate Each Alternative ..................................................................590 Analyze Results ..................................................................................591 Implementation Phase............................................................................................591 Goal for Achievement.......................................................................................592 Developing a Plan.............................................................................................592 Evaluation of the System..................................................................................593 Understanding the Principles............................................................................593 Organization ......................................................................................................594 Attitude..............................................................................................................596 Value Council....................................................................................................596 Audit Results.....................................................................................................597 Project Selection ....................................................................................................597 Concluding Comments ..........................................................................................598 References..............................................................................................................598 Selected Bibliography............................................................................................598 Chapter 13 Project Management (PM).............................................................599 What Is a Project? .................................................................................................599 The Process of Project Management.....................................................................601 Key Integrative Processes......................................................................................602 Project Management and Quality..........................................................................603

SL3151 FMFrame Page 47 Friday, September 27, 2002 3:14 PM

A Generic Seven-Step Approach to Project Management....................................603 Phase 1. Define the Project ..............................................................................603 Step 1. Describe the Project ........................................................................603 Step 2. Appoint the Planning Team.............................................................604 Step 3. Define the Work...............................................................................604 Phase 2. Plan the Project ..................................................................................604 Step 4. Estimate Tasks .................................................................................604 Step 5. Calculate the Schedule and Budgets...............................................604 Phase 3. Implement the Plan ............................................................................605 Step 6. Start the Project ...............................................................................605 Phase 4. Complete the Project..........................................................................605 Step 7. Track Progress and Finish the Project ............................................605 A Generic Application of Project Management in Implementing Six Sigma and DFSS ...............................................................................................................605 The Value of Project Management in the Implementation Process ................607 Planning the Process ....................................................................................607 Goal Setting..................................................................................................608 PM and Six Sigma/DFSS .................................................................................608 Project Justification and Prioritization Techniques .....................................610 Benefit-Cost Analysis..............................................................................610 Return on Assets (ROA).....................................................................610 Return on Investment (ROI)...............................................................610 Net Present Value (NPV) Method......................................................611 Internal Rate of Return (IRR) Method ..............................................611 Payback Period Method .....................................................................612 Project Decision Analysis .......................................................................612 Why Project Management Succeeds .....................................................................613 References..............................................................................................................615 Selected Bibliography............................................................................................615 Chapter 14 Limited Mathematical Background for Design for Six Sigma (DFSS) ...........................................................................................617 Exponential Distribution and Reliability...............................................................617 Exponential Distribution...................................................................................617 Probability Density Function and Cumulative Distribution Function ........618 Probability Density Function (Decay Time)...........................................618 Cumulative Distribution Function (Rise Time) ......................................618 Reliability Problems.....................................................................................618 Constant Rate Failure .......................................................................................619 Example........................................................................................................619 Probability of Reliability ..................................................................................621 Control Charts...................................................................................................621 Continuous Time Waveform ........................................................................621 Discrete Time Samples ................................................................................621 Digital Signal Processing ........................................................................622

SL3151 FMFrame Page 48 Friday, September 27, 2002 3:14 PM

Sample Space ....................................................................................................622 Assigning Probability to Sets ...........................................................................624 Gamma Distribution ..............................................................................................625 Gamma Distribution (pdf) ................................................................................625 Gamma Function...............................................................................................626 Properties of Gamma Functions ..................................................................626 Gamma Distribution and Reliability............................................................627 Example 1: Time to Total System Failure.......................................................627 Gamma Distribution and Reliability............................................................628 Reliability Relationships ..............................................................................632 Reliability Function......................................................................................632 Data Failure Distribution ..................................................................................633 Failure Rate or Density Function .....................................................................633 Hazard Rate Function .......................................................................................634 Relations between Reliability and Hazard Functions ......................................634 Poisson Process.................................................................................................635 Characteristics of Poisson Process ..............................................................636 Poisson Distribution .....................................................................................636 Example........................................................................................................639 Weibull Distribution...............................................................................................640 Three-Parameter Weibull Distribution..............................................................643 Taylor Series Expansion ........................................................................................644 Taylor Series Expansion ...................................................................................645 Partial Derivatives ........................................................................................649 Taylor Series in Two-Dimensions................................................................649 Taylor Series of Random Variable (RV) Functions.....................................650 Variance and Covariance..............................................................................650 Functions of Random Variables...................................................................651 Division of Random Variables .....................................................................651 Powers of a Random Variable .....................................................................652 Exponential of a Random Variable..............................................................652 Constant Raised to RV Power .....................................................................653 Logarithm of Random Variable ...................................................................653 Example: Horizontal Beam Deflection........................................................654 Example: Difference between Two Means..................................................655 Miscellaneous ........................................................................................................656 Closing Remarks....................................................................................................658 Selected Bibliography............................................................................................658 Chapter 15 Fundamentals of Finance and Accounting for Champions, Master Blacks, and Black Belts ............................................................................661 The Theory of the Firm.........................................................................................661 Budgets ..................................................................................................................662 Our Romance with Growth ...................................................................................663 The New Industrial State.......................................................................................663

SL3151 FMFrame Page 49 Friday, September 27, 2002 3:14 PM

Behavioral Theory .................................................................................................663 Accounting Fundamentals .....................................................................................664 Accounting’s Role in Business.........................................................................664 Financial Reports ..............................................................................................664 The Balance Sheet .......................................................................................664 Current Assets and Liabilities .................................................................665 Fixed Assets.............................................................................................665 Other Slow Assets ...................................................................................666 Current Liabilities ...................................................................................666 Working Capital Format..........................................................................666 Noncurrent Assets ...................................................................................667 Noncurrent Liabilities .............................................................................667 Shareholders’ Equity ...............................................................................667 The Income Statement ............................................................................667 Gross Profit..............................................................................................668 A Gaggle of Profits .................................................................................668 Earnings per Share ..................................................................................669 The Statement of Changes ......................................................................669 Sources of Funds or Cash .......................................................................669 Use of Funds ...........................................................................................670 Changes in Working Capital Items .........................................................670 The Footnotes ..........................................................................................670 Accountants’ Report.....................................................................................671 How to Look at an Annual Report ..............................................................671 Recording Business Transactions .....................................................................672 Debits and Credits........................................................................................673 Sources and Uses of Cash.......................................................................673 How Debits and Credits Are Used .........................................................673 The Balance Sheet Equations ......................................................................673 Classification of Accounts ................................................................................674 Recording Transactions................................................................................675 The Two Books of Account.........................................................................675 The Trial Balance ....................................................................................676 The Mirror Image....................................................................................676 Accrual Basis of Accounting............................................................................676 Accrual Basis versus Cash Basis.................................................................677 Details, Details .............................................................................................677 Birth of the Balance Sheet...........................................................................678 Profits versus Cash.......................................................................................678 Things Are Measured in Money..................................................................678 Values Are Based on Historical Costs.........................................................678 Understanding Financial Statements .....................................................................679 Assets ................................................................................................................679 The Inflation Effect...........................................................................................679 Summary of Valuation Methods .......................................................................679 Historical Cost..............................................................................................679

SL3151 FMFrame Page 50 Friday, September 27, 2002 3:14 PM

Liquidation Value .........................................................................................679 Investment or Intrinsic Value .......................................................................680 Psychic Value ...............................................................................................680 Current Value or Replacement Cost ............................................................680 Assets versus Expenses................................................................................680 Types of Assets .................................................................................................681 Financial Assets............................................................................................681 Physical Assets.............................................................................................681 Operating Leverage .................................................................................682 Determining the Value of Inventory .......................................................682 FIFO....................................................................................................682 LIFO ...................................................................................................682 Weighted Average...............................................................................683 Depreciation ............................................................................................683 Useful Life Concept ................................................................................683 Depreciation as an Expense ...............................................................684 Depreciation as a Valuation Reserve..................................................684 Depreciation as a Tax Strategy ..........................................................684 Depreciation as Part of Cash Flow ....................................................685 Straight Line .......................................................................................685 Sum-of-the-Years’ Digits (SYD)........................................................686 Double Declining Balance (DDB) .....................................................686 Unit of Production..............................................................................687 Replacement Cost...............................................................................687 Advantages of Accelerated Depreciation ...........................................687 Financial Statement Analysis ................................................................................688 Ratio Analysis ...................................................................................................688 Liquidity Ratios............................................................................................691 Financial Leverage .......................................................................................692 Coverage Ratios ...........................................................................................692 Earnings........................................................................................................692 Earnings Ratios ............................................................................................693 Le ROI .....................................................................................................693 ROE: Return on Equity ......................................................................694 ROA: Return on Assets ......................................................................694 ROS: Return on Sales ........................................................................694 Other Return Ratios............................................................................694 Financial Rating Systems ......................................................................................695 Bond Rating Companies...................................................................................695 Moody’s et al. ..............................................................................................695 Moody’s...................................................................................................696 Standard and Poor’s ................................................................................696 Ratings on Common Stocks .............................................................................696 The S&P Rating Method .............................................................................697 The Value Line Method ...............................................................................697 Good Ole Ben Graham ................................................................................697

SL3151 FMFrame Page 51 Friday, September 27, 2002 3:14 PM

Commercial Credit Ratings ..............................................................................698 Dun & Bradstreet .........................................................................................698 Other Systems ..............................................................................................699 Company and Product Life Cycle.........................................................................699 Cash Flow .........................................................................................................700 A Final Thought about Cash Flow...................................................................701 A Handy Guide to Cost Terms.........................................................................703 Useful Concepts for Financial Decisions..............................................................704 The Modified duPont Formula .........................................................................704 Breakeven Analysis...........................................................................................705 Contribution Margin Analysis ..........................................................................706 Price–Volume Variance Analysis ......................................................................707 Inventory’s EOQ Model....................................................................................707 Return on Investment Analysis.........................................................................708 Net Present Value (NPV) .............................................................................709 Internal Rate of Return (IRR)......................................................................709 Profit Planning .......................................................................................................710 The Nature of Sales Forecasting ......................................................................710 The Plans Up Form......................................................................................710 Statistical Analysis .......................................................................................711 Compound Growth Rates........................................................................711 Regression Analysis ................................................................................711 Revenues and Costs.................................................................................711 Departmental Budgets .............................................................................711 How to Budget ........................................................................................712 Zero-Growth Budgeting ..........................................................................712 Selected Bibliography............................................................................................712 Chapter 16 Closing Thoughts about Design for Six Sigma (DFSS) ...............715 Appendix The Four Stages of Quality Function Deployment ..........................725 Stage 1: Establish Targets......................................................................................725 Stage 2: Finalize Design Timetables and Prototype Plans ...................................725 Stage 3: Establish Conditions of Production ........................................................725 Stage 4: Begin Mass Production Startup ..............................................................726 Tangible Benefits ...................................................................................................726 Intangible Benefits .................................................................................................727 Summary Value......................................................................................................727 The QFD Process...................................................................................................727 Managing the Process............................................................................................728 Selected Bibliography ..........................................................................................731 Index ......................................................................................................................737

SL3151 FMFrame Page 52 Friday, September 27, 2002 3:14 PM

SL3151Ch00Frame Page 1 Thursday, September 12, 2002 6:15 PM

Introduction — Understanding the Six Sigma Philosophy Much discussion in recent years has been devoted to the concept of “six sigma” quality. The company most often associated with this philosophy is Motorola, Inc., whose definition of this principle is stated by Harry (1997, p. 3) as follows: A product is said to have six sigma quality when it exhibits no more than 3.4 npmo at the part and process step levels.

Confusion often exists about the relationship between six sigma and this definition of producing no more than 3.4 nonconformities per million opportunities. From a typical normal distribution table, one may find that the area underneath the normal curve beyond six sigma from the average is 1.248 × 10–9 or .001248 ppm, which is about 1 part per billion. Considering both tails of the process distribution, this would be a total of .002 ppm. This process has the potential capability of fitting two six sigma spreads within the tolerance, or equivalently, having 12 σ equal the tolerance. However, the 3.4 ppm value corresponds to the area under the curve at a distance of only 4.5 sigma from the process average. Why this apparent discrepancy? It is due to the difference between a static and a dynamic process. (The reader is encouraged to review Volume I of this series.)

A STATIC VERSUS A DYNAMIC PROCESS If a process is static, meaning the process average remains centered at the middle of the tolerance, then approximately .002 ppm will be produced. But under the six sigma concept, the process is considered to be dynamic, implying that over time, the process average will move both higher and lower because of many small changes in material, operators, environmental factors, tools, etc. Most small shifts in the process average will go undetected by the control chart. For an n of 4, there is only a 50 percent chance a 1.5-sigma shift in µ is detected by the next subgroup after this change. By the time this next subgroup is collected, it may have returned to its original position. Thus, this process change will never be noticed on the chart, which means that no corrective action will be implemented. However, this movement has caused the actual long-term process variation to increase somewhat because betweensubgroup variation is greater than within-subgroup variation. Note that estimates of short-term process variation are not impacted because they are determined only from within-subgroup variation. 1

SL3151Ch00Frame Page 2 Thursday, September 12, 2002 6:15 PM

2

Six Sigma and Beyond

Based on studies analyzing the effect of these changes on process variation (Bender, 1962, 1968; Evans, 1970, 1974, 1975a and b; Gilson, 1951), the six sigma principle acknowledges the likelihood of undetected shifts in the process average of up to ±1.5 sigma. Because shifts in the average greater than 1.5 sigma are expected to be caught, and six is assumed not to change, the worst case for the production of nonconforming parts happens when the process average has shifted either the full 1.5 sigma above the middle of the tolerance or the full 1.5 sigma below it. For this worst case, there would be only 4.5 sigma (6 sigma minus 1.5 sigma) remaining between the process average and the nearest specification limit. This reduced Z value of 4.5 for the dynamic model corresponds to 3.4 ppm. When this size of shift occurs, the Z value for the other specification limit becomes 7.5, which means essentially 0 ppm are outside this limit. Because the process average can shift in only one direction at a time, the maximum number of nonconforming parts produced is 3.4 ppm. Notice that most of the time the average should be closer to the middle of the tolerance, resulting in far fewer than 3.4 ppm actually being manufactured. To achieve a goal of 3.4 ppm, the process average must be no closer than 4.5 sigma to a specification limit. Assuming the average could drift by as much as 1.5 sigma, potential capability must be at least 6.06 (4.56 plus 1.5 sigma) to compensate for shifts in the process average of up to 1.56, yet still be able to produce the desired quality level. The required 4.56 plus this added buffer of 1.5 sigma create the 6 σ requirement, and thereby generate the label “six sigma.” (Here it must be noted that the 4.5 shift is allegedly an empirical value for the electronic industry. In the automotive industry, for years the shift has been identified as only 1 sigma — a shift from a Ppk of 1.33 to a Cpk of 1.67 i.e., from 4 sigma to 5 sigma. The point is that every industry should identify its own shift and use it accordingly. It is unfortunate that the 4.5 shift has become the default value for everything. For a detailed explanation on the difference between Ppk and Cpk, the reader is encouraged to review Volumes I and IV of this series.) To counter the effect of shifts in µ, a buffer of 1.5 standard deviations can be added to other capability goals as well. If no more than 32 ppm are desired outside either specification, the goal would be to have ±4.06 fit within the tolerance, assuming no change in the process average. This target equates to a Cp of 1.33 (4.0/3). Under the static model, this potential capability goal translates into 32 ppm outside each specification when the average is centered at M. But with the inevitable 1.56 drifts in µ occurring with the dynamic process model, the average could move as close as 2.56 (4.5 sigma minus 1.5 sigma) to a specification limit before triggering any type of corrective action. This change in centering would cause as many as 6210 ppm to be produced, quite a bit more than the desired maximum of 32 ppm.

PRODUCTS WITH MULTIPLE CHARACTERISTICS Extremely low ppm levels are imperative for producing high quality products possessing many characteristics (or components). Table I.1 compares the probability of manufacturing a product with all characteristics inside their respective specifications when each is produced with ±4 sigma (Cp = 1.33) versus ±6 sigma (Cp = 2.00)

SL3151Ch00Frame Page 3 Thursday, September 12, 2002 6:15 PM

Introduction — Understanding the Six Sigma Philosophy

3

TABLE I.1 Probability of a Completely Conforming Product With 1.56 Shift Number of Characteristics

C, = 1.33 (±46)

C,. = 2.00 (±6a)

1 2 5 10 25 50 100 150 250 500

99.3790 98.7618 96.9333 93.9607 85.5787 73.2371 53.6367 39.2820 21.0696 4.4393

99.99966 99.99932 99.9983 99.9966 99.9915 99.9830 99.9660 99.9490 99.9150 99.8301

capability. The processes producing the features are assumed to be dynamic, with up to a 1.5-sigma shift in average possible. Suppose a product has only one feature, which is produced on a process having ±4 sigma potential capability. We can then calculate that a maximum of .6210 percent of these parts will be non-conforming under the dynamic model. Conversely, at least 99.3790 percent will be conforming, as is listed in the first line of Table I.1. If this single characteristic is instead produced on a process with ±6 sigma potential capability, at most .00034 percent of the finished product will be out of specification, with at least 99.99966 percent within specification. If a product has two characteristics, the probability that both are within specification (assuming independence) is .993790 times .993790, or 98.7618 percent when each is produced on a ±4 sigma process. If they are produced on a ±6 sigma process, this probability increases to 99.99932 percent (.9999966 times .9999966). The remainder of the table is computed in a similar manner. When each characteristic is produced with ±4 sigma capability (and assuming a maximum drift of 1.5 sigma), a product with 10 characteristics will average about 939 conforming parts out of every 1000 made, with the 61 nonconforming ones having at least one characteristic out of specification. If all characteristics are manufactured with ±6 sigma capability, it would be very unlikely to see even one nonconforming part out of these 1000. For a product having 50 characteristics, 268 out of 1000 parts will have at least one nonconforming characteristic when each is produced with ±4 sigma capability. If these 50 characteristics were manufactured with ±6 sigma capability, it would still be improbable to see one nonconforming part. In fact, with ±6 sigma capability, a product must have 150 characteristics before you would expect to find even one nonconforming part out of 1000. Contrast this to the ±4 sigma capability level, where 60.7 percent of these parts would be rejected, and the rationale for adopting the six sigma philosophy becomes quite evident.

SL3151Ch00Frame Page 4 Thursday, September 12, 2002 6:15 PM

4

Six Sigma and Beyond

SHORT- AND LONG-TERM SIX SIGMA CAPABILITY The six sigma approach also differentiates between short- and long-term process variation. Just as in the past, the short-term standard deviation has been estimated from within-subgroup variation, usually from R, and the long-term standard deviation incorporates both the short-term variation and any additional variation in the process introduced by the small, undetected shifts in the process average that occur over time. Although no exact relationship between these two types of variation applies to every kind of process, the six sigma philosophy ties them together with this general equation (Harry and Lawson, 1992, pp. 6–8). σ LT = cσ ST

As c is affected by shifts in the process average, it is related to the k factor, which quantifies how far the process average is from the middle of the tolerance.

c=

µ−M 1 k= (USL − LSL ) / 2 1− k

If a process has a Cp of 2.00 and is centered at the middle of the tolerance, then there is a distance of 6σST from the average to the USL. When the process average shifts up by 1.5σST, it has moved off target by 25 percent of one-half the tolerance (1.5/6.0 = .25). For this k factor of .25, c is calculated as 1.33. C = 1/(1 – .25) = 1/.75 = 1.33 The long-term standard deviation for this process would then be estimated from σST, as: ˆ LT = cσ ST = 1.33σ ST σ

The value 1.33 is quite commonly adopted as the relationship between shortand long-term process variation (Koons, 1992). This factor implies that long-term variation is approximately 33 percent greater than short-term variation. Other authors are more conservative and assume a c factor between 1.40 and 1.60, which translates to a k factor ranging from .286 to .375 (Harry and Lawson, 1992, pp. 6–12, 7–6). For a c factor of 1.50, k is .333. 1.50 = 1/(1 – k) 1 – k = 1/1.50 k = 1 – (1/1.50) = .333

SL3151Ch00Frame Page 5 Thursday, September 12, 2002 6:15 PM

Introduction — Understanding the Six Sigma Philosophy

5

This assumption expects up to a 33.3 percent shift in the process average. With six sigma capability, there is 6σST from M to the specification limit, a distance that equals one-half the tolerance. A k factor of .333 represents a maximum shift in the process average of 2.0σST, a number derived by multiplying one-half the tolerance, or 6σST, by .333.

DESIGN FOR SIX SIGMA AND THE SIX SIGMA PHILOSOPHY The six sigma philosophy is becoming more and more popular in the quality field, especially with companies in the electronics industry (de Treville et al., 1995). Organizations striving to attain the quality levels required with the six sigma system usually adopt the following three recommended strategies for accomplishing this goal (Tomas (1991) offers a six-step approach). Improving an existing process to the six sigma level of quality would be very difficult, if not impossible. That is why Fan (1990) insists this type of thinking must already be incorporated into the original design of new products and the processes that will manufacture them if there is to be any chance of achieving six sigma quality. The three recommended strategies are as follows:

DESIGN PHASE 1. Design in ±6σ tolerances for all critical product and process parameters. For additional information on this topic, read Six Sigma Mechanical Design Tolerancing by Harry and Stewart (1988). 2. Develop designs robust to unexpected changes in both manufacturing and customer environments (see Harry and Lawson, 1992). 3. Minimize part count and number of processing steps. 4. Standardize parts and processes. Knowing the process capability of current manufacturing operations will greatly aid designers in accomplishing this first step. And of course, good designs will positively influence the capability of future processes. Once a new product is released for production, the designed-in quality levels must be maintained, and even improved upon, by working to reduce (or eliminate) both assignable and common causes of process variation. McFadden (1993) lists several additional key components of a six sigma quality program specifically targeted at manufacturing.

INTERNAL MANUFACTURING 1. Standardize manufacturing practices. 2. Audit the manufacturing system. Pena (1990) provides a detailed audit checklist for this purpose.

SL3151Ch00Frame Page 6 Thursday, September 12, 2002 6:15 PM

6

Six Sigma and Beyond

3. Use SPC to control, identify, and eliminate causes of variation in the manufacturing process. Mader et al. (1993) have written a book entitled Process Control Methods to help with this step. The reader may also review Volume IV of this series. 4. Measure process capability and compare to goals. Koons’ (1992) and Bothe’s (1997) books on capability indices are useful here. 5. Consider the effects of random sampling variation on all six sigma estimates and apply the proper confidence bounds. The reference by Tavormina and Buckley (1992) would be helpful here. 6. Kelly and Seymour (1993), Bothe (1993), and Delott and Gupta (1990) reveal how the application of statistical techniques helped achieve six sigma quality levels for copper plating ceramic substrates. Harry (1994) provides several examples of applying design of experiments to improve quality in the electronics industry. A special warning here is appropriate. Even if the first two strategies are adopted, a company will never achieve six sigma quality unless it has the full cooperation and participation of all its suppliers.

EXTERNAL MANUFACTURING 1. 2. 3. 4. 5.

Qualify suppliers. Minimize the number of suppliers. Develop long-term partnerships with remaining suppliers. Require documented process control plans. Insist on continuous process improvement.

Craig (1993) shows how Dupont Connector Systems utilized this set of strategies to introduce new products into the data processing and telecommunications industries. Noguera (1992) discusses how the six sigma doctrine applies to chip connection technology in electronics manufacturing, while Fontenot et al. (1994) explain how these six sigma principles pertain to improving customer service. Daskalantonakis et al. (1990–1991) describe how software measurement technology can identify areas of improvement and help track progress toward attaining six sigma quality in software development. As all these authors conclude, the rewards for achieving the six sigma quality goals are shorter cycle times, shorter lead times, reduced costs, higher yields, improved product reliability, increased profitability, and most important of all, highly satisfied customers. We have reviewed the principles of six sigma here to make sure the reader understands the ramifications of poor quality and the significance of implementing the six sigma philosophy. In Volume I of this series, we discussed this philosophy in much more detail. However, it is imperative to summarize some of the inherent advantages, as follows:

SL3151Ch00Frame Page 7 Thursday, September 12, 2002 6:15 PM

Introduction — Understanding the Six Sigma Philosophy

7

1. As quality improves to the six sigma level, profits will follow with a margin of about 8% higher prices. 2. The difference between a six sigma company and a non–six sigma company is that the six sigma company is three times more profitable. Most of that profitability is through elimination of variability — waste. 3. Companies with improved quality gain market share continuously at the expense of companies that do not improve. The focus of all these great results is in the manufacturing. However, most of the cost reduction is not in manufacturing. We know from many studies and the experience of management consultants that about 80% of quality problems are actually designed into the product without any conscious attempt to do so. We also know that about 70% of a product’s cost is determined by its design. Yet, most of the “hoopla” about six sigma in the last several years has been about the DMAIC model. To be sure, in the absence of anything else, the DMAIC model is great. But it still focuses on after-the-fact problems, issues, and concerns. As we keep on fixing problems, we continually generate problems to be fixed. That is why Stamatis (2000) and Tavormina and Buckley (1994) and the first volume of this series proclaimed that six sigma is not any different from any other tool already in the tool box of the practitioner. We still believe that, but with a major caveat. The benefit of the six sigma philosophy and its application is in the design phase of the product or service. It is unconscionable to think that in this day and age there are organizations that allow their people to chase their tails and give accolades to so many for fixing problems. Never mind that the problems they are fixing are repeatable problems. It is an abomination to think that the more we talk about quality, the more it seems that we regress. We believe that a certification program will do its magic when in fact nothing will lead to real improvement unless we focus on the design. This volume is dedicated to the Design for Six Sigma, and we are going to talk about some of the most essential tools for improvement in “real” terms. Specifically, we are going to focus on resource efficiency, robust designs, and production of products and services that are directly correlated with customer needs, wants, and expectations.

REFERENCES Bender, A, Bendarizing Tolerances — A Simple Practical Probability Method of Handling Tolerances for Limit-Stack-Ups. Graphic Science, Dec. 1962, pp. 17–21. Bender, A., Statistical Tolerancing as It Relates to Quality Control and the Designer, SAE Paper No. 680490, Society of Automotive Engineers, Southfield, MI, 1968. Bothe, D.R., Reducing Process Variation, International Quality Institute., Inc., Sacramento, CA, 1993. Bothe, D.R., Measuring Process Capability, McGraw-Hill. New York, 1997. Craig, R.J., Six-Sigma Quality, the Key to Customer Satisfaction, 47th ASQC Annual Quality Congress Transactions, Boston, 1993, pp. 206–212.

SL3151Ch00Frame Page 8 Thursday, September 12, 2002 6:15 PM

8

Six Sigma and Beyond

Daskalantonakis, M.K., Yacobellis, R.H., and Basili, V.R., A method for assessing software measurement technology, Quality Engineering 3, 27–40, 1990–1991. Delott, C. and Gupta, P., Characterization of copperplating process for ceramic substrates, Quality Engineering, 2, 269–284, 1990. de Treville, S., Edelson, N.M., and Watson, R., Getting six sigma back to basics, Quality Digest, 15, 42–47, 1995. Evans, D.H., Statistical tolerancing formulation, Journal of Quality Technology, 2, 188–195, 1970. Evans, D.H., Statistical tolerancing: the state of the art, Part I: Background, Journal of Quality Technology, 6, 188–195, 1974. Evans, D.H., Statistical tolerancing: the state of the art, Part II: Methods for estimating moments, Journal of Quality Technology, 7, 1–12. 1975 (a). Evans, D.H., Statistical tolerancing: the state of the art, Part III: Shifts and drifts, Journal of Quality Technology, 7, 72–76, 1975 (b). Fan, John Y. (May 1990). Achieving Six Sigma in Design, 44th ASQC Annual Quality Congress Transactions, San Francisco, May 1990, pp. 851–856. Fontenot, G., Behara, R., and Gresham, A., Six sigma in customer satisfaction, Quality Progress, 27, 73–76, 1994. Gilson, J., A New Approach to Engineering Tolerances, Machinery Publishing Co., London, 1951. Harry, M., The Nature of Six Sigma Quality, Motorola Univ. Press, Schaumburg, IL, 1997. Harry, M. and Stewart, R., Six Sigma Mechanical Design Tolerancing, Motorola University Press, Schaumburg, IL, 1988. Harry, M., The Vision of Six Sigma: Case Studies and Applications, 2nd ed., Sigma Publishing Co., Phoenix, 1994. Harry, M. and Lawson, J.R., Six Sigma Producibility Analysis and Process Characterization, Addison-Wesley Publishing Co., Reading, MA, 1992. Kelly, H.W. and Seymour, L.A., Data Display. Addison-Wesley Publishing Co., Reading, MA, 1993. Koons, J., Indices of Capability: Classical and Six Sigma Tools, Addison-Wesley Publishing Co., Reading, MA, 1992. Mader, D.P., Seymour, L.A., Brauer, D.C., and Gallemore, M.L., Process Control Methods, Addison-Wesley Publishing Co., Reading, MA, 1993. McFadden, F.R., Six-sigma quality programs, Quality Progress, 26, 37–42, 1993. Noguera, J., Implementing Six Sigma for Interconnect Technology, 46th ASQC Annual Quality Congress Transactions, Nashville, TN, May 1992, pp. 538–544. Pena, E., Motorola’s secret to total quality control, Quality Progress, 23, 43–45, 1990. Stamatis, D.H., Six sigma: point/counterpoint: who needs six sigma anyway, Quality Digest, 33–38, May, 2000. Tadikamalla, P.R., The confusion over six-sigma quality, Quality Progress, 21, 83–85, 1994. Tavormina, J.J., and Buckley, S., SPC and six-sigma, Quality, 31, 47, 1992. Tomas, S., Motorola’s Six Steps to Six Sigma, 34th International Conference Proceedings, APICS, Seattle, WA, 1991. pp. 166–169.

SL3151Ch01Frame Page 9 Thursday, September 12, 2002 6:15 PM

1

Prerequisites to Design for Six Sigma (DFSS)

So far in this series we have presented an overview of the six sigma methodology (DMAIC) and some of the tools and specific methodologies for addressing problems in manufacturing. Although this is a commendable endeavor for anyone to pursue — as mentioned in Volume I of this series — it is not an efficient way to use resources to pursue improvement. The reason for this is the same as the reason you do not apply an atomic bomb to demolish a two-story building. It can be done, but it is a very expensive way to go. As we proposed in Volume I, if an organization really means business and wants quality improvement to go beyond six sigma constraints, it must focus on the design phase of its products or services. It is the design that produces results. It is the design that allows the organization to have flexibility. It is the design that convinces the customer of the existence of quality in a product. Of course, in order for this design to be appropriate and applicable for customer use, it must be perceived by the customer as functional, not by the organization’s definition but by the customer’s personal perceived understanding and application of that product or service. Design for Six Sigma (DFSS) is an approach in which engineers interpret and design the functionality of the customer need, want, and expectation into requirements that are based on a win-win proposition between customer and organization. Why is this important? It is important because only through improved quality and perceived value will the customer be satisfied. In turn, only if the customer is satisfied will the competitive advantage of a given organization increase. There are four prerequisites to DFSS and beyond. The first is the recognition that improvement must be a collaboration between organization and supplier (partnering). The second is the recognition that true DFSS and beyond will only be achieved if in a given organization there are “real” teams and those teams are really “robust.” The third prerequisite is that improvement on such a large scale can only be achieved by recognizing that systems engineering must be in place. Its function has to be to make sure that the customer’s needs, wants, and expectations are cascaded all the way to the component level. The fourth prerequisite is the implementation of at least a rudimentary system of Advanced Quality Planning (AQP). In this chapter we will address each of these prerequisites in a cursory format. (Here we must note that these prerequisites have also been called the “recognize” phase of the six sigma methodology.) In the follow-up chapters, we will discuss specific tools that we need in the pursuit of DFSS and beyond.

PARTNERING Partnering and cooperation must be our watchwords. In any industry, better communication up and down the supply chain is mandatory. In the past — in few 9

SL3151Ch01Frame Page 10 Thursday, September 12, 2002 6:15 PM

10

Six Sigma and Beyond

instances even today — U.S. companies have bought almost solely on the basis of price through competitive bidding. We need to change our attitude. Price is important, but it is not the only consideration. Partnering with both customers and suppliers is just as important. The Japanese have created a competitive edge through vertical integration. We can learn from them by establishing “virtual” vertical integration through partnering with customers and suppliers. Just as in a marriage, we need to give more than we get and believe that it will all work out better in the end. We need to give preferential treatment to local suppliers. We should take a long-term view, understanding their need for profitability and looking beyond this year’s buy. To begin our thinking in that direction we must change our current paradigm. The first paradigm shift must be in the following definitions: 1. Vendors must be viewed as suppliers. 2. Procurement must be viewed as business strategy. These are small changes indeed but they mean totally different things. For example: “supplier” implies working together in a win-win situation, while “vendor” implies a one-time benefit — usually price. “Procurement” implies price orientation based on bidding of some sort, while “business strategy” takes into account the concern(s) of the entire organization. We all know that price alone is not the sole reason we buy. If we do buy on the basis of price alone, we pay the consequences later on. So, what is partnering? Partnering is a business culture that fosters open communication and mutually beneficial relationships in a supportive environment built on trust. Partnering relationships stimulate continuous quality improvement and a reduction in the total cost of ownership. Partnering starts with: 1. An attitude and behavioral change at the top of the organization 2. Recognition of long-term mutual dependencies internal and external to the organization 3. A commitment to this change being understood and valued at all levels within the organization At the core or basic level, partnering: 1. Fosters excellence throughout the organization 2. Encourages open communication in a beneficial, supportive, and nonadversarial environment of mutual trust and respect 3. Carries this positive environment outward from the organization to its customers and suppliers At an expanded level, partnering involves:

SL3151Ch01Frame Page 11 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

1. 2. 3. 4.

11

Teaming Sharing resources Melding of customer and supplier Eliminating the we/they approach to conducting business

By the same token, partnering is not: 1. A negotiation or purchasing tool to be used as leverage against the supplier 2. A business guarantee However, in all cases, partnering promotes: 1. 2. 3. 4.

Customer satisfaction Mutual profitability Improved product, service, and operational quality A desire for and a commitment to excellence through continuous improvements in communication skills, quality, delivery, administration, and service performance 5. The factors that contribute to customer satisfaction and the lowest total cost of ownership 6. A situation in which each partner enhances its own competitive position through the knowledge and resources shared by the other

THE PRINCIPLES

OF

PARTNERING

Effective partnering has its foundation in the basic principles of economics, marketing, business, humanities, and sociology. The customer develops a set of business and technical desires, needs, requirements, and expectations in a competitive global market. The supplier most closely meeting those business and technical needs will be successful. The supplier asks the customer what is wanted rather than telling the customer what is available. The customer recognizes and understands the supplier’s business and technical requirements, allowing the supplier to be a viable and successful source to the industry. All transactions are honorable and fair. The parties are not trying to take advantage of each other. Functioning interchangeably each day as customer and supplier, internally within the organization and externally with customers and suppliers, every person in a strong supply chain recognizes mutual dependencies. All transactions must be mutually beneficial, with each person encouraging open communication and operating with integrity, mutual trust, cooperation, and respect.

VIEW

OF

BUYER/SUPPLIER RELATIONSHIP: A PARADIGM SHIFT

Partnering involves an expanded view of the buyer/supplier relationship, as shown here:

SL3151Ch01Frame Page 12 Thursday, September 12, 2002 6:15 PM

12

Six Sigma and Beyond

Traditional

Expanded

Lowest price Specification-driven Short-term, reacts to market Trouble avoidance Purchasing’s responsibility Tactical Little sharing of information on both sides

Total cost of ownership End customer–driven Long-term Opportunity maximization Cross-functional teams and top management involvement Strategic Both supplier and buyer share short- and long-term plans Share risk and opportunity Standardization Joint venture Share data

How can this partnership develop? There are prerequisites. Some are listed here. The prerequisites for basic partnering include: 1. 2. 3. 4. 5.

Mutual respect Honesty Trust Open and frequent communication Understanding of each other’s needs

Additional prerequisites for expanded partnering include: 6. 7. 8. 9. 10. 11. 12.

Long-term commitment Recognition of continuing improvement — objective and factual Passion to help each other succeed High priority on relationship Shared risk and opportunity Shared strategies/technology road maps Management commitment

CHARACTERISTICS

OF

EXPANDED PARTNERING

Expanded partnering promotes dedication, desire, and commitment to product and service excellence through improvements in technology, skills, quality, delivery, administration, responsiveness, and total cost of ownership. All these are imperative requirements for DFSS. In other words, expanded partnering: 1. 2. 3. 4.

Builds on basic partnering Is a long-term relationship process Provides focus on mutual strategic and tactical goals Includes customer/supplier team support to promote mutual success and profitability.

Of course, there are different levels of partnering just as there are different levels of results. For example:

SL3151Ch01Frame Page 13 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS) Results

Partnering Focus

Sale only Loyalty/trust Secured volumes Mutual improvements Mutual breakthrough

Short term Product Product and service Process or system Continual improvement

13 Stage 1 2 3 4 5

Why is partnering so important in the DFSS even though it may mean different things to different people? It is because the purposes or goal of most customers who advocate “partnerships” are to reduce the time to get a new product to market by eliminating the bid cycle and to extend the customer’s capability without adding personnel. Partnering is joining together to accomplish an objective that can best be met by two individuals or corporations rather than one. For a partnership to work well, it requires that both partners understand the objective, each partner complements the other in skills necessary to meet the objective, and each recognizes the value of the other in the relationship. A true partnership occurs when both partners make a conscious decision to enter into a unique relationship. As the partnership develops, trust and respect build to a degree that both share the joy and rewards of success and, when things do not go so well, both work hard together to resolve the issues to mutual satisfaction. In a customer/supplier partnership, the customer must define the objective (or the scope of the project) and identify the needs. The supplier must have the capability to meet the customer’s needs and become an extension of the customer’s resources. To be more specific, the customer must be able to quantify and share the desired needs in terms of the quantity of services required, the timeline or critical path desired, and targeted costs — including up-front engineering as well as unit cost and capital investment. The supplier must determine whether it can commit the resources required to meet those needs and whether it is capable of reaching the targets. A mutual commitment must be made early in the program, and it must be for the life of the program. In a more practical sense, the customer in a customer/supplier partnership must be the leader and be in a position to guide the partners to the objective — no different than a project leader or a team leader of a program that is 100 percent internal to the customer. The leader also must monitor the progress in terms of cost and time with input from the supplier. Our experience would indicate that longer projects should be broken into “phases” so that there are milestones that are mutually agreed to in advance by the partners and that mark the points at which the supplier is paid for its services. For a partnership to work well, customer/supplier communications must be open and frequent. With the availability of CAD, e-mail, Internet, Web sites, fax, and voice mail, there should be no reason not to communicate within minutes of recognition of an issue critical to the program, but there is also a need for regular meetings at predetermined intervals at either the customer’s or supplier’s location (probably with some meetings at each location to expose both partners to as many of the team players as possible).

SL3151Ch01Frame Page 14 Thursday, September 12, 2002 6:15 PM

14

Six Sigma and Beyond

I am sure there is more to be said as to why partnership and DFSS work in tandem and why both strive for mutual benefits, but I hope these thoughts gave some idea of the significance that both have for each other.

EVALUATING SUPPLIERS

AND

SELECTING SUPPLIER PARTNERS

There are many schemes to evaluate suppliers, and each of them has advantages and disadvantages. We believe, however, that each organization should take the time to generate its own criteria in at least two dimensions. The first should be the supplier’s situation and the second the purchaser’s situation. Within each category, levels of satisfaction may be assessed as total dissatisfaction, partial satisfaction, or total satisfaction, or numerical values may be used. The higher the number, the more qualified the supplier is. This may be done with either a questionnaire or a matrix. In either case, this task should be performed by a team of people from various functional areas, such as purchasing, engineering, finance, quality, and legal. The important point is to evaluate key suppliers for a fit with your company’s needs.

IMPLEMENTING PARTNERING There are five steps to partnering. They are: 1. Establish Top Management Enrollment (Role of Top Management — Leadership) The senior management, in the role of an executive customer partner or executive supplier partner (champion): 1. 2. 3. 4. 5. 6. 7. 8. 9.

Serves in a long-term assignment for each expanded partnering relationship Is available to support prompt issue resolution Establishes strong counterpart relationships with key customers and suppliers Provides for and supports decision-making authority at the lowest practical levels Provides partnering progress updates for executive management review Encourages and supports prompt responsiveness to communications affecting customer/supplier relationships Maintains a rapid management approval cycle, providing an ombudsman when required Commits adequate time to the partnering process Ensures that cohesive internal, cross-functional teams are in place to support the partnering process

2. Establish Internal Organization There are several options in this phase. However, the most common are: Option 1: Supplier Partnering Manager A staff supplier partnering manager is appointed to a full-time position (for a minimum of two years). This manager will be responsible for:

SL3151Ch01Frame Page 15 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

15

1. Working with purchasing/commodity team management 2. Instilling the partnering principles into the company culture 3. Implementing the partnering process with company management and suppliers 4. Reviewing progress during customer/supplier review sessions 5. Working the issues specific to the partnering process Option 2: Supplier Council/Team A supplier partnering council or team is established within the organizational and operational structure that “owns” the resources required to support the partnering process. The functions are the same as for the supplier partnering manager but are assigned to several individuals. Typically, the council or team is made up of purchasing, quality, product engineering, and manufacturing management with additional resources available from finance, law, training, and other departments as required. Option 3: Commodity Management Organization A line organization consisting of a commodity manager and staff is created to manage the commodity and the partnering activities described in Option 1. Support is received from the operational groups as required. 3. Establish Supplier Involvement To have an effective partnering involvement is of paramount importance. This involvement may be encouraged and helped to grow by having open communication. Communication may be conducted in a variety of forums or as scheduled periodic meetings — see Table 1.1. 4. Establish Responsibility for Implementation Identify roles and responsibilities of the partnering process manager: 1. 2. 3. 4. 5.

Serve as customer representative. Serve as supplier advocate. (Avoid conflict of interest.) Focus participants on long-term success. Accelerate and route communications (good news, bad news). Perform meeting planning (with supplier) and facilitation function.

Perhaps one of the most important functions in this step is to establish credibility with each other as well as confidentiality requirements. The process of this exchange must be truthful and full of integrity. Some characteristics of this exchange are: 1. Each party provides the other with the information needed to be successful. 2. The supplier needs to know the customer’s requirements and expectations in order to meet them on a long-term basis.

Establish/update mutual key results, goals, objectives, action plans Discuss issues Review performance Review/discuss on-time deliveries Required actions of both parties Quality indicators Quality action plan Business issues

Purchasing Technical Quality/reliability (Other team members)

Monthly Team Meeting

Major issues Performance review “Health check” Objectives Expectations Actual performance Technology trends Business trends Program direction

Purchasing Technical Quality/reliability (Other team members) Executive partnersb

Quarterly/Semiannual Management Meeting

At supplier location and tour Maintain key contacts Major performance review

Purchasing Technical Quality/reliability (Other team members) Executive partners

Annual Management Review

Team includes personnel from Purchasing, Quality, Material Control, Engineering. When needed, also can include personnel from Sales, Safety, Manufacturing, Process Area Management, Planning, Training, Legal, Risk Management, Finance, Project Management. b Optional as part of quarterly and semiannual meetings.

Introduce program Obtain mutual agreement and commitment Identify teams Introduce/suggest executive partners Present/discuss customer objectives Supplier objectives Proposed objectives Business objectives Definition of responsibilities Expectations

Customer team Supplier team Executive partners (if appointed)

Kick-off Meeting

Meetings

16

a

Meeting Topics Partner meeting Meeting purpose Objectives Issues Participant responsibilities

Participants Customer teama Supplier teama

Internal Preparation Meeting

TABLE 1.1 Customer/Supplier Expanded Partnering Interface Meetings

SL3151Ch01Frame Page 16 Thursday, September 12, 2002 6:15 PM

Six Sigma and Beyond

SL3151Ch01Frame Page 17 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

17

To be successful in this exchange requires time. The reason for this is that building trust is a function of time. The longer you work with someone the more you get to know that person. To expedite the process of gaining trust, suppliers and customers may want to share in: 1. 2. 3. 4. 5. 6. 7. 8.

Non-disclosure agreements Quality improvement process Technology development roadmaps Specification development Should-cost/Total-cost model Forecasts/Frozen schedules Executive partners Job rotation with suppliers

Be aware of, adhere to, and respect the sensitive/confidential nature of proprietary information, both yours and your partner’s. Always remember: recognize the differences in company cultures. Find ways to do things without imposing your value system. Compromise... Find the common ground... Work out the differences... Move forward… Negotiate... COOPERATE! 5. Reevaluate the Partnering Process People cannot improve unless they know where they are. Evaluation of the partnering process is a way to benchmark the progress of the relationship and to set priorities for future improvement. Questionnaires with five-point rating criteria provide a means for this evaluation in which both customers and suppliers take an active role. A typical questionnaire may look like Table 1.2. Sometimes the questionnaires provide detailed definitions of certain words or criteria that are being used in the instrument. The following is a brief supplement to explain/define the rating categories and some of the terms used in Table 1.2: Ratings 1. Does not meet — Failing to satisfy requirements, unacceptable performance 2. Marginally meets — Performance is not fully acceptable, needs improvement 3. Meets — Fulfills basic requirements, satisfactory 4. Exceeds — Surpasses normal requirements 5. Superior — Consistently excels above and beyond expectations, “worldclass” performance

SL3151Ch01Frame Page 18 Thursday, September 12, 2002 6:15 PM

18

Six Sigma and Beyond

TABLE 1.2 A Typical Questionnaire Please select one of the following ratings for each question: Ratings: (1) Does not meet (2) Marginally meets (3) Meets (4) Exceeds (5) Superior 1. Rate the relationship’s impact in focusing both parties on strategic and tactical goals to foster mutual success. Strategic Tactical

1 1

2 2

3 3

4 4

5 5

Comments: 2. Have all established communication channels within Intel, from executive sponsor down, enabled the partners to improve their effectiveness/competitiveness as a company? Technical Issues Business Issues

1 1

2 2

3 3

4 4

5 5

1 1 1

2 2 2

3 3 3

4 4 4

5 5 5

1

2

3

4

5

Comments: 3. Rate the effectiveness of the team structure. Management Team Working Team Performance Reviews (Both Parties) Follow-Up on Action Items Comments: 4. Rate the effectiveness of the Key Supplier Program team in generating high quality solutions. Time of Solutions Quality of Solutions Cost-Effective Solutions Comments:

1 1 1

2 2 2

3 3 3

4 4 4

5 5 5

1 1

2 2

3 3

4 4

5 5

4 4 4 4 4 4

5 5 5 5 5 5

5. Does the Executive Partner provide meaningful support? Customer Supplier Comments: 6. Is the Key Supplier Program process formally managed in an effective manner? Customer Resource Commitment Supplier Resource Commitment Formal Communication Tools Information Sharing Total Cost Focus Dealing with “Me Best“ Comments:

1 1 1 1 1 1

2 2 2 2 2 2

3 3. 3 3 3 3

SL3151Ch01Frame Page 19 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

19

Terms Used in Specific Questions Question 1 Strategic Goals — Long-range objectives (i.e., next-generation technology) Tactical Goals — Operational, day-to-day problem solving, etc. Question 3 Management Team — Executive sponsors plus upper/middle managers Working Team — Commodity/product teams, task forces, user groups Performance Reviews — Grading joint MBOs, other indicators (e.g., quality, customer satisfaction survey) Question 4 Time of Solution — Meets or exceeds time requirements/expectations Quality of Solution — Meets or exceeds quality requirements/expectations Cost-Effective Solution — Improves total cost effectiveness/fosters mutual profitability Question 5 Meaningful Support — Active participation and involvement during and between business meetings Question 6 Resource Commitment — Adequate support (people, tools, space...) to allow successful results Formal Communication Tools — Meetings, reports, MBO’s technology exchange; correct topics, timely, worthwhile Information Sharing — Plans, technology, data; useful, timely, fosters profitability Total Cost Focus — Model in place and used to support decisions to apply resources Dealing with “The Best” — Process contributes to world-class performance Another general questionnaire evaluating the partnering process is shown in Table 1.3.

MAJOR ISSUES

WITH

SUPPLIER PARTNERING RELATIONSHIPS

In any relationship that one may think of, issues and concerns exist. Partnering is no different. Some of the areas that might be of general concern include the following: 1. 2. 3. 4. 5. 6.

Issues Issues Issues Issues Issues Other

or or or or or

concerns concerns concerns concerns concerns

within the customer’s company within the supplier’s company of a competitive nature of a political or legal nature of a technological nature

SL3151Ch01Frame Page 20 Thursday, September 12, 2002 6:15 PM

20

Six Sigma and Beyond

TABLE 1.3 A General Questionnaire Evaluate the following categories based on a rating of 1 to 5, with 1 being low and 5 being excellent. (Yet another variation of the criteria may be 1 = Much improvement needed, 5 = Little or no improvement needed.) Executive commitment to the process Recognition of mutual dependencies Mutually defined and shared expectations/objectives Executive partners/sponsors Quick issue resolution (break down roadblocks) Understanding and sharing of risks Sharing of technical roadmaps/competitive analysis/business plans Openness, honesty, respect Formal and frequent communication/feedback process Access to data Establish clear definition of responsibility (project leadership)

Issues or concerns of specific nature may develop when any of the following situations exist: 1. Support on either side is insufficient. 2. Something has caused one party to consider abandoning the partnering relationship. 3. A “better deal” or innovation threatens the partnering relationship. 4. Unequal benefits or conflicting incentives exist. 5. There are forced requirements under the guise of a partnering relationship and fear on the part of the supplier to decline or dissent, particularly if the supplier is small. 6. Key players change or there is a change of ownership.

HOW CAN WE IMPROVE? A fundamental question that needs to be answered from a customer’s perspective is “How can we improve?” The answer is by establishing a process with strategic importance of “key” relationships. Once this process is identified then it needs recognition — the more the better. How do we do that? We can do it by: 1. 2. 3. 4.

Establishing upper management involvement Sharing information: technology exchanges Showing suppliers how to use the data Educating suppliers in tools and methodologies

We can benefit from creating a “mentoring” attitude toward our suppliers. Traditionally we say, “Do this because we need it.” Start saying (and thinking), “Do

SL3151Ch01Frame Page 21 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

21

this because it will make you a stronger company, and that will in turn make us a stronger company.” Become a mentor in the Partnering for Total Quality assessment process with your suppliers. Clearly define expectations by: 1. Mutually developing short- and long-term objectives for each relationship 2. Increasing the concentration on areas for mutual success; reducing the concentration on terms and conditions 3. Making decisions based on total cost; increasing the involvement and awareness of suppliers in this process In the final analysis, in order for a successful partnership to flourish both partners — customer and supplier — must recognize that change is imminent, at least in the following areas: 1. 2. 3. 4. 5.

Organization itself Internal, interfunctional communication Customer orientation World-class definition Skills development

Are there indicators of a successful partnering process? We believe that there are. Typical indicators are the existence of: 1. 2. 3. 4. 5. 6. 7. 8. 9.

Formal communication processes Commitment to the suppliers’ success Stable relationships, not dependent on a few personalities Consistent and specific feedback on supplier performance Realistic expectations Employee accountability for ethical business conduct Meaningful information sharing Guidance to supplier in defining improvement efforts Non-adversarial negotiations and decisions based on total cost of ownership 10. Employees empowered to do the right thing

BASIC PARTNERING CHECKLIST The basic partnering principles below may be applied to any customer/supplier relationship, regardless of size of company and number of employees. The principles also apply to relationships within the organization. The investment is primarily an attitude and behavioral change to bring about six sigma quality and beyond. 1. Leadership Our management:

SL3151Ch01Frame Page 22 Thursday, September 12, 2002 6:15 PM

22

Six Sigma and Beyond

1. Is personally committed to the principles of the partnering process 2. Has directed organization-wide commitment, adoption, and execution of the partnering principles and philosophy 3. Is committed to generating accurate forecasts to improve delivery schedule stability with our suppliers 4. Ensures that the partnering principles flourish even in stressful times 5. Seeks mutually profitable arrangements with our suppliers 6. Is involved in high-level review of the partnering process. 2. Information and Analysis Our organization: 1. Has standardized measurements and performance for products, processes, service, and administration 2. Respects the protection of intellectual property 3. Treats information gained in open exchanges with respect and confidentiality 4. Provides consistent and specific feedback on supplier performance 3. Strategic Quality Planning Our organization: 1. Avoids short-term solutions at the expense of long-term viability 2. Places more emphasis on overall needs and mutual expectations, less on legal or formal aspects of the relationship 3. Uses reasonable and realistic expectations and milestones with our customers and suppliers 4. Demonstrates a commitment to continuous improvement in all facets of our business 4. Human Resource Development and Management Our organization: 1. Promotes employee accountability for ethical business conduct through performance reviews, holding supervisors accountable for promoting such practices 2. Helps employees understand their roles as customer and supplier internal and external to the organization 3. Trains employees on business practices that are ethical, open, professional, and of high integrity 4. Provides position descriptions with a clear definition of responsibility 5. Supports decision-making authority at the lowest practical level

SL3151Ch01Frame Page 23 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

23

5. Management of Process Quality Our organization: 1. Shares basic evaluation criteria with our customers and suppliers 2. Has methods for ensuring quality of components, processes, administration, service, and final product. 3. Checks periodically with our customers to verify that our quality meets their expectations 6. Quality and Operational Results Our organization: 1. Shares meaningful information and data with our customers and suppliers, with frequent and timely feedback on problems as well as successes 2. Provides guidance to suppliers in defining improvement efforts that address all problems 7. Customer Focus and Satisfaction Our organization: 1. Recognizes mutual dependencies with our customers and the need to work together; understands that partnering does not end with the signing of the purchase order. 2. Engages in win/win, non-adversarial negotiations and purchasing decisions based on total cost of ownership 3. Provides prompt disclosure to customers of any inability of the organization to meet current or future requirements; makes realistic commitments to customers

EXPANDED PARTNERING CHECKLIST In addition to the basic partnering principles, expanded partnering recognizes the need for mutual support based on such factors as cost, risk, criticalness, and actual performance. The investment involves an application of resources from both the customer and the supplier. Customer resource availability limits the number of expanded partnering relationships in which any organization can be simultaneously engaged. 1. Leadership Our senior management, in the role of an executive customer partner or executive supplier partner (champion):

SL3151Ch01Frame Page 24 Thursday, September 12, 2002 6:15 PM

24

Six Sigma and Beyond

1. Serves in a long-term assignment for each expanded partner relationship 2. Is available to support prompt issue resolution 3. Establishes strong counterpoint relationships with our key customers and suppliers 4. Provides for and supports decision-making authority at the lowest practical levels 5. Encourages and supports prompt responsiveness to communications affecting customer/supplier relationships 6. Maintains a rapid management approval cycle, providing an ombudsman when required 7. Commits adequate time to the partnering process 8. Ensures that cohesive, internal, cross-functional teams are in place to support the partnering process 2. Information and Analysis Our organization, with our suppliers: 1. Uses positive encouragement and support to improve performance and total cost of ownership 2. Participates in joint information-sharing activities to develop value analysis models 3. Shares technical roadmaps, competitive analyses, and plans 4. Focuses on clearly defined, complete, achievable requirements, with less emphasis on contractual terms and conditions 5. Ensures that suppliers understand our long-term procurement strategy 3. Strategic Quality Planning Our organization: 1. Shares short- and long-term improvement plans and priorities with suppliers and customers 2. Works with customers and suppliers to understand their quality needs and plans for continuous improvements 4. Human Resource Development and Management Our company management: 1. Has established technical advisory boards to support supplier activities 2. Communicates regularly with customer and supplier management to understand mutual needs and possible areas for cooperation 3. Encourages employees to submit suggestions for continuous quality improvements 4. Offers the same quality training to supplier personnel as we provide to our own employees

SL3151Ch01Frame Page 25 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

25

5. Management of Process Quality Our organization works with customers and suppliers to: 1. Share mutual joint performance measures that are written, measured, and tracked 2. Work toward standardization of quality and certification programs 3. Develop and implement valid quality assurance systems for products, processes, service, and administration 6. Quality and Operational Results Our organization works with customers and suppliers to: 1. Develop joint quality and yield improvement processes 2. Provide access to process data for tool and material development and refinement 7. Customer Focus and Satisfaction Our organization works with customers to: 1. Mutually define expectations, understand mutual requirements, and share risks 2. Ensure that partnering survives lapses in missed generation orders 3. Establish formal, frequent communications as part of the management process

THE ROBUST TEAM: A QUALITY ENGINEERING APPROACH In general, the traditional approach to evaluating the performance of groups in process has been twofold. The first has been to use a developmental model that provides a summary of the different phases or stages in the life cycle of a group. A popular example of this approach is the forming, storming, norming, performing model of group development. Each phase corresponds to a stage in the group life cycle — review Volume I, Part II of this series. The second model has emphasized structural patterns of a group or team. These may be construed in terms of gender, experience, length of service, or positional roles (leader, secretary, or assistant, for example). Using the structural approach, the team can also be analyzed in terms of process; the “peacemaker,” the “aggressor,” the “blocker,” or the “help-seeker,” for example, or Resource Investigator, Coordinator, and so on. Both these models have proven to be useful when trying to describe some aspects of group dynamics, and it may be possible to identify colleagues who fulfill some of these roles or identify teams that have passed through these different

SL3151Ch01Frame Page 26 Thursday, September 12, 2002 6:15 PM

26

Six Sigma and Beyond

stages of development. Unfortunately, such a restricted approach to monitoring team process does not provide any feedback as to whether the team is producing predictable results, nor does it identify problems or opportunities for improvement — especially breakthrough opportunities. Specifically, no opportunity exists to determine whether team process is “in control” (capable) or whether the group is “out of control” (chaotic and falling far short of what it could achieve). Some of these issues were addressed in Volumes I and II of this series, and perhaps the reader may want to review them at this time. A further shortcoming in non-systems approaches to team building concerns team process improvement. As long as the team is operating within “acceptable” parameters, no opportunity or drive to improve or maximize the performance of the team exists. Furthermore, the team usually does not have the ability or training to self-regulate and, through self-regulation, to begin to change and adapt to the continual change taking place in the workplace. These and other considerations suggest that a systems approach to team building may have considerable advantages. The robust team involves an examination of teams as systems in conjunction with more detailed parallels between a team systems approach and the model put forward by Taguchi as part of his quality engineering methodology (see Volume V of this series). Using this viewpoint, a system is considered as a means by which a user’s intention is transformed into a perceived result. Therefore, if teams are considered in terms of how successfully they transfer energy when they function, it should be evident that there will be parallels between their functioning and the functioning of an engineered system — as in the P diagram for example. After all, in many ways, a team shares similar features to the manufacturing process of a particular product. Specifications are drawn up (objectives, time scales, etc. are established); the production machinery is put in place (team members are selected); the production process is designed and implemented (teams meet, establish norms, set agendas, and engage in problem solving, decision making, and planning activities, etc.) and the system is regulated by performance criteria (by the individual members’ expectations, assessments, performance appraisals, etc.). In manufacturing, it is important not to separate the performance of the component from its interaction with other components and its integration into large subsystems of the whole process or product. In teams, it is important not to separate the performance of the individual from his/her relationships to other team members, their interactions, and their membership in sub-teams and the team as a whole — rather it is of paramount importance to view them as a team system.

TEAM SYSTEMS Many social psychologists only consider a collection of people to be a group if their activities relate to one another in a systematic fashion. However, it is easier to define a group as a collection of individuals. The word “team,” however, as mentioned in Volume I, Part II, is reserved for those groups that constitute a system whose parts interrelate and whose members share a common goal. Some groups can easily be viewed according to this criterion. A soccer club, its manager, and its players constitute a set of parts necessary to the functioning of the whole — the common

SL3151Ch01Frame Page 27 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

27

aim being to win soccer games. However, when does a newly established team become a good or effective team? To see the answer to this question let us examine the team from a systems approach. Input A team has an input or signal. The input is the information, energy, resources, etc., that enter into the system and are transformed through its structures and processes. A broad spectrum of inputs into the system can exist and, depending on the perspective one chooses to take, the boundaries that are drawn around the system can be more or less inclusive of these elements. A system in which the boundary is closely defined will have only the fixed structures and extant processes within it and will have a wide range of inputs, many of which may enter the system simultaneously. A system that has a very broad boundary might include people, materials, resources, and most information as a part of the system, with the input defined very narrowly as a discrete piece of information or energy. Signal The signal as developed in the Taguchi model has a more specific and limited definition. It is an input into the system, but it is limited to the means by which the user conveys to the system a deliberate intention to change (or adjust) the system output. In more general terms, it is the variable to which the system must respond in order to fulfill the user’s intent. From this perspective, most of what are traditionally considered inputs into the system, i.e., people, materials, information, and so on, are already part of the system itself, and the signal is the discrete piece of information that determines the amount of energy transformed by the system. The System The structure of a system comprises aspects of the system that are relatively static or enduring. Process, on the other hand, refers to the behavior of the system. Consequently, process refers to those relatively dynamic or transient aspects of a system that are observable by virtue of change or instability. Traditional models of a system are based upon an input-process-output model. The system acts to transform the energy from the input into the output. This process, once established, is subject to variation due to internal and external factors that produce “error states” or outputs other than the desired output. These outputs can simply be wasted energy or may actually reduce the functional ability of the system itself. If a particular team has a task to perform, e.g., solving a problem, you can consider the team to be a system that has inputs, output, and a process that allows the team members to transform their energy into the desired outputs. Team process can be defined as any activity (for example, meetings) that utilizes resources (the team) to transform inputs (ideas, skills, and qualities of team members) into outputs (discoveries, solutions to problems, proposals, actions, design ideas, products, etc.).

SL3151Ch01Frame Page 28 Thursday, September 12, 2002 6:15 PM

28

Six Sigma and Beyond

Often the energy that the team brings to the process is not used to best effect. For example, in a meeting, time may be wasted reiterating points because individuals have not paid attention to what is being discussed or because there is cross talk. This in turn leaves people annoyed and frustrated. These are examples of “error states” or undesirable outputs from the team process. Output/Response In traditional systems models, the output is whatever the system transforms, produces, or expresses into the environment as a consequence of the impact its structures and processes have on the input. An output can be anything from a newborn baby to well done barbecued ribs to a presentation to a text return. This is very important to understand because teams, by their nature, are complex and multifunctional. They cannot and should not be configured to produce one kind of response. Most teams will have a whole range of outputs with accompanying measures that will be used to identify how successful they are and how effective they are in transferring energy. The key is to identify appropriate measures that can be used to monitor the team’s progress. The Environment It is important in attempting to maximize the performance of a team to identify factors that may have an impact on the performance of the team and its ability to maximize the transfer of its input into desired output and over which the team has little or no control. (Remember, the output of the team will be a new design — however defined — and it is up to the team to make that design “wanted” in the preset environment. This is not a small feat.) These factors are designated as internal or external to the system. It is these factors that cause energy to be wasted and undesirable output (error states) to occur. External Variation In teams, external variation factors may include such things as change in team membership, the environment in which the team is working, changing demands from management, corporate cultural, racial, and gender factors, and so on. In developing a group process, it is important to develop group systems and processes that are robust to these factors. In addition, team goals exert a considerable influence on the behavior of individual members, and goals can vary enormously. They could be output targets that will vary in accordance with the team’s task — problem-solving teams puzzling over the root cause of a problem; design teams considering the optimization of a particular system design to achieve robustness; a marketing team attempting to understand the exact details of customer requirements; or sports teams, each of which will have an entirely different set of performance goals depending upon the sport: soccer, football, tennis, golf, and so on. Any analysis of working teams should take into account the objectives of the team and the situation in which the team performs because both will have a profound effect on the team functioning.

SL3151Ch01Frame Page 29 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

29

Internal Variation Internal variation, on the other hand, relates to factors that are in the team system and its members. People may bring predetermined ideas about the correct design solution. They may have biases about other team members depending on their race, gender, function, grade, and so on. Certain team members may not get along with other team members and will regularly question, challenge, or contradict the others for no apparent reason. The team may not manage its time well and consequently may find itself chronically short of time at the end of meetings. Team members may not know how to ask open questions that will open up fresh avenues of information. Closed questions will result in familiar dead ends or nonproductive and previously rejected ideas. Team members may not know how to build on the ideas of other team members and, consequently, good ideas may be regularly lost. If the reader needs help in this area, we recommend a review of Volume I, Part II. The Boundary At the simplest level, boundaries can be put around almost anything, thereby defining it as a system. In practice, the identification of the boundary is the key to successful system analysis. The classification of factors (signal, control, and variation) that impact on the system is dependent on the way in which the boundary is defined. For example, by setting the boundary of the system fairly wide, to include the team members, environment, resources, information, and so on, leaving only the directive from the champion or the monthly output target outside, more factors would be considered as control factors and fewer as variation. In this case, the directive from the champion would be the signal factor. The team members, environment (or aspect of it), and so on would be control factors. External variations would then include disruptions to the team process from sources outside the team boundary. Internal variations would include attitudinal, cultural, and intellectual variations among and between team members and variations in environmental conditions (e.g., temperature). By setting a narrower boundary, many of the factors such as environment and resources would be considered external to the system and therefore would become noise factors rather than control factors. These issues are important because they determine the team’s strategy for dealing with variations and establishing a means of becoming robust to them.

CONTROLLING

A

TEAM PROCESS: CONFORMANCE

IN

TEAMS

A tale in Hellenic mythology describes the behavior of Procrustes — an innkeeper by the Corinthean peninsula. Procrustes took his clientele, people of definite natural shape and size, and either stretched or truncated their limbs so that they might fit the mattresses he provided. There are many echoes here of the original approach to quality, “We know what you want, we will design it, you will buy it and you will like it.” Or the now famous quality euphemism, “We are not sure of what is really quality, but we sure know it, if we ever see it.” Fortunately, this philosophy is being transformed into a “customer-driven approach” and the pursuit of Total Quality Excellence through DFSS. It is not entirely

SL3151Ch01Frame Page 30 Thursday, September 12, 2002 6:15 PM

30

Six Sigma and Beyond

unreasonable, therefore, when it comes to monitoring groups or teams, to identify an alternative to the current emphasis on fitting the behavior of team members into behavioral roles through a “Procrustean” method, that is, by squeezing identity and function into personality models like those of Belbin, Myers-Briggs, Bion, and so on, through normalization and pressure to conform. Remember, one of the diversity issues is the fact that everyone is different and we are all much better because of that difference. This is particularly the case when old norms are not questioned and challenged regularly or when personality models are used to avoid genuine personal contact or in place of a genuine understanding of the uniqueness of others.

STRATEGIES

FOR

DEALING

WITH

VARIATION

There are four basic strategies for dealing with variation and its effect on the performance of a system: ignore the variation, attempt to control or eliminate the variation, compensate for the variation, or minimize the effect of the variation by making the system robust to it. Adopting the first of these strategies would mean accepting that teams will never function efficiently, but hoping that they will “do the best they can under the circumstances.” As with an engineering system, this strategy would result in a lot of unhappy customers. Generally, with engineering systems, you are encouraged to adopt strategy four first, reverting only to strategies two and three as a last resort because they are difficult and expensive to implement. While strategy four should also be chosen in the case of the team system wherever possible, you have greater flexibility in many cases to consider the other two options. Controlling or Eliminating Variation Procrustes’ behavior is an example of controlling inner variation. While this approach to variation might be considered extreme, you may have some scope for selecting team members with the right characteristics for effective teamwork as well as the necessary technical expertise. External variations are perhaps a little easier to deal with. For example, you could ensure that meetings are held away from the shop floor to reduce distractions due to noise (in the audible sense!) or hold them at an off-site location to minimize interruption. Compensating for Variation The principal means of compensating for variation is by providing some feedback on its effect on system output. The link between structure and process —the way in which structure determines process, and for your purposes perhaps more importantly, the way that process determines structure — is found in the concept of feedback loops. Feedback loops are so named because they are circular interrelationships that feed information from output back to input. Information is transmitted within the system and is used to maintain stability, to bring about structural changes, and to facilitate interaction with other systems.

SL3151Ch01Frame Page 31 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

31

Even the simplest model of the effective team includes this concept of feedback loops. By employing information feedback loops, systems may behave in ways that can be described as “goal seeking” or “purposive.” Negative feedback allows a system to maintain stability as in the case of the most commonly quoted example, a thermostat. A thermostat is controlled by negative feedback so that when the temperature increases above a certain level the heating is switched off, but when the temperature decreases sufficiently the heating is switched on. The process of maintaining stability is called “homeostasis.” The capacity for such control is engineered into some mechanical systems and occurs naturally in all biological and social systems. Threats to the stability of the system will be countered in a powerful attempt to maintain homeostasis. System Feedback One alternative approach is to monitor those aspects of team behavior that are observable (i.e., gather “the voice of the process”). Descriptive Feedback offers a non-judgmental method of monitoring what happens in working groups. It allows team members to notice when team process is in control and meeting or exceeding predetermined expectations or drifting out of control and reducing potential. Descriptive Feedback provides three basic functions: 1. It makes explicit what is happening during team process. 2. It describes those characteristics of team process behavior, relationships, and feelings that may degrade or go out of control and inhibit the potential of the team. 3. It determines what, if anything, needs to be changed in order to facilitate continuous improvement in team process. Feedback over time enables a team to establish performance-based control limits. By using these data, specific characteristics or variables relating to team process can be plotted over time. This will identify patterns that emerge and that can be used to identify and capture the degree of variability of the team. Some patterns are related to “in control” conditions, others to “out of control” conditions, just as the patterns of points on a control chart can be used to establish whether a manufacturing process is in control or out of control. Based on feedback that describes what people notice and how they feel, the team is able to regulate its process and identify opportunities for improvement. Minimizing the Effect of Variation The Parameter Design approach used in quality engineering — see Volume V of this series — is concerned with minimizing the effect of variation factors by making the system robust. This involves identifying control factors — in this case, aspects of the team process that are within the control of the team and that can be used to reduce the impact of variation factors without eliminating or controlling the variation factors themselves. An example of a “control factor” functioning in this way is the

SL3151Ch01Frame Page 32 Thursday, September 12, 2002 6:15 PM

32

Six Sigma and Beyond

use of Warm-Up and its consideration of “place” (layout, heating, lighting, ventilation) so that best use is made of the facility provided and distractions are minimized, even though the place itself and many of its features cannot be changed. The key to a successful team lies not only in identifying those parameters that are critical for the efficient transformation of inputs to the team process into outputs but also in doing this with minimal loss of energy in error states and maximum robustness to variation factors in the environment. Different types of teams with different outputs required of them would have different parameters established for their most efficient performance. Many of the structures, processes and skills that could be used as control factors in a team process have been identified in Volume I, Part II of this series. Through this process of observation, it is possible to establish control limits in a wider area of team performance. A number of the factors that have an impact on team performance can be observed and regulated through feedback, and “tolerance” for them can be established depending upon the makeup and objective of the team. These factors include warming up and down, place, task, maintenance, process management, team roles, agenda management, communication skills, speaking guidelines, meeting management, exploratory thinking guidelines, experimental thinking guidelines, change management, action planning, and team parameters. The traditional approach to engineering waits until the end of the design process to address the optimization of a system’s performance — in other words, after parameter values are selected and tolerances determined, often at the extremes of conditions and often without considering interactions among different components or subsystems. When the components and subsystems are integrated, and if performance does not meet the target value or the customer’s requirements, parameter values are altered. Consequently, though the system may be adjusted to operate within tolerance, this process does not guarantee that the system is producing its ideal performance. Similarly, traditional approaches to building teams have selected team members according to a number of factors: predetermined skills and knowledge, established roles for team members, and implemented structured norms. They also have waited until the end of the process of team design in order to optimize performance. If the team does not perform within the accepted values of these parameters, then it is adjusted: team members are changed, roles are redefined, norms are more strictly enforced. This, however, is against performance criteria that do not necessarily optimize the team’s performance nor add to the motivation or job satisfaction of the team members. The shift suggested in Parameter Design in engineering (and that may be applied to teams as well) is to move from establishing parameter values to identifying those parameters that are most important for the function of the process and then determine through experimental design the correct values for those parameters. The key is to establish the values that use the energy of the system most efficiently and that are resistant to uncontrollable impact from other factors internal or external to the system itself.

SL3151Ch01Frame Page 33 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

33

MONITORING TEAM PERFORMANCE One way of monitoring team performance has already been suggested, namely the use of Descriptive Feedback. Gathering “the voice of the process” enables the team to evaluate its performance and to continuously improve its efficiency and hence its effectiveness, before completing the task. Preliminary work in using process-control charting from Statistical Process Control suggests that there is opportunity for application to group process. This provides a second means of monitoring and continuously improving the team’s performance. Critical “control factors,” identified using the Parameter Design approach, could be measured and monitored in this way. Based upon further refinement, it may be possible to establish control limits, targets, and tolerances for these factors. System Interrelationships A systems model of processes differs from traditional models in many ways, one of which is the notion of circular causality. In the non-systems view, every event has its cause or causes in preceding events and its effects on subsequent events: the scientist seeks the cause or effect. Using the linear method of causality, ultimate causes are sought by tracing back through proximate causes. However, many phenomena do not “fit” the linear model: the relationships between them — and the relationships between the attributes or characteristics of the elements — do not conform to this linear approach to causality. In engineering systems, a direct cause and effect relationship often exists between the component of the system and the transformation of the input into an output. A steering wheel channels the input of the vehicle operator directly into the output of the system. That is, turning the steering wheel to the right or left actually turns the wheels of the vehicle to the right or the left. However, it is equally clear that error states or phenomena are nowhere near as simple or linear in the causal relationship. Feedback loops and circular causality create very complex interactions. Similarly, the choice of lubricants may not affect the performance of the system until months or years later, when early deterioration of a transmission would result in difficulty shifting gears. Similarly, in teams, some cause and effect relationships are clearly related in time and others are not. Interventions by a timekeeper will affect the ability of the team to stick to its agenda. But other factors have more circular relationships. In a global problem-solving team, changing seating arrangements from the long-tabled boardroom style to a circular arrangement will result in more universal eye contact among team members, which may increase the team’s communication. This leads to enhanced exchange of information, which may lead to a clearer identification of the problem which will, in turn, lead to a more targeted search for relevant data, which will finally lead to a root-cause identification for the problem. Changing the seating arrangement may enhance finding a root cause more quickly than might have been the case in boardroom seating, and the cause and effect chain may be quite intricate.

SL3151Ch01Frame Page 34 Thursday, September 12, 2002 6:15 PM

34

Six Sigma and Beyond

SYSTEMS ENGINEERING An emerging basis for unifying and relating the complexities of managerial problems is the system concept and its methodology. This concept has been applied more to the analysis of productive systems than to other fields, but it is clear that the value of the concept in management is pervasive. The word “system” has become so commonplace in the general literature as well as in the field that one often wants to scream, for its common use almost depreciates its value. Yet the word itself is so descriptive of the general interacting nature of the myriad of elements that enter managerial problems that we can no longer talk of complex problems without using the term “systems.” Indeed, we must learn to distinguish the general use of the term from its specific use as a mode of structuring and analyzing problems. One of the great values of the system concept is that it helps us to take a very complex situation and lend order and structure to it by using statistics, probability, and mathematical modeling. A major contribution of the concept is the reduction of complexity in managerial problems to a block diagram showing the relationship and interacting effects of the various elements that affect the problem at hand. At its present state of development and application, the systems concept is most useful in helping us gain insight into problems. At a second and very powerful level of contribution, however, systems analysis is gaining prominence as a basis for generating solutions to problems and evaluating their effects, and for designing alternate systems.

“SYSTEMS“ DEFINED We have been using the term systems without defining it. Though nearly everyone may have a general understanding of the term, it may be useful to be more precise. Webster defines a system as a regularly interacting or interdependent group of items forming a unified whole. Thus a system may have many components and objects, but they are united in the pursuit of some common goal. They are in some sense unified, organized, or coordinated. The components of a system contribute to the production of a set of outputs from given inputs that may or may not be optimal or best with respect to some appropriate measure of effectiveness. Systems are often complex, although the definition does not specify that they need to be. It is probably correct to say that some of the most interesting systems for study are complex and that a change in one variable within the system will affect many other variables of the system. Thus in productive systems, a change in production rate may affect inventories, hours worked per week, overtime hours, facility layout, and so on. Understanding and predicting these complex interactions among variables is one of our main objectives in this section. One of the elusive aspects of the systems concept is in the definition of a specific system. The fact that we can define the system that we wish to consider and draw boundaries around it is important. We can then look inside the defined system to see what happens, but it is just as important to see how the system is affected by its environment.

SL3151Ch01Frame Page 35 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

35

Thus, invariably, every system can be thought of as a part of an even larger system. One of the dangers of defining systems that are too narrow in scope is that we may fail to see broader implications. On the other hand, a broad definition runs the risk of leaving out important details involved in the functioning of the system. Obviously, there is a large element of “art” in the application of systems concepts. Systems can be open or closed. An open system is one characterized by outputs that respond to inputs but where the outputs are isolated from and have no influence on the inputs. An open system is not aware of its own performance. In an open system, past performance does not control future performance. A closed system (sometimes called a feedback system), on the other hand, is influenced by its own behavior. A feedback system has a closed loop structure that brings results from past action of the system back to control future action. There are two types of feedback systems: the negative feedback, which seeks a goal and responds as a consequence of failure to achieve the goal, and the positive feedback, which generates growth processes wherein action builds a result that generates still greater action. Unfortunately most of the feedback systems in managerial problems are of the negative feedback type where the objective is to control a process.

IMPLICATIONS

OF THE

SYSTEMS CONCEPT

FOR THE

MANAGER

Managers who put the systems concept to work are rewarded initially by the development of a deeper understanding of the systems that they manage. By developing the structure of the interacting effects of system components and the various feedback control loops in the system, managers can see better which “handles” to twist in order to keep themselves in control. Indeed, with a knowledge of the system structure, a manager can see how it might be possible to restructure the system in order to create the most effective feedback control mechanisms. With the availability of large-scale system models (simulation, statistical, reliability, and mathematical models) a manager is better able to assess the effects of changes in one division component on another and on the organization as a whole. Furthermore, the managers of any of the productive operations are better able to see how their units fit into the whole and to understand the kinds of trade-offs that are often made by higher level management and that sometimes seemingly affect one unit adversely. Perhaps one of the most important contributions of systems thinking is in the concept of suboptimization. Suboptimization often occurs when one views a problem narrowly. For example, one can construct mathematical formulas to determine the minimum cost (optimum) quantity of products or parts to manufacture at one time, which results in a supposedly optimum inventory level. If one broadens the definition of the system under study, however, and includes not just the inventory and reorder subsystem but the production and warehousing subsystems as well, it may turn out that the inventory-connected costs are a measure of only part of the problem. If the product exhibits seasonal sales, the costs of changing production levels may be significant enough to warrant carrying extra inventories to smooth production and employment. In such a situation, the minimum cost inventory model would be a suboptimal policy.

SL3151Ch01Frame Page 36 Thursday, September 12, 2002 6:15 PM

36

Six Sigma and Beyond

Organizational suboptimization often occurs when the production and distribution functions of an enterprise are operated as essentially two different businesses. The factory manager will be faced with minimizing production cost while the sales/distribution manager will be faced mainly with an inventory management, shipping, and customer service problem. Each suborganization attempting to optimize separately will likely result in a combined cost somewhat larger than if the attempt were made to optimize the combined system. The reasons are fairly obvious, since in minimizing the costs of inventories, the sales function transmits directly to the factory most of the effects of sales fluctuations instead of absorbing these fluctuations through buffer inventories. Suboptimization is the result. By coordinating the efforts of the production and distribution managers, however, it may be possible to achieve some balance between inventory costs and the costs of production fluctuation. Another way to view suboptimization is through the “hidden factory” — the terminology of six sigma. If we take for example the issue of safety, let us examine what is really at stake. No one will deny that the bottom line of all safety programs is injury prevention, more often called “loss control.” To appreciate the concept of “loss control,” however, we must look at the direct and indirect costs (often called the hidden costs) associated with an on-the-job injury. The direct costs are: 1. Medical 2. Compensation The indirect costs are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Time lost from work by injured Loss in earning power Economic loss to injured’s family Lost time by fellow workers Loss of efficiency due to breakup of crew Lost time by supervision Cost of breaking in new worker Damage to tools and equipment Time damaged equipment is out of service Spoiled work Loss of production Spoilage — fire, water, chemical, explosives, and so on Failure to fill orders Overhead cost (while work was disrupted) Miscellaneous (There are at least 100 other items of cost that appear one or more times with every incident in which a worker is injured.)

The point here is that with most injuries the focus becomes the direct cost, thereby dismissing the indirect costs. It has been estimated time and again that the cost relationship of direct to indirect cost is one to three, yet we continue to ignore the real problems of injury. An appropriate system design for injury prevention would

SL3151Ch01Frame Page 37 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

37

minimize if not eliminate the hidden costs. Generally speaking, the system should include (a) engineering, (b) education, and (c) enforcement considerations. Some specific considerations should be: 1. 2. 3. 4.

Workers will not be injured or killed Property and materials will not be destroyed Production will flow more smoothly You will have more time for the other management duties of your job

DEFINING SYSTEMS ENGINEERING A simple definition of systems engineering is: A customer/requirements–driven engineering and management process which transforms the voice of the customer(s) into a feasible and verified product/process of appropriate configuration, capability, and cost/price. A system is always greater than the sum of its parts and is no better than the weakest link. The derivative of that, of course, is that optimizing a part does not optimize the whole. This was brought out by Mayne et al. (2001), when they reported that 37% of all the automotive warranty for model year 2000 was in interfacing of parts rather than individual components. The message of Mayne and coworkers and most of us in the quality field has been and continues to be: interactions determine the performance of the system. We cannot, no matter how hard we try, fully understand the whole by breaking down and analyzing parts — yet design is historically done precisely that way. Systems engineering builds on the fact that the whole is the most important entity and that integration to meet cost, schedule, and technical performance is dependent on both technical and management intervention. Ultimately, systems engineering is a team-based activity. This is very important because as we move into the future we see that: 1. Quality is becoming more customer dependent rather than definitional from the provider’s point of view. In other words, we must specify what the product or service must do and how well it must do it, then verify the design to those requirements. 2. Products/services are becoming more sophisticated (complex). 3. Traditionally, product development has been very serial with designs thrown over the imaginary wall to manufacturing — something that today is not working very well. This has resulted in late changes and ultimately higher costs. Systems engineering is based on the notion that design may be on a parallel development process and with a strong consideration for its total life cycle — manufacture, delivery, maintenance, decommission, and recycling. For systems engineering to be effective in any organization, that organization must be committed to integration of several items including timing of development and specific delivery(ies) at specific milestones. A generic approach to facilitate this is the following model, involving the steps of pre-feasibility analysis, requirement analysis, design synthesis, and verification.

SL3151Ch01Frame Page 38 Thursday, September 12, 2002 6:15 PM

38

Six Sigma and Beyond

PRE-FEASIBILITY ANALYSIS Before the actual analysis takes place there is a preliminary trade-off analysis as to what the customer needs and wants and what the organization is willing to provide. This is done under the rubric of preliminary feasibility. When the feasibility is complete, then the actual requirement analysis takes place.

REQUIREMENT ANALYSIS The requirement analysis involves the following steps: 1. Collect the requirements — the customer’s needs, wants, and expectations are collected at every level. 2. Organize requirements — group the information in such a way that requirements are easy to address. Determine if the requirements are complete. 3. Translate into more precise terms — cascade the definitions to precise terms, honing their definition to the best possible correlation of real world usage. 4. Develop verification requirements — preliminary verification tests are discussed and proposed here to make sure that they are in fact doable. At the end of the requirement analysis the results are moved to the second stage of the system engineering model, which is design synthesis. However, before the synthesis actually takes place, another feasibility analysis is completed to find out whether the organization is capable of designing the requirements of the customer. This feasibility analysis takes into consideration the organization’s knowledge from previous or similar designs and incorporates it into the new. The idea of this feasibility analysis is to make sure the designers optimize reusability and carry over parts and/or complete designs.

DESIGN SYNTHESIS Design synthesis involves the following steps: 1. Generate alternatives — the more alternatives the better. The alternatives are generated with functionality in mind from the customer’s perspective as reflected in the system specifications. Remember that the ultimate design is indeed a trade-off design. 2. Evaluate alternatives — the generated alternatives are evaluated with appropriate benchmarking data and integrated into the design based on the customer’s requirements. 3. Generate sub element requirements — big chunks or sub elements are chosen and requirements cascaded to each sub element. As the cascading process continues, verification requirements are also developed to test the overall system integrity as more and more sub elements are integrated into the total system.

SL3151Ch01Frame Page 39 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

39

At the end of the design synthesis a very important analysis takes place. This analysis tests the integrity of the design against the customer’s requirements. If it is found that the requirements are not addressed (design gap), a redesign or a review takes place and a fix is issued. If, on the other hand, everything is as planned, the process moves to the third link of the model — verification.

VERIFICATION The final stage is verification. It involves the following: 1. Verify that requirements are complete — a review of all requirements from both design and the customer takes place with appropriate tests and correlated to real world usage. 2. Verify that design meets customer’s requirements — CAE tools, labs, rigs, simulations, and key life testing are some of the verification methodologies used at this stage. The intent here is to verify that the selected system and cascaded requirements will meet the customer’s requirements and provide a balanced optimum design from the customer’s perspective. At the end of this stage, if problems are found they (the problems) revert back to the design; if there are no problems, the design goes to manufacturing, with a design ready to fulfill the customer’s expectations. This final stage in essence tests the integrity of the design against the actual hardware. In other words, the questions often heard in verification are: Does the design work? Can you prove it? The beauty of this model is that it is an iterative model, meaning that the process — no matter where you are in the model — iterates until a balanced optimum design is achieved. This is because the goal is to design a customer-friendly design with compatibility, carryover, reusability, and low complexity requirements compared to other, similar designs. Iterations happen because of human oversights, poorly defined requirements, or an increase in knowledge. Another special characteristic of systems engineering is the notion of traceability. Traceability is reverse cascading and is used throughout the design process to make sure that the voices of the customer, regulator, and corporate or lower-level design are heard and accounted for in the overall design. With traceability, extra caution is given to the trade-off analysis. This is because by definition trade-off analysis accounts for designs with certain priority levels among the needs and wants of the customer. In a trade-off analysis, we choose among stated design alternatives. However, a trade-off analysis is also an iterative process, and usually none of the alternatives is perfect [R(t) = 1 – F(t)]. This is important to remember because all trade-off analyses assess risk, both external and internal, of the given alternatives so as to make robust designs. A final word about verification and systems engineering: As we already mentioned, the intent of verification is to make sure that the hardware meets the requirements of the design. The process for conducting this verification is done — generally — in five steps, which are:

SL3151Ch01Frame Page 40 Thursday, September 12, 2002 6:15 PM

40

Six Sigma and Beyond

1. Plan — Review all requirements and make a preliminary assessment as to their impact. At the end of this evaluation, take ownership of important requirements and begin the assessment of specific tests and methods. It is not unusual at this stage to review the plan again and perhaps combine, consolidate, or even adjust the plan completely. In this stage we select attribute data, as well, monitor the “unselected” requirement, schedule preliminary tests, and approve the testing schedule. As you begin to zero in on specific targets, you may want to take into consideration features of the proposed design and benchmarking data so that your targets become of value to the customer. If this plan is rich in information, it is possible to begin predicting and formulating prototype(s). 2. Execute — In this stage, the engineer in charge will determine which test(s) to run, when to run them, what the data should look like, and what to expect. Proper test execution is of importance here. 3. Analyze/revise — Analyze the test results, and see if the design has changed in any way. Determine whether to redo the test if the design changed during the test for any reason. At this stage you expect no design changes, only testing revisions and modifications. 4. Sign-off — This is the most common ending for a verification process. In this stage final approvals are given, usually several months before production begins. 5. Archive — This is a step that most engineers do not do, yet it is a very important step in the process. The idea of archiving or documenting is to make sure that key events are appropriately documented for future use. You may want to document unusual tests, time frames of specific tests, or any specific requirements that you had the intention of verifying but could not verify using the planned method. In essence, this phase of verification consists of lessons learned that need to be carried forward to the next design. The focus of this process is to make sure that the requirements are driving the process and not the tests regardless of how sophisticated they are. To be sure, tests are an integral part of verification, but they are the means not the end. The intent of the tests is to verify each requirement, and there is no wrong way as long as the testing method is linked to real world usage. The reason for doing all this is to: 1. 2. 3. 4.

Reduce workload in design verification Improve prototype and testing efficiency by avoiding duplication Improve testing quality resulting in higher sign-off confidence Improve communication and stronger relationships between customer and suppliers

ADVANCED QUALITY PLANNING Before we address the “why” of planning, we assume that things do go wrong. But why do they go wrong? Obviously, many specific answers address this question. Often the answer falls into one of these four categories:

SL3151Ch01Frame Page 41 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

41

1. We never have enough time, so things are omitted. 2. We have done this, this way, in order to minimize the effort. 3. We assume that we know what has been requested, so we do not listen carefully. 4. We assume that because we finish a project, improvement will indeed follow, so we bypass the improvement steps. In essence then, the customer appears satisfied, but a product, service, or process is not improved at all. This is precisely why it is imperative for organizations to look at quality planning as a totally integrated activity that involves the entire organization. The organization must expect changes in its operations by employing cross-functional and multidisciplinary teams to exceed customer desires — not just meet requirements. A quality plan includes, but is not limited to: • • • • • • •

A team to manage the plan Timing to monitor progress Procedures to define operating policies Standards to clarify requirements Controls to stay on course Data and feedback to verify and to provide direction An action plan to initiate change

Advanced quality planning (AQP), then, is a methodology that yields a quality plan for the creation of a process, product, or service consistent with customer requirements. It allows for maximum quality in the workplace by planning and documenting the process of improvement. AQP is the essential discipline that offers both the customer and the supplier a systematic approach to quality planning, to defect prevention, and to continual improvement. Some specific uses are: 1. In the auto industry, demand is so high that Chrysler, Ford, and General Motors have developed a standardized approach to AQP. That standardized approach is a requirement for the QS-9000 and/or ISO/TS19469 certification. In addition, each company has its own way of measuring success in the implementation and reporting phase of AQP tasks. 2. Auto suppliers are expected to demonstrate the ability to participate in early design activities from concept through prototype and on to production. 3. Quality planning is initiated as early as possible, well before print release. 4. Planning for quality is needed particularly when a company’s management establishes a policy of “prevention” as opposed to “detection.” 5. When you use AQP, you provide for the organization and resources needed to accomplish the quality improvement task. 6. Early planning prevents waste (scrap, rework, and repair), identifies required engineering changes, improves timing for new product introduction, and lowers costs.

SL3151Ch01Frame Page 42 Thursday, September 12, 2002 6:15 PM

42

Six Sigma and Beyond

7. AQP is used to facilitate communication with all individuals involved in a program and to ensure that all required steps are completed on time at acceptable cost and quality levels. 8. AQP is used to provide a structured tool for management that enforces the inclusion of quality principles in program planning.

WHEN DO WE USE AQP? We use AQP when we need to meet or exceed expectations in the following situations: 1. 2. 3. 4. 5.

During the development of new processes and products Prior to changes in processes and products When reacting to processes or products with reported quality concerns Before tooling is transferred to new producers or new plants Prior to process or product changes affecting product safety or compliance to regulations

The supplier — as in the case of certification programs such as ISO 9000, QS9000, ISO/TS19469, and so on — is to maintain evidence of the use of defect prevention techniques prior to production launch. The defect prevention methods used are to be implemented as soon as possible in the new product development cycle. It follows then, that the basic requirements for appropriate and complete AQP are: 1. Team approach 2. Systematic development of products/services and processes 3. Reduction in variation (this must be done, even before the customer requests improvement of any kind) 4. Development of a control plan As AQP is continuously used in a given organization, the obvious need for its implementation becomes stronger and stronger. That need may be demonstrated through: 1. Minimizing the present level of problems and errors 2. Yielding a methodology that integrates customer and supplier development activities as well as concerns 3. Exceeding present reliability/durability levels to surpass the competition’s and customer’s expectations 4. Reinforcing the integration of quality tools with the latest management techniques for total improvement 5. Exceeding the limits set for cycle time and delivery time 6. Developing new and improving existing methods of communicating the results of quality processes for a positive impact throughout the organization

SL3151Ch01Frame Page 43 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

WHAT IS

THE

DIFFERENCE

BETWEEN

AQP

AND

43

APQP?

AQP is the generic methodology for all quality planning activities in all industries. APQP is AQP; however, it emphasizes the product orientation of quality. APQP is used specifically in the automotive industry. In this book, both terms are used interchangeably.

HOW DO WE MAKE AQP WORK? There are no guarantees for making AQP work. However, three basic characteristics are essential and must be adhered to for AQP to work. They are: 1. Activities must be measured based on who, what, where, and when. 2. Activities must be tracked based on shared information (how and why), as well as work schedules and objectives. 3. Activities must be focused on the goal of quality-cost-delivery, using information and consensus to improve quality. As long as our focus is on the triad of quality-cost-delivery, AQP can produce positive results. After all, we all need to reduce cost while we increase quality and reduce lead time. That is the focus of an AQP program, and the more we understand it, the more likely we are to have a workable plan.

ARE THERE PITFALLS

IN

PLANNING?

Just like everything else, planning has pitfalls. However, if one considers the alternatives, there is no doubt that planning will win out by far. To be sure, perhaps one of the greatest pitfalls in planning is the lack of support by management and a hostile climate for its practice. So, the question is not really whether any pitfalls exist, but why such support is quite often withheld and why such climates arise in organizations that claim to be “quality oriented.” Some specific pitfalls in any planning environment may have to do with commitment, time allocation, objective interpretations, tendency toward conservatism, and an obsession with control. All these elements breed a climate of conformity and inflexibility that favors incremental changes for the short term but ignores the potential of large changes in the long run. Of these, the most misunderstood element is commitment. The assumption is that with the support of management, all will be well. This assumption is based in the axiom of F. Taylor at the turn of the 20th century, which is “there is one best way.” Planning is assumed to generate the one best way not only to formulate, but to implement, a particular idea, product, and so on. Sometimes, this notion is not correct. In today’s “agile world,” we must be prepared to evaluate several alternatives of equal value. (See the section on system engineering). As a consequence, the issue is not simply whether management is committed to planning. It is also, as Mintzberg (1994) has observed, (1) whether planning is committed to management, (2) whether commitment to planning engenders commitment

SL3151Ch01Frame Page 44 Thursday, September 12, 2002 6:15 PM

44

Six Sigma and Beyond

to the process of strategy making, to the strategies that result from that process, and ultimately to the taking of effective actions by the organization, and (3) whether the very nature of planning actually fosters managerial commitment to itself. Another pitfall, of equal importance, is the cultural attitude of “fighting fires.” In most organizations, we reward problem solvers rather than planners. As a consequence, in most organizations the emphasis is on low-risk “fire fighting,” when in fact it should be on planning a course of action that will be realistic, productive, and effective. Planning may be tedious in the early stages of conceptual design, but it is certainly less expensive and much more effective than corrective action in the implementation stage of any product or service development.

DO WE REALLY NEED ANOTHER QUALITATIVE TOOL TO GAUGE QUALITY? While quantitative methods are excellent ways to address the “who,” “what,” “when,” and “where,” qualitative study focuses on the “why.” It is in this “why” that the focus of advanced quality planning contributes the most results, especially in the exploratory feasibility phase of our projects. So, the answer to the question is a categorical “yes” because the aim of qualitative study is to understand rather than to measure. It is used to increase knowledge, clarify issues, define problems, formulate hypotheses, and generate ideas. Using qualitative methodology in advanced quality planning endeavors will indeed lead to a more holistic, empathetic customer portrait than can be achieved through quantitative study, which in turn can lead to enlightened engineering and production decisions as well as advertising campaigns.

HOW DO WE USE

THE

QUALITATIVE METHODOLOGY

IN AN

AQP SETTING?

Since this volume focuses on the applicability of tools rather than on the details of the tools, the methodology is summarized in seven steps: 1. Begin with the end in mind. This may be obvious; however, it is how most goals are achieved. This is the stage where the experimenter determines how the study results will be implemented. What courses of action can the customer take and how will they be influenced by the study results? Clearly understanding the goal defines the study problem and report structure. To ensure implementation, determine what the report should look like and what it should contain. 2. Determine what is important. All resources are limited and therefore we cannot do everything. However, we can do the most important things. We must learn to use the Pareto principle (the vital few as opposed to the trivial many). To identify what is important, we have many methods, including asking about advantages and disadvantages, benefits desired, likes and dislikes, importance ratings, preference regression, key driver analysis, conjoint and discrete choice analysis, force field analysis, value analysis, and many others. The focus of these approaches is to improve performance in areas in which a competitor is ahead or in areas where

SL3151Ch01Frame Page 45 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

3.

4.

5.

6.

7.

45

your organization is determined to hold the lead in a particular product or service. Use segmentation strategies. Not everyone wants the same thing. Learn to segment markets for specific products or services that deliver value to your customer. By segmenting based on wants, engineering and product development can develop action oriented recommendations for specific markets and therefore contribute to customer satisfaction. Use action standards. To be successful, standards must be used, but with diagnostics. Standards must be defined at the outset. They are always considered as the minimum requirements. Then when the results come in, there will be an identified action to be taken, even if it is to do nothing. List the possible results and the corresponding actions that could be taken for each. Diagnostics, on the other hand, provide the “what if” questions that one considers in pursuing the standards. Usually, they provide alternatives through a set of questions specific to the standard. If you cannot list actions, you have not designed an actionable study. Better design it again. Develop optimals. Everyone wants to be the best. The problem with this statement is that there is only room for one best. All other choices are second best. When an organization focuses on being the best in everything, that organization is asking for failure. No one can be the best in everything and sustain it. What we can do is focus on the optimal combination of choices. By doing so, we usually have a usable recommendation based on a course of action that is reasonable and within the constraints of the organization. Give grasp-at-a-glance results. The focus of any study is to turn people into numbers (wants into requirements), numbers into a story (requirements into specifications), and that story into action (specifications into products or services). But the story must be easy to understand. The results must be clear and well-organized so that they and their implications can be grasped at a glance. Recommend clearly. Once you have a basis for an action, recommend that action clearly. You do not want a doctor to order tests and then hand you the laboratory report. You want to be told what is wrong and how to fix it. From an advanced quality planning perspective, we want the same. That is, we want to know where the bottlenecks are, what kind of problems we will encounter, and how we will overcome them for a successful delivery.

APQP INITIATIVE

AND

RELATIONSHIP

TO

DFSS

The APQP initiative in any organization is important in that it demonstrates our continuing effort to achieve the goal of becoming a quality leader in the given industry. Inherent in the structure of APQP are the following underlying value-added goals:

SL3151Ch01Frame Page 46 Thursday, September 12, 2002 6:15 PM

46

Six Sigma and Beyond

1. Reinforces the company’s focus on continuous improvement in quality, cost, and delivery 2. Provides the ability to look at an entirely new program as a single unit • Preparing for every step in the creation • Identifying where the greatest amount of effort must be centered • Creating a new product with efficiency and quality 3. Provides a better method for balancing the targets for quality, cost, and timing 4. Allows for deployment of the targets using detailed practical deliverables with specific timing schedule requirements 5. Provides a tool for program management to follow up all program planning processes. The APQP initiative explicitly focuses on basic engineering activities to avoid concerns rather than focusing on the results in the product throughout all phases. Based on the fact that the deliverables are clearly defined between departments (supplier/customer relationships), program concerns and issues can be solved efficiently. The APQP initiative also is forceful in viewing the review process at the end of the cycle as unacceptable. Rather, the review must be done at the end of each planning step. This provides a critical step-by-step review of how the organizations are following best possible practices. Also, the APQP initiative has a serious impact on stabilizing the program timing and content. Stabilization results in cost improvement opportunities including reduction of special sample test trials. Understanding the program requirements for each APQP element from the beginning provides the following advantages: • • • • •

Clarifies the program content Controls the sourcing decision dates Identifies customer-related significant/critical characteristics Evaluates and avoids risks to quality, cost, and timing Clarifies for all organizations product specifications using a common control plan concept

Application of APQP in the DFSS process provides a company with the opportunity to achieve the following benefits: 1. It provides a value-added tool allowing program management to track and follow up on all the program planning processes — focusing on engineering method and quality results. 2. It provides a critical review of how each organization is following best possible practices by focusing on each planning step. 3. It identifies the complete program content upon program initiation, viewing all elements of the process as a whole (AIAG 1995; Stamatis 1998).

SL3151Ch01Frame Page 47 Thursday, September 12, 2002 6:15 PM

Prerequisites to Design for Six Sigma (DFSS)

47

Once program content has been clarified, the following information can be discerned: 1. 2. 3. 4.

Sourcing decision dates are identified. Customer-related significant/critical characteristics are specified. Quality, cost, and timing risks are evaluated and avoided. Product specifications are established for all organizations using a common control plan concept.

Using the APQP process to stabilize program timing and content, the opportunities for cost improvement are dramatically increased. When we are aware of the timing and concerns that may occur during the course of a program, it provides us the opportunity to reduce costs in the following areas: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Product changes during the program development phase Engineering tests Special samples Number of verification units to be built (prototypes, first preproduction units, and so on) Number of concerns identified and reduced Fixture and tooling modification costs Fixture and tooling trials Number of meetings for concern resolution Overtime Program development time and deliverables (an essential aspect of both APQP and DFSS)

For a very detailed discussion of APQP see Stamatis (1998).

REFERENCES Automotive Industry Action Group (AIAG), Advanced Product Quality Planning and Control Plan. Chrysler Co., Ford Motor Co., and General Motors. Distributed by AIAG, Southfield, MI, 1995. Mayne, E. et al., Quality Crunch, Ward’s AUTOWORLD, July 2001, pp. 14–18. Mintzberg, H., The Rise and Fall of Strategic Planning, New York Free Press, New York, 1994. Stamatis, D.H., Advanced Quality Planning. A Commonsense Guide to AQP and APQP, Quality Resources, New York, 1998.

SELECTED BIBLIOGRAPHY Bossert, J., Considerations for Global Supplier Quality, Quality Progress, Jan. 1998, pp. 29–34. Brown, J.O., A Practical Approach to Service: Supplier Certification, Quality Progress, Jan. 1998, pp. 35–40.

SL3151Ch01Frame Page 48 Thursday, September 12, 2002 6:15 PM

48

Six Sigma and Beyond

Forcinio, H., Supply Chain Visibility: Is It Really Possible? Managing Automation, July 2001, pp. 24–28 Gurwitz, P.M., Six Questions to Ask Your Supplier About Multivariate Analysis, Quirk’s Marketing Review, Feb. 1991, pp 8–9, 23. Mehta, P.V. and Scheffler, J.M., Getting Suppliers in on the Quality Act, Quality Progress, Jan. 1998, pp. 21–28. Schoenfeldt, T., Building Effective Supplier Relationships, Automotive Excellence, Winter 1999, pp.17–25.

SL3151Ch02Frame Page 49 Thursday, September 12, 2002 6:13 PM

2

Customer Understanding

In Volume I of this series, we made a point to discuss the difference between “customer satisfaction” and “loyalty.” We said that they are not the same and that most organizations are interested in loyalty. We are going to pursue this discussion in this chapter because, as we have been saying all along, understanding the difference between customer service and customer satisfaction can provide marketers with the competitive advantage necessary to retain existing customers and attract new ones. “Understanding” what satisfaction is and what the customer is looking for can provide the engineer with a competitive advantage to design a product and or service second to none. At first glance, service and satisfaction may appear to mean the same thing, but they do not; service is what the marketer provides and what the customer gets, and satisfaction is the customer’s evaluation of the level of service received based on preconceived assumptions and the customer’s own definition of “functionalities.” The satisfaction level is determined by comparing expected service to delivered service. Four outcomes are possible: 1. 2. 3. 4.

Delight — positive disconfirmation (a pleasant surprise) Dissatisfaction — negative disconfirmation (an unpleasant surprise) Satisfaction — positive confirmation (expected level of service) Negative confirmation, which suggests that you are neither managing expectations properly nor delivering good service

In managing service delivery, relying solely on the objective aspects of service is a mistake. Customers base future behavior on their evaluation of the experience they actually had, which is in effect their degree of satisfaction or dissatisfaction. In addition to determining that satisfaction degree, marketers should seek to learn the reasons underlying customers’ feelings (the insight) in order to tell the engineers what, how and when to make changes and maintain high satisfaction levels when they are achieved. In researching these areas, marketers should note that the why is not the what; nor is it the how. That is, what happened and how it made customers feel does not tell us why they felt as they did. And not knowing that, managing not only the service that customers experience but also their expectations becomes difficult, if not impossible. At times, service providers and customers tend to think differently. Consider this dealership example: After conducting 10 focus groups for an automotive company in a medium-size Midwestern city, the researchers discussed the findings in a review meeting with the head of marketing for the company. The researchers noted that, after having spoken with more than 100 recent customers, they had learned

49

SL3151Ch02Frame Page 50 Thursday, September 12, 2002 6:13 PM

50

Six Sigma and Beyond

that the vast majority were frustrated and unhappy about having to wait more than 15 minutes before getting attended to. The marketing executive interrupted, saying, “Those customers should consider themselves lucky; if they were in the dealerships of one of our competitors, they would have to wait 20 to 30 minutes before they were seen by the service manager.” This example includes all the information needed to explain the difference between customer service and customer satisfaction. The customers in this example defined their personal expectations about the service — their waiting time experience — and clearly, a conflict existed between their service expectation (a short wait before being seen by a service manager) and their service experience (waits of more than 15 minutes). Customers then were dissatisfied with the waiting rooms and the dealerships in general. The marketing manager’s response to customer dissatisfaction was to note that the waits could have been worse: He knew that competitors’ dealerships were worse. He also knew customer waits of more than 30 minutes were not uncommon. In light of these data, he judged the 15- to 20-minute waits in the waiting room acceptable. Herein lies the conflict between service delivery and customer satisfaction. The important concept for this marketing executive — and for all marketers — is that customers define their own satisfaction standards. The customers in this example did not go to the competitors’ dealerships; instead, they came to the marketer’s dealership with a set of their own expectations in a preconceived environment. When the marketer used his service delivery criteria to defend the waiting time, he simply missed the point. Unfortunately, this illustration is typical of how many marketers think about customer satisfaction. They tend to relate customer satisfaction directly to their own service standards and goals rather than to their customers’ expectations, whether or not those expectations are realistic. To assess satisfaction, marketers must look beyond their own assessments, tapping into the customers’ evaluations of their service experience. Consider, for example, a bank that thought it was doing a good job of measuring service satisfaction but really was too focused on service delivery. This bank had developed a policy that time spent in the lobby room should be less than 15 minutes for all customers. A customer came into the office and waited 12 minutes in the reception area for a mortgage application. Then she waited another five minutes for the loan officer to clear all the papers from his desk from the previous customer and an additional three minutes for him to get the file and all the pertinent information for the current application. As this customer was leaving, she was asked to fill out a customer satisfaction questionnaire. Under the category for reception area waiting time, she checked off that she had waited less than 15 minutes. Based on this response, the bank’s marketing director assumed the customer was satisfied, but she was not; the customer had been told that if she came in for the mortgage loan during her lunch hour, she would be taken care of right away. Instead, she waited a total of 20 minutes for her application process to begin. She did not have time to shop for the gift her son needed that night for a birthday party, and her entire schedule was in disarray. She left dissatisfied.

SL3151Ch02Frame Page 51 Thursday, September 12, 2002 6:13 PM

Customer Understanding

51

Understanding the difference between service and satisfaction is the first step in developing a successful customer satisfaction program, and all marketers must share the same understanding. Only customers can define what satisfaction means to them. Here are some practical ways to understand customers’ expectations: • Ask customers to reflect on their experiences with your services and their needs, wants, and expectations. • Talk to customers face to face through focus groups, as well as through questionnaires. A wealth of information can be collected this way. • Talk with your staff about what they hear from customers about their expectations and experience with service delivery. • Review warranty data. Remember the three words that can help you learn from your customers: What, how and why. That is, what service did you experience, how did it make you feel, and why did you feel that way? Continual probing with these three perspectives will deliver the answers you need to better manage service to generate customer satisfaction. As Harry (1997 p. 2.20) has pointed out: • • • • • •

We do not know what we do not know We cannot act on what we do not know We do not know until we search We will not search for what we do not question We do not question what we do not measure Hence, we just do not know

Therefore, part of this understanding is to identify a transfer function. That is, you need a bridge (quantitatively or qualitatively) that will define and explain the dependent variable (the customer’s needs, wishes, and excitements) with the independent variable(s) (the actual requirements that are needed from an engineering perspective to satisfy the dependent variable). The transfer function may be a linear one (the simplest form) or a polynomial one (a very complex form). Typical equations expressing transfer functions may look something like: Y = a + bx Y = f(x1, x2…xn) − sin θ  df a  r  a sin θ  sin cos + + =r θ + θ     dθ g cos θ  cos θ  cos2 θ  g cos2 θ  They can be derived from:

SL3151Ch02Frame Page 52 Thursday, September 12, 2002 6:13 PM

52

Six Sigma and Beyond

• • • • • •

Known equations that describe the function Finite element analysis and other analytical models Simulation and modeling Drawing of parts and systems Design of experiments Regression and correlation analysis

In DFSS, the transfer function is used to estimate both mean (sensitivity) and variance (variability) of Ys and ys. When we do not know what the Y is, it is acceptable to use surrogate metrics. However, it must be recognized from the beginning that not all variables should be included in a transfer function. Priorities should be set based on appropriate trade-off analysis. This is because DFSS is meant to emphasize only what is critical, and that means we must understand the mathematical concept of measurement. The focus of understanding customer satisfaction has been captured by Rechtin and Hair (1998), when they wrote that “an insight is worth a thousand market surveys.” It is that insight that DFSS is looking for before the requirements are discussed and ultimately set. This will help us in identifying what is really going on with the customer. Let us look at the function first.

THE CONCEPT OF FUNCTION In any business environment, there may be no more powerful concept than that of function. To understand why this is a potent notion, we need to consider what we mean by “function.” What is “function?” Let us start with a common definition: “The natural, proper, or characteristic action of any thing...” This is the Webster’s New Collegiate Dictionary definition, and it is quite representative of what you will find in most dictionaries. This is actually a powerful and insightful definition. Think about any product or service that you purchase. What is it about the product or service (I will use the term “product” from here on, although every issue that will be discussed will be equally valid for services) that causes you to exchange money, goods, or some other scarce resource for it? Ultimately, it is because you want the “characteristic actions” that the product provides. These “actions” may be simple or complex, utilitarian or capricious, Spartan or gilded — but in each transaction, you enter with a set of unfulfilled wants and needs that you attempt to satisfy. Moreover, if the product you purchase actually manages to fulfill the wants and needs that you perceive, you are more likely to be satisfied with your purchase than when the product fails to satisfy your desires. Within these few short sentences, we have the fundamental principles that underpin three of the most powerful tools in the modern pantheon of quality, productivity, and profitability: Quality Function Deployment, Value Analysis, and Failure Modes & Effects Analysis. To put the concept of function into action, we need to refine our definition. The expanded definition we would like you to consider is “The characteristic actions that a system, part, service, or manufacturing process generates to satisfy customers.”

SL3151Ch02Frame Page 53 Thursday, September 12, 2002 6:13 PM

Customer Understanding

53

In this expanded definition, you cannot only see the concept of function at work, but you may be able to recognize the essential abstraction of a process. In a process, some type of input is transformed into an output. As a simple equation, we might say that Input(s) + Transformation = Output(s) In the case of function, the inputs are the unfulfilled wants and needs that a customer or a prospective customer has. These can be and often are intricate; this is why the discipline of marketing is still more art than science. (We will have more to say about this issue in just a moment.) Nevertheless, there exist multiple sets of unfulfilled wants and needs that are open to the lures and attractions provided by the marketplace. In this very broad model, the transformation is provided by the producer. With one, ten, or hundreds of internal processes (within any discussion of process, there is always the “Russian doll” image: processes within processes within processes), business organizations attempt to determine the unmet wants and needs that customers have. The producer then must design and develop products and delivery processes that will provide tangible and/or intangible media of exchange that will assuage the unmet needs or need sets. Finally, the external processes that involve exchange of the producer’s goods for money or other barter provide the customer with varying degrees of satisfaction. The gratification (or lack of satisfaction) that results can then be viewed as the output of the general process. In business, the inputs are not within the control of producers. As a result, producers need powerful tools to understand, delineate, and plan for ways to meet these needs. This can be thought of as the domain of the Kano model or Quality Function Deployment. The transformational activities, however, are within the control of the producer. These “controlled” activities include planning efforts to deliver “function” at a satisfactory price; the nuances and subtleties of this activity can be strongly influenced or even controlled by the discipline of Value Analysis. In addition, fulfillment of marketplace “need sets” also implies that this fulfillment will occur without unpleasant surprises. Unwanted, incomplete, or otherwise unacceptable attempts to produce “function” often result in failure. This implies that producers have a need to systematically analyze and plan for a reduction in the propensity to deliver unpleasant surprises. This planning activity can be greatly aided by the application of Failure Modes and Effects Analysis techniques. To see how these ideas mesh, we need to consider how “function” can be comprehensively mapped. This will require several steps. To apply what will be discussed in the rest of this section, we need to emphasize the importance of choosing the proper scale for any analysis. The probability is that you will choose too broad a view or too much detail; we will try to provide guidance on this issue during our discussion of methods.

SL3151Ch02Frame Page 54 Thursday, September 12, 2002 6:13 PM

54

UNDERSTANDING CUSTOMER WANTS

Six Sigma and Beyond AND

NEEDS

The nature of customer wants and needs is complex, deceptive, and difficult to discern. Nevertheless, the prediction of future wants and needs in the marketplace is perhaps the most important precursor to financial success that exists. Knowing and doing something that is profitable are two very separate (but not completely independent) aspects of this challenge. The first task that must be undertaken is to list the customers that we are interested in. Virtually no business is universal in terms of target market. Moreover, in today’s highly differentiated world, it is likely an act of folly to suggest that any product would have universal appeal. (Even an idealized product such as a capsule that, when ingested, yields immortality would have its detractors and would be rejected by some elements of humanity.) So, we need to start by cataloging the customers that we might wish to serve. In this effort, however, we need to recognize that there is a chain of customers. This is often seen in discussions of the “value chain,” a concept explored in detail by Porter (1985). For example, Porter discusses the concept of “channel value,” wherein channels of distribution “perform additional activities that affect the buyer, as well as influence the firm’s own activities.” This means that there are several dimensions on which we will discover important customers. First, there are market segments and niches. These can be geographic, demographic, or even psychographic in nature. Second, there are many intermediary customers, who have an important influence on ultimate purchases in the marketplace. Finally, there are what might be called “overarching” customers — persons or entities that must be satisfied even in the absence of any purchasing power — to enable or permit the sale of goods and services. This is readily visible in the auto industry. From the standpoint of a major parts manufacturer, say United Technologies, Johnson Controls, Dana Corporation, or Federal-Mogul, there are legions of important customers. In the market segment category, there are the vehicle manufacturers, including GM, Ford, Toyota, and all of the others. Contained within this category of customers are many sub-customers, including purchasing agents, engineers, and quality system specialists. As far as intermediary customers are concerned, we can consider perhaps a dozen or more important players. We need to consider the transportation firms that carry the parts from the parts plant to the assembly plant. We also need to think about the people and the equipment within the assembly plant that facilitate the assembly of the part into a vehicle. (If anyone doubts this is important, they have never tried to sell a part to an assembly plant where the assembly workers truly dislike some aspect of the part.) The auto dealer is yet another step in this array of hurdles, and mechanics and service technicians constitute still one more customer who must, in some way, be reasonably satisfied if commercial success is to spring forth. In addition, the auto industry has a web of regulatory and statutory requirements that govern its operation. These include safety regulations, emission standards, fleet mileage laws, and the general requirements of contract law. Behind these government requirements are still more governmental prerequisites, including occupational safety law, environmental law as applied to manufacturing, and labor law. This means that

SL3151Ch02Frame Page 55 Thursday, September 12, 2002 6:13 PM

Customer Understanding

55

the governmental agencies and political constituencies that administer these laws can be seen as the “overarching” customers described previously. Ultimately, vehicle purchasers themselves are the critical endpoint in this chain of evaluation. And within this category of customers are the many segments and niches that car makers discuss, such as entry level, luxury, sport utility, and the many other differentiation patterns that auto marketers employ. Only when a product passes through the entire sequence will it have a reasonable chance of successfully and repeatedly generating revenue for the producer. This provides a critical insight about function. Function is only meaningful through some transactional event involving one or more customers. Only customers can judge whether a product delivers desired or unanticipated-yet-delightful function. In many cases (in fact, most), firms simply do not consider all of these customers. As a result, they are often surprised when problems arise. Moreover, they suffer financial impediments as a result — even though they may simply budget some degree of failure expectation into overhead calculations. A rational assessment of this situation means that the first requirement for understanding function is a comprehensive listing of customers. Frankly, this is very hard work, and it requires time, dedication, and effort. Regardless, understanding the customers that you wish to serve is an essential prerequisite to comprehension of function.

CREATING

A

FUNCTION DIAGRAM

If you want to understand function, the first requirement is the use of a special language. Function must be described using an active verb and a measurable noun. Fowler (1990) calls this linguistic construction a “functive” — a function described in direct terms that are, to the greatest degree possible, unambiguous. In a functive, the verb should be active and direct. How can you tell if the verb meets this test? Can you subject the action described by the verb to reasonable verification? One of the difficulties with this approach is the widespread affinity for ambiguity, the evil spawn of corporate life. To reduce ambiguity, you must avoid “nerd” verbs such as provide, be, supply, facilitate, and allow. Since most people pepper their business speech with these verbs, how can you avoid using them? If you cannot avoid “nerd verbs,” then you might try to convert the noun to a verb. Instead of “allow adjustment” think about what it is that you are adjusting. For example, you could easily restate this “nerd verb” functive with “adjust clearance.” Whenever a “nerd verb” comes up, try converting the noun that goes with the nerd verb to a verb, and then select the appropriate measurable noun. Most of the time, this will reduce the ambiguity. The measurable noun also must be reasonably precise. In particular, it should be relatively unchanging in usage and should rarely be the name of a part, operation, or activity used to generate the product or service under consideration. The test for a measurable noun is very simple: can you measure the noun? Bear in mind, however, that the measurement may be as simple as counting — or it can be a detailed statement of a technical or engineering expectation of the degree to which a function can be fulfilled. Ultimately, the combination of an active verb and measurable noun will give rise to an extent — the degree to which the functive is executed.

SL3151Ch02Frame Page 56 Thursday, September 12, 2002 6:13 PM

56

Six Sigma and Beyond

For example, let us consider a simple mechanical pencil. The mechanism of the pencil must feed lead at a controlled rate. This also means that there must be a specific position for the lead. If the lead is fed too far, the lead will break. If the feed is not far enough, the pencil may not be able to make marks. As a result, one function that we can consider is “position lead.” The measurement is the length of exposed lead, and the desired extent of the positioning function may be 5 mm from the barrel end of the pencil. If there are limits on the extent in the form of tolerance, this is a good time to think about these limits as well.* While you are describing function in terms of an active verb and measurable noun, it is very important to maintain a customer frame of reference. Do not forget that function is only meaningful in terms of customer perception. No matter how much you may be enamored of a product feature or service issue, you must decide if the target customer will perceive your product in the same way.**

THE PRODUCT FLOW DIAGRAM

AND THE

CONCEPT

OF

FUNCTIVES

Now that we understand the essential issues involved in describing function, we can learn more about techniques for understanding the many complex functions that exist in a product. If products had just one or two functions, it would be easy to understand the issues that motivate purchase behavior. In today’s complex world, though, products seem to have more features (and hence more functions) nearly every day. How can we understand this complexity? Fortunately, there are common patterns that exist in the functionality of any product. We can see this through the creation of a product flow diagram. A product flow diagram uses simple, direct language to delineate function. This is a valuable aid to help you understand what your product provides to customers. We can start our efforts to develop this diagram by identifying functions. In practice, this is best done by a group or team, and it should be done after all participants have become familiar with the list of customers at whom the product is aimed. A general list of functions can then be developed using brainstorming techniques or other group-based creativity tools. There are a few issues that you should keep in mind while simply listing functions. Functions must describe customer wants and needs from the viewpoint of the customer. A common problem is to confuse product functions with functions being performed by the customer, the designer, the engineer who created the product, or the manufacturer who produces the product. Again, think about a mechanical pencil. Many people will start by describing the function of the pencil as “write notes.” However, the pencil, by itself, cannott write anything. (If you can invent a * When you do this, you have created a “specification” for this function. ** One of the most common and debilitating errors in market analysis is to assume that others will respond the same way that you do. This is a simple but profound delusion. Most of us think that we are normal, typical people. When we awaken in the morning, we look in the mirror and see a normal (although perhaps disheveled if we look before the second cup of coffee) person. Thus, we think, “I like this widget. Since I am normal, most other people will like this widget, too. Therefore, my tastes are likely to be a good guideline to what my customers will want.” In most cases, even if you really are “normal” and even “typical,” this easy generalization is dangerously false.

SL3151Ch02Frame Page 57 Thursday, September 12, 2002 6:13 PM

Customer Understanding

57

pencil that will write notes without a writer attached you will probably become rich.) The function that is more appropriate for the pencil is “make marks.” The best way to start is to simply brainstorm as many functions as you can using active verbs and measurable nouns. There are many ways to brainstorm; in this case, it is usually easiest to have everyone involved use index cards or sticky notes to record their ideas. Remember that brainstorming should not be interrupted by criticism; just let the ideas flow. You will get things that do not apply, and, until you gain experience, you will not always use the “functive” structure that is ultimately important. Do not worry about these issues during the idea-generation phase of this process. Once you have a nice pile of cards or notes, start by sifting and sorting the ideas into categories. In any pile of ideas, there will be natural “groupings” of the cards. Determine these categories and then sort the cards. This can be thought of as “affinity diagramming” of the ideas. You will find some duplicates and some weird things that probably do not belong in the pile.* Discard the excess baggage and look at the categories. Are there any important functions you have missed? Do not hesitate to add new ideas to the categories, either. Finally, you are ready to bear down on the linguistic issues. Make sure that all of the ideas are expressed in terms of active verbs and measurable nouns. Change the idea to a “functive” construction, and then look for the “nerd” verb cards. Convert all of the “nerd verb” functions into true functives, with fully active verbs and measurable nouns. When you are done, you will have an interesting and important preliminary output. Now, count the cards again. If you have more than 20 to 30 cards, you have probably tackled too complex a subject or viewpoint. For example, a commercial airline has thousands — even hundreds of thousands — of functions. If you wanted to analyze function on the widest scale, you would probably be guilty of too much detail if you listed more than 30 functives. On the broadest scales of view, you may only list a handful of functions. Nothing is wrong with a short list, especially for the broadest view. If you have trouble, we can suggest some “function questions” that can assist you in your brainstorming. Try these questions: • What does it do? • If a product feature is deleted, what functions disappear? • If you were this element, what are you supposed to accomplish? Why do you exist? Ask the function questions in this order: • The entire scope of the project • A “system” view • Each element of the project * Do not automatically toss out strange ideas — see if the team can reword or express more clearly the idea that underlies the oddball cards or notes. Some percentage of these cards will have important information. Many will be eventual discards, but do not jump to conclusions.

SL3151Ch02Frame Page 58 Thursday, September 12, 2002 6:13 PM

58

Six Sigma and Beyond

• A “part” view • Each sub-element of the project • A “component” view Finally, we can start our next task, which consists of arranging functions into logical groups that show interrelationships. In addition, this next “arranging” step will allow us to test for completeness of function identification and improve team communication. We start by asking “What is the reason for the existence of the product or service?” This function represents the fundamental need of the customer. Example: a vacuum cleaner “sucks air” but the customer really needs “remove debris.” Whatever this reason for being is, we need to identify this particular function, which we call the task function. You must identify the task function from all of the functions you have listed. If you happen to find more than one task function, it is quite likely that you have actually taken on two products. For example, a clock-radio has two task functions: tell time and play music. However, you would be far better served by breaking your analysis into two components — one for telling time, the other for playing music. Alternatively, this product could be considered on a broader basis, as a system — in which case the task function might be “inform user,” with subordinate functions of “tell time” and “play music.” In any event, once you have identified the task function, you will realize that there are many functions other than the task function. Divide the remaining functions by asking: “Is the function required for the performance of the task function?”

If the answer to this question is yes, then the function can be termed essential. If the answer is no, then the function can be considered enhancing. All functions other than the task function must be either essential (necessary to the task function) or enhancing. So, your next task is to divide all of the remaining functions into these two general categories. You can further divide the enhancing functions — the functions that are not essential to the task function. Enhancing functions influence customer satisfaction and purchase decisions. Enhancing functions always divide into four categories: 1. 2. 3. 4.

Ensure dependability Ensure convenience Please senses Delight the customer*

* “Delight the customer” is actually quite rare — most enhancing functions fit one of the other three categories. If you do find a “delight the customer” function, try comparing this with an “excitement” feature in a Kano analysis; you should find that the function fits both descriptions.

SL3151Ch02Frame Page 59 Thursday, September 12, 2002 6:13 PM

Customer Understanding

59

None of these categories is needed to accomplish the task function. In fact, if you do not have a task function (and the associated essential functions), you probably do not have a product. The enhancing functions are those issues that purchasers weigh once they have determined that the task function will likely be fulfilled by your product. So, divide all of the enhancing functions into these four categories. Your next challenge is to create a function hierarchy that will, in finished form, be a function diagram. Start by asking this question: how does the product perform the task function? Primary essential functions provide a direct answer to this question without conditions or ambiguity. Secondary functions explain how primary functions are performed. Continue until the answer to “how” requires using a part name, labor operation, or activity, or you deplete your reserve of essential function cards. Now, you must reverse this process. Ask “why” in the reverse direction. For example, for a mechanical pencil, the task function is “make marks.” One of the functions you must perform to make marks is “support lead.” How do you support the lead? You do it by supporting the internal barrel tube (support tube) that carries the lead and by positioning this tube (position tube). Why do you support the tube and position the tube? You do this to support the lead. Why do you support the lead? You support the lead in order to make marks. The “chain” of function is driven by the how questions from the task function to primary then secondary functions — while this same chain is driven in reverse by why questions from secondary to primary to task function. As you progress, you will notice that you may be missing functions. If you find that you are, add additional functions as needed. After you have completed building “trees” of functions with the essential functions, repeat this process with the enhancing functions. The only difference is that the primary enhancing functions — ensure convenience, ensure dependability, please the senses, and delight the customer — have already been chosen. When you have finished, you will have a completed product flow diagram. At this point, try to delineate the extent of each function (range, target, specification, etc.) for each of the functions. Do not forget: Extent also tests “measurability” of each active verb–measurable noun combination. For example, for the mechanical pencil, the assembly may look like Figure 2.1. The sorted brainstorm list of functives may look like this: Entire project scope: • Make marks • Erase marks • Fit hand • Fit pocket • Show name • Display advertising • Convey message • Maintain point Tube assembly: • Store lead • Position lead

SL3151Ch02Frame Page 60 Thursday, September 12, 2002 6:13 PM

60

Six Sigma and Beyond

Eraser

Tube Assembly Lead Barrel EAT AT JOE’S

0,7 mm

Clip FIGURE 2.1 Paper pencil assembly.

• Feed lead • Reposition lead • Support lead • Locate eraser • Position tube • Generate force • Hold eraser Lead: • Make marks • Maintain point Eraser: • Erase marks • Locate eraser Barrel: • Support tube • Support lead • Position lead • Protect lead • Position tube • Position eraser • Show name • Display advertising • Convey message • Fit hand • Enhance feel • Provide instructions Clip: • Generate force • Position clip • Retain clip And, finally, the function diagram (only one possibility among many, many different results) may look like Figure 2.2.

SL3151Ch02Frame Page 61 Thursday, September 12, 2002 6:13 PM

Customer Understanding

61

Maintain point Support lead Basic Functions

Position lead

Support tube Position tube Store lead Feed lead Re-position lead

Make Marks

Ensure convenience

Erase marks

Hold eraser Retain clip

Fit pocket

Supporting Functions

Position clip

Ensure dependability

Provide Instructions

Please the senses

Enhance feel

Fit hand

Delight the customer

Display advertising Convey message

Show name

Position eraser Locate eraser Generate force

FIGURE 2.2 Function diagram for a mechanical pencil.

THE PROCESS FLOW DIAGRAM If you are working with a process rather than a product, you need to create a broad viewpoint “map” that shows how the activities in the process are accomplished. This can be done quickly and easily with a process flow diagram. The difficulty with most process flow diagrams is that they quickly bog down in too much detail. Whenever the detail gets too extensive, people lose interest (except for those who created the chart, but they are only part of the audience). Even though we need detail, we must avoid placing all of the details into one flow chart — at least if we want people to use the resulting charts. So, we will employ a “10 × 10” method that will aid in both communicating and managing the level of detail in a flow chart. If you keep the number of boxes in a flow chart to ten or fewer, most people will find your chart easy to read and understand. You can also use a “standard” symbol set for flow charting. After a great deal of trial and error from our experience, we have found that a simple set of ten symbols will explain almost any business process and provide enough options so that any team can easily illustrate what is going on — see Figure 2.3. By using some of the American National Standards Institute (ANSI) symbols and judiciously mixing in some easy-to-remember shapes, anyone can learn to flow chart a process in just a few minutes. The first step is to select a simple basis or point of view for your flow charts. This could be the view of the process operator, the work piece, or the process owner. (Be careful — if you

SL3151Ch02Frame Page 62 Thursday, September 12, 2002 6:13 PM

62

Six Sigma and Beyond

Significant

Incidental

move

move

Output

Process

Input

Decision

Store

Delay

Inspect

Document

FIGURE 2.3 Ten symbols for process flow charting.

confuse your viewpoint while developing a flow chart, you will quickly become confused about the process functions.) Inputs and outputs are the easiest steps to understand. You start with an input and you end with an output. A document may be a special kind of input or output — it can appear at the beginning, at the end, or during the overall process. The process box is the most common box; it describes transformations that occur within the process. Decisions are represented by a diamond shape, and an inspection step (in the shape of a stop sign) is just a special kind of decision. If you delay a process, you use a yield sign. If you store information, you use an inverted yield sign — a pile. Movement is also important. If a move is incidental, you tie the associated boxes together with a simple arrow. However, if a movement is complex (say, sending a courier package to Hong Kong as opposed to handing it to your next-cubicle neighbor), then you may have a special transformation or process step that we call a “significant” move, i.e., a large horizontal arrow. Let us look at a simple process for handling complaints. Your office deals with customer complaints, but you have a local factory (where your office is) and a factory in Japan. How you handle a complaint might look like Figure 2.4. This flow chart shows many of the symbols noted above, but it is not the only way that the process could be flow charted. However, if the team that developed the chart (once again, a team approach is likely to be the most effective technique) can reach a high level of consensus, then the communication of these ideas to others will be powerful and comprehensive. Now that the basics of 10 × 10 (ten steps or fewer using ten or fewer symbols) are apparent, it becomes possible to construct a “hierarchy” of flow charts that will fill in missing details that may have been skirted with the “10 step limit.” The next step is to create a new 10 × 10 flow chart for each box in the top level flow chart that requires additional explanation to reach the desired level of detail. These next flow charts (typically three to five of the boxes require additional detail) make up the second level flow charts. Wherever necessary, go to another level of flow charts; continue creating 10 × 10 flow charts until you have a hierarchy of flow charts that directly addresses all of the details that you feel are important. Finally, for each process box on each flow chart, you will have a process purpose. Why did you do this step? Simple — you had one (or possibly two) purposes in

SL3151Ch02Frame Page 63 Thursday, September 12, 2002 6:13 PM

Customer Understanding

63

Phone notice of customer complaint

Local or overseas factory?

Log complaint into database

Compile information for notice

Pending file

Local

Local factory is notified of complaint

Complaint notice

Overseas Send by courier to Japan

Factory is notified of complaint

FIGURE 2.4 Process flow for complaint handling.

mind when you designed this step into your process. Process purpose can be easily described using the language of function. Once again, you must use an active verb and a measurable noun. Often, a team can move directly to listing process functions from the flow charts. However, especially in manufacturing, it is common for the level of detail hidden in flow charts to be large, especially with intricate or subtle fabrication procedures. You may need to use an additional tool for teasing the “function” information from a flow chart called a “characteristic matrix.” A characteristic matrix is a reasonably simple analysis tool. The purpose of the matrix is to show the relationships between product characteristics and manufacturing steps. The importance of product characteristics in this matrix is significant; by considering the impact of a manufacturing step on product characteristics, we again focus our attention on customer requirements. Too often, manufacturing emphasis turns inward; it is critical that the focus be constantly directed at customers. Of course, there are “internal” customers as well. It is certainly important that intermediate characteristics, necessary for facilitating additional fabrication or assembly activities, be included in the analysis of function. For example, a simple machining process could have the characteristic matrix shown in Table 2.1. In this example, a simple machining step could be shown on a process flow chart with a process box that describes the machining operation as “CNC Lathe” or something similar. However, the lathe operation creates several important dimensions, or product characteristics, that are needed to meet customer expectations. These characteristics are sufficiently varied and complex that an additional level of detail is necessary. Some of these characteristics are important to the end customer; some are important to internal or “next step” process stations. For this example, the three left hand columns establish important functional information. The product characteristic is essentially the “measurable noun” (an

SL3151Ch02Frame Page 64 Thursday, September 12, 2002 6:13 PM

64

Six Sigma and Beyond

TABLE 2.1 Characteristic Matrix for a Machining Process Product Characteristic

Target Value

Tolerance

Process Operations Lathe Turn 10

Diameter “A”

6.22 mm

±0.25 mm

Diameter “B”

3.25 mm

±0.1 mm

Shoulder “C”

12.2 mm

±0.5 mm

Radius “D”

0.5 mm

±0.05 mm

Lathe Turn 20

X

Face Cut 30

Deburr 40

C

L

X

C

L

X

C

Cut Radius 50

L X

X = Characteristic Created By This Operation C = Characteristic Used For Clamp Down In This Operation L = Characteristics Used As Locating Datum In This Operation

occasional adjective is acceptable in a functive if there are several identical nouns, such as diameter in this case). The extent is shown in the target dimension and tolerance columns, and the “active verbs” can be constructed or deduced from the “code letters” inserted in the matrix cells in the “Process Operation” columns. In any event, whether you are able to determine functions directly from a process flow chart or whether you find the use of characteristic matrices important, you need to end with a comprehensive listing of function. The important aspect of process function is to use a flow charting technique of some type to assist in reaching the comprehensive assessment of function that is similar to the point-by-point listing that can be achieved by the product flow diagram technique.

USING FUNCTION CONCEPTS QUALITY METHODOLOGIES

WITH

PRODUCTIVITY

AND

Earlier, we suggested that function concepts form a powerful fundamental basis for three major productivity and quality methodologies: • Quality Function Deployment (QFD) • Failure Modes and Effects Analysis (FMEA) • Value Analysis (VA) While we do not intend to explain these techniques fully in this context (however, they will be explained later), we would like to address the usefulness of function concepts in these methodologies. In these discussions, we are assuming that you have a passing or even detailed familiarity with these tools. If not, you may wish to pass over to the discussion of QFD later in this chapter or to Chapters 6 and 12 for lengthy discussions of FMEA and VA.

SL3151Ch02Frame Page 65 Thursday, September 12, 2002 6:13 PM

Customer Understanding

65

For Quality Function Deployment, the most challenging issue is the one that we have just explored: how can one determine the functions that must be analyzed for deployment? In other forms, this is the same question facing practitioners in FMEA and VA. Clearly, the product flow diagram provides several instrumental techniques for improving these activities. A major difficulty in QFD is the often overwhelming complexity of the “House of Quality” approach. Constructing the first house, using conventional QFD techniques, is often the start of the complexity. Many different customer “wants” are listed. This is occasionally done as a “pre-planning” matrix. Moreover, the linguistic construction for these “wants” is undisciplined and subjective. Similarly, in FMEA, the initial list of failure modes is difficult to obtain. In VA, determining the “baseline” value assessment can also be difficult.* The techniques for developing a function diagram, especially the informal suggestions about “sizing” a project, can be very helpful in this regard. QFD, like FMEA and VA, typically fails to deliver the results expected because the project selected is too complex. A QFD study on a car or truck, for example, could easily contain hundreds of thousands of pages of information. That is not to say that the information in this study would not be valuable or that it should not be done; the issue is how complexity of this type should be dealt with. If you start with a systemwide view and construct a function diagram of the limited size previously discussed (20–30 functions maximum, even fewer are better), then this will provide a first level in a “hierarchy” of function diagrams. Subsequent analysis of various subsystems, then components and parts, and finally processes will complete the analysis. While the end result (for a car) would conceivably be of the same magnitude, the belief that all of the work must be done within the same team or by the same organization would be quickly abandoned. Moreover, the knowledge and understanding that is developed is generated at the hierarchical level (in the supply chain) of greatest importance, utility, and impact. Moreover, using the “functive” combination of active verbs and measurable nouns will assist in making QFD a useful tool. The vague, imprecise, or even confusing descriptions of function that are often used in QFD contribute to the difficulty in usage. A vehicle planning team may carry out a QFD study on the overall vehicle, assessing the major issues regarding the vehicle; these could include size, styling motifs, performance themes, and target markets. Subsequently, a study of the powertrain (engine, transmission, and axles) could be completed by another team. The engine itself could then be divided into major components: block, pistons, electronic controls, and so forth. Ultimately, suppliers of major and minor components alike would be asked to carry out QFD studies on each element. The multiplicity of information is still present, but it is no longer generated in some centralized form. This means that accessibility, usefulness, and the likelihood of beneficial deployment of the findings are much greater.** * In Value Analysis, the Function Analysis System Technique or “FAST,” a close cousin of the function diagram, is typically used to establish the initial functional baseline for value calculations. ** If the reader sees an “echo” of the hierarchy of flow charts, this is not coincidental.

SL3151Ch02Frame Page 66 Thursday, September 12, 2002 6:13 PM

66

Six Sigma and Beyond

As an added benefit, starting QFD using this approach provides benefits in the completion of FMEA and VA studies, since a consistent set of functions will be used as a basis for each technique. We will next consider each of these in turn. We will start with FMEA, because the importance of function in this methodology is not widely understood or appreciated. In FMEA, determining all of the appropriate failure modes is usually a great challenge. This obstacle is reflected in the widespread difficulty in understanding what is a failure mode and what is an effect. For example, the effect “customer is dissatisfied” is often found in FMEA studies. While this is likely to be true, it is an effect of little or no worth in developing and improving products and processes. Similarly, failure modes are often confused with effect. This can be illustrated with another common product, a disposable razor. How can we determine a comprehensive list of failure modes? Simply start with an appropriate function diagram. For each function, we need to consider how these functions can go astray. There are a limited number of ways that this can occur, all related to function. If you consider the completion of a function (at the desired extent) to be the absence of failure, then pose these questions about each function in the function diagram: • • • • • •

What would constitute an absence of function? What would occur if the function were incomplete? What would demonstrate a partial function? What would be observed if there was excess function? What would a decayed function consist of? What would happen if a function occurs too soon or too late (out of desired sequence)? • Could there be an additional unwanted function?

Each of these conditions establishes a possible failure mode. For the disposable razor, the task function is generally understood to be “cut hair” (not, of course, to shave). The failure mode that is most obvious is an additional unwanted function, namely “cut skin.” Notice that the mode of failure is not “feel pain” or “bleed;” these are failure effects. To make use of these ideas in the context of the function diagram, we must next define “terminus” functions. Terminus functions are simply those functions at the right hand (or “how”) end of any function chain in the function diagram. In the mechanical pencil example, two terminus functions would be “position eraser” and “locate eraser.” Why do you position and locate the eraser? To hold the eraser. Why do you hold the eraser? To erase marks. Why do you erase marks? To ensure correctness. Since this chain is one of enhancing functions, we do not directly modify the task function. Start your analysis of failure modes by testing each of the possible conditions listed above against the terminus functions. After you have completed the terminus functions, move one step in the “why” direction. However, as you move to the left, you will find that you frequently discover the same modes for the other functions. Since the function chain shows the interrelated nature of the functions, this should

SL3151Ch02Frame Page 67 Thursday, September 12, 2002 6:13 PM

Customer Understanding

67

not be surprising. As a rule, you will get most (if not all) of the relevant failure modes from the terminus functions.* So, starting with the terminus functions will speed your work and reduce redundancy. By working through each function chain in the function diagram, a comprehensive list of failure modes can be developed. This listing of failure modes then alters the approach to FMEA substantially; modes are clear, and cause-effect relationships are easier to understand. Moreover, developing FMEA studies using function diagrams that were originally constructed as part of the QFD discipline assures that product development activities continue to reflect the initial assumptions incorporated in the conceptual planning phase of the development process.** Once you have identified failure modes in association with functions, the remainder of the FMEA study — though still involved — is rather mechanical. For each failure mode, you must examine the likely effects that will result from this mode. With a clear mode statement, this is much simpler, and you are much less likely to confuse mode and effect issues. The effects can then be rated for severity using an appropriate table. With the effects in hand, causes can next be established and the occurrence rating estimated. Notice that this sequence of events makes the confusion of cause and effect much more difficult; in many cases, the logical improbability of reversal of cause and effect statements is so obvious that you simply cannot reverse these two issues. Finally, you can conclude the fundamental analysis with an evaluation of controls and detection. Once again, starting with a statement of function makes this clearer and less subject to ambiguity. Understanding the progression from function to mode to cause to effect sets the stage. What is it that you expect to detect? Is it a mode? In practice, detecting modes is extremely unlikely. You are more likely to detect effects. However, are effects what you want to detect? Once an effect is seen the failure has already occurred, and costs associated with the failure must already be absorbed. Let us return to the disposable razor to understand this. If the failure mode is “cut skin,” we must recognize that detecting “cut skin” is extremely difficult. You are much more likely to detect an effect — namely, pain or bleeding. Now, we recognize that we really do not want to detect failures at this point. Instead, we need to ask what are the possible causes of this failure mode. In this simple example, two different causes are readily apparent. From a design standpoint, the blades of the razor could be designed at the wrong angle to the shaver head. Even if the manufacturing were 100% accurate, a design that sets the blade angle incorrectly would have terrible consequences. On the other hand, the design could be correct; the blade angle could be specified at the optimum angle, but it could be assembled at an incorrect angle. Detection would best be aimed at testing the design angle*** and

* This is even more true for a system FMEA than for a design FMEA study. ** Of course, any change that is made in concept during development activities requires a continuous updating of the function diagrams under consideration. *** In the ISO and QS-9000 systems, we can think of this in terms of design verification.

SL3151Ch02Frame Page 68 Thursday, September 12, 2002 6:13 PM

68

Six Sigma and Beyond

at controlling the manufacturing process so that the optimum design angle would be repeatable (within limits) in production.* Finally, the Value Analysis process can also make use of the function diagrams that serve in the QFD and FMEA processes. In VA, the essence of the technique is the association of cost with function. Once this is accomplished, the method of functional realization can be considered in a variety of “what if” conditions. If there is a comprehensive statement of function, VA teams can be reasonably sure that ongoing value assessments, based on the ratio of function to cost, have a consistent and rational foundation. Moreover, the teams have a much higher confidence that these “what if” questions take customer issues into proper account. Too often, VA activities are carried out as if function is well understood and only cost matters. In too many cases, no function analysis is even performed. Despite the long-standing cautions against this, this alluring shortcut is often taken to save time, money, or both. The shortcomings of skipping function analysis in VA are not trivial. More disappointing results in usage of the VA methodology have probably been obtained because function was not fully and comprehensively understood. At a very fundamental level, how can a value ratio analysis be performed without a full statement of function? This is like calculating a return on investment without knowing the investment. Moreover, the analysis of value ratio can be misleading if the function issue is not well defined. It is easy to reduce cost. You simply eliminate features and functions from a product. Soon, you will not even be able to accomplish the task function. (In practice, “functionless” VA studies typically eliminate important enhancing functions that make a critical difference in the marketplace, and customers consequently pronounce unfavorable judgments on “decontented” products. VA then gets the blame.) Since value studies typically occur subsequent to QFD and FMEA in product development activities, the difficulty of understanding function is eliminated if function is fully defined and even specified during these earlier activities. By using function as the basis for product and manufacturing activities, a degree of focus and understanding of customer wants and needs is preserved not only during VA activities but throughout the product life cycle.

KANO MODEL The tool of choice that is preferred for understanding the “function” is the Kano model. A typical framework of the model is shown in Figure 2.5. The Kano model identifies three aspects of quality, each having a different effect on customer satisfaction. They are: 1. Basic quality — take for granted they exist 2. Performance quality — the more principle 3. Excitement quality — the wow * This is the issue of “process control” in the ISO and QS-9000 systems — in QS-9000, it goes to the heart of the control plan itself. Also, this is a simplified example. In more detail, the failure mode of “cut skin” can even occur when the blade angle is correct both in design and execution. A deeper examination of these issues quickly leads to the consideration of “robustness” in the design itself.

SL3151Ch02Frame Page 69 Thursday, September 12, 2002 6:13 PM

Customer Understanding

69

+ Y – axis (Customer satisfaction)

-

+ X – axis (product functionality)

-

FIGURE 2.5 Kano model framework.

The more we find out about these three aspects from the customer, the more successful we are going to be in our DFSS venture. (Caution: It is imperative to understand that the customer talks in everyday language, and that this language may or may not be acceptable from a design perspective. It is the engineer’s responsibility to translate the language data into a form that may prove worthwhile in requirements as well as verification. A good source for more detailed information is the 1993 book by Shoji.)

BASIC QUALITY “Basic” quality refers to items that the customer is dissatisfied with when the product performs poorly but is not more satisfied with when the product performs well. Fixing these items will not raise satisfaction beyond a minimum point. These items may be identified in the Kano model as in Figure 2.6. Some sources for the basic quality characteristics are: things gone right, things gone wrong, surrogate data, surveys, warranty, and market research.

PERFORMANCE QUALITY “Performance” quality refers to items that the customer is more satisfied with more of. In other words, the better the product performs the more satisfied the customer. The worse the product performs, the less satisfied the customer. Attributes that can be classified as linear satisfiers fall into this category. A typical depiction is shown in Figure 2.7. Some sources for performance quality characteristics are: internal satisfaction analysis, customer interviews, corporate targets/goals, competition, and benchmarking.

EXCITEMENT QUALITY “Excitement” quality refers to items that the customer is more satisfied with when the product is more functional but is not less satisfied with when it is not. This is the area where the customer can be really surprised and delighted. A typical depiction of these attributes is shown in Figure 2.8. Some sources for excitement quality characteristics are: customer insight, technology, interviews with comments such as high % or better than expected.

SL3151Ch02Frame Page 70 Thursday, September 12, 2002 6:13 PM

70

Six Sigma and Beyond

+ Y – axis (Customer satisfaction)

-

+ product functionality

Brakes Horn Windshield wipers -

FIGURE 2.6 Basic quality depicted in the Kano model.

Performance Customer satisfaction + Quiet gear shift

+ X – axis (product functionality)

Wind noise Power -

Fuel economy

FIGURE 2.7 Performance quality depicted in the Kano model.

Customer satisfaction + Style Ride Features -

+ X – axis (product functionality)

-

FIGURE 2.8 Excitement quality depicted in the Kano model.

Items that are identified as surprise/delight candidates are very fickle in the sense that they may change without warning. Indeed, they become expectations. The engineer must be very cautious here because items that are identified as excitement items now may not predict excitement at some future date. In fact, we already know that over time the surprised/delighted items become performance items, the performance items become basic, and the basic items become inherent attributes of the product. A classic example is the brakes of an automobile. The traditional brakes were the default item. However, when disc brakes came in as a new technology, they were indeed the excitement item of the hour. They were replaced, however,

SL3151Ch02Frame Page 71 Thursday, September 12, 2002 6:13 PM

Customer Understanding Surprise/delight

71 Customer satisfaction +

-

Performance

+ product functionality

Excitement quality over time -

FIGURE 2.9 Excitement quality depicted over time in the Kano model.

with the ABS brake system, and now even this is about to be displaced by the electronic brake system. This evolution may be seen in the Kano model in Figure 2.9. Developing these “surprised and delighted” items requires activities that gain insight into the customers’ emotions and motivations. It requires an investment of time to talk with and observe the customer in the customer’s own setting, and the use of the potential product. Above all, it requires the ability to read the customer’s latent needs and unspoken requirements. Is there a way to sustain the delight of the customer? We believe that there is. Once the attributes have been identified, a robust design must be initiated with two objectives in mind. 1. Minimize the degradation of these items. 2. Preserve the basic quality beyond expectations. These two steps will create an outstanding reliability and durability reputation.

QUALITY FUNCTION DEPLOYMENT (QFD) Now that we have finished the Kano analysis, and we know pretty much what the customer sees as functional and value added items, we are ready to organize all these attributes and then prioritize them. The methodology used is that of QFD. QFD is a planning tool that incorporates the voice of the customer into features that satisfy the customer. It does this by portraying the relationships between product or process whats and hows in a matrix form. The matrix form in its entirety is called the House of Quality — see Figure 2.10. One of the reasons why QFD is used is because it allows us to organize the Ys and ys and xs into a workable framework of understanding. QFD does not generate the Ys, ys, or xs. Ultimately, however, QFD will help in identifying the transfer function in the form Y = f(x, n) QFD was developed in Japan, with the intent to achieve competitive advantage in quality, cost, and timing. To understand this need, one must comprehend what quality control is all about from Japan’s point of view. Japan’s industrial standards define Quality Control (QC) as a system of means to economically produce goods and/or services that satisfy customer requirements. It is this definition of QC that

SL3151Ch02Frame Page 72 Thursday, September 12, 2002 6:13 PM

72

Six Sigma and Beyond

Correlation matrix HOW

Importance What

I M P O R T A N C E

Relationship matrix

I M P O R T A N C E

Competitive assessment

Technical difficulty How much Competitive assessment Important control items Importance

FIGURE 2.10 A typical House of Quality matrix.

propelled the Japanese to find not only a tool but a planning tool that implements the business objectives, of which the right application is product development. The definition of QFD is a systematic approach for translating customer wants/requirements into company-wide requirements. This translation takes place at each stage from research and development to engineering and manufacturing to marketing and sales and distribution. The QFD system concept is based on four key documents: 1. Overall customer requirement planning matrix. This document provides a way of turning general customer requirements into specified final product control characteristics. 2. Final product characteristic deployment matrix. This document translates the output of the planning matrix into critical component characteristics. 3. Process plan and quality control charts. These documents identify critical product and process parameters as well as benchmarks for each of those parameters.

SL3151Ch02Frame Page 73 Thursday, September 12, 2002 6:13 PM

Customer Understanding

73

4. Operating instructions. These documents identify operations to be performed by plant personnel to assure that the important parameters are achieved.

TERMS ASSOCIATED

WITH

QFD

There are six key terms associated with QFD: Quality function deployment — An overall concept that provides a means of translating customer requirements into the appropriate technical requirements for each stage of product development and production (i.e., marketing strategies, planning, product design and engineering, prototype evaluation, production process development, production, sales). This concept is further broken down into “product quality deployment” and “deployment of the quality function” (described below). Voice of the customer — The customers’ requirements expressed in their own terms. Counterpart characteristics — An expression of the voice of the customer in technical language that specifies customer-required quality; counterpart characteristics are critical final product control characteristics. Product quality deployment — Activities needed to translate the voice of the customer into counterpart characteristics. Deployment of the quality function — Activities needed to ensure that customer-required quality is achieved; the assignment of specific quality responsibilities to specific departments. (The phrase “quality function” does not refer to the quality department, but rather to any activity needed to ensure that quality is achieved, no matter which department performs the activity.) Quality tables — A series of matrices used to translate the voice of the customer into final product control characteristics.

BENEFITS

OF

QFD

QFD certainly appears to be a sensible approach to defining and executing the myriad of details embodied in the product development process, but it also appears to be a great deal of extra work. What is it really worth? Setting the logical arguments aside, there are a number of demonstrated benefits resulting from the use of QFD: • • • • • • • •

Demonstrated results Preservation of knowledge Fewer startup problems Lower startup cost Shorter lead time Warranty reduction Customer satisfaction Marketing advantage

SL3151Ch02Frame Page 74 Thursday, September 12, 2002 6:13 PM

74

Six Sigma and Beyond

Preservation of knowledge — The QFD charts form a repository of knowledge, which may (and should) be used in future design efforts. For example: Toyota is convinced that the QFD process will make good engineers into excellent engineers. An American engineering expert once commented, “There isn’t anything in the QFD chart I don’t already know.” Upon reflection, he realized that few other engineers knew everything on that chart. The QFD charts can be a knowledge base from which to train engineers. Fewer startup problems/lower startup cost — Toyota and other Japanese automobile manufacturers have found that the use of QFD more effectively “front loads” the engineering effort. This has substantially reduced the number of costly engineering changes at startup through a marked reduction of problems at startup. QFD has helped to identify potential problems early in design or avoid oversights through its disciplined approach. Shorter lead time — Toyota has reduced its product development cycle to less than 24 months. Warranty reduction — The corrosion problems with Japanese cars of the 1960s and 1970s led to enormous warranty expenses, significantly impacting profitability. The Toyota rust QFD study resulted in virtually eliminating corrosion and the resulting warranty expense. Customer satisfaction — The Japanese automobile manufacturers tend to focus on products that satisfy customers (as opposed to eliminating problems). The QFD approach has greatly facilitated the satisfying of customer wants. Domestic customer satisfaction surveys show that Japanese products have consistently scored higher than many American products. Marketing advantage — A Japanese manufacturer of earth moving equipment introduced a series of five new models that offered substantial advantages over their Caterpillar corporation counterparts, resulting in redistribution of market share. QFD brings several benefits to companies willing to undertake the study and training required to put the system in place. Some of these benefits as they relate to marketing advantage are: • Product objectives based on customers’ requirements are not misinterpreted at subsequent stops. • Particular marketing strategies’ “sales points” do not become lost or blurred during the translation process from marketing through planning and on to execution. • Important production control points are not overlooked. Everything necessary to achieve the desired outcome is understood and in place. • Tremendous efficiency is achieved because misinterpretation of program objectives, marketing strategy, and critical control points is minimized. See Figure 2.2. All of the above translate into significant marketing advantages, that is, speedy introduction of products that satisfy customers without problems. In addition to all the benefits already mentioned, Table 2.2 shows some of the benefits from the total development process perspective, which is a synergistic result starting with QFD.

SL3151Ch02Frame Page 75 Thursday, September 12, 2002 6:13 PM

Customer Understanding

75

TABLE 2.2 Benefits of Improved Total Development Process Cash Drain

Old Process

Technology push, but where’s the pull? Disregard for voice of the customer

Concepts with no needs, needs with no concept The voice of the engineer and other corporate specialists is emphasized Mad dash with singular concept, usually vulnerable Initial design is not production intent and emphasizes newness rather than superior design Make it look good for demonstration Large number of highly overlapped prototype iterations leaves little time for improvement Product is developed, then factory reacts to it Old process parameters used repetitiously without design improvement Inspection creates scrap, rework, adjustments, and field quality loss Lack of teamwork

Eureka concept Pretend designs

Pampered product Hardware swamps

Here is the product; where is the factory? We have always made it this way Inspection

Give me my targets, let me do my thing

ISSUES

WITH

Improved Process Technology strategy and technology transfer bring right technology to the product House of Quality and all steps of QFD deploy the voice of the customer throughout the process Pugh process converges on consensus and commitment to invulnerable concept Two step design and design competitive benchmarking lead to superior design

Taguchi optimization positions product as far as possible away from potential problems Only four iterations, each planned to make maximum contribution to optimization

One total development process, product, and production capability Taguchi process parameter improves quality, reduces cycle times Taguchi’s optimal checking and adjusting minimizes costs of inspection Teamwork and competitive benchmarking beat contracts, and targets lead the process, do not manage problems

TRADITIONAL QFD

The use of traditional QFD raises several issues for business people, including the following: 1. Change is uncomfortable. Counterpoint: There is an old saying, “If we do what we have done, we will get what we have.” To truly improve, we must explore new patterns of logical thinking and let go of outdated ways. We must be willing to change. 2. Success is not realized until the product is released. Counterpoint: The truest measure of customer satisfaction comes after the product or service is introduced. It is easy to lose sight of improvements that do not materialize until years after the improvement effort. We

SL3151Ch02Frame Page 76 Thursday, September 12, 2002 6:13 PM

76

Six Sigma and Beyond

3.

4.

5.

6.

would be remiss not to seek ways to achieve the end goal of customer satisfaction in our design and development process. QFD is a long process. Counterpoint: QFD saves the team’s time and resources with new approaches and tools. Avoiding multiple redesigns and multiple prototype levels in response to customer input recovers the time spent on QFD. The upstream time saves multiples of downstream time. It is not as much fun as “fire fighting.” Counterpoint. Finding and fixing problems may be personally gratifying. It is the stuff from which heroes/heroines are made. But emergencies are not in the company’s best interest and certainly not in the customer’s interest. Management must provide a system that rewards problem prevention as well as problem solving. The relation to the traditional product development process is not understood. Counterpoint: QFD replaces some traditional product design and development events, i.e., target setting and functional assumptions, and thereby does not add time. It is difficult to accept customer input when the “voice of the engineer” contradicts. Counterpoint: Engineering has delivered about 80% customer satisfaction; getting to 90–95% is a tough challenge requiring enhancements to current methods for achieving quality.

PROCESS OVERVIEW The easiest way to think of QFD is to think of it as a process consisting of linked spreadsheets arranged along a horizontal (Customer) axis and intersecting vertical (Technical) axis. Important details include the following: • From a macro perspective, the horizontal arrangement is referred to as the Customer Axis because it organizes the Customer Wants. • Customers are the people external to the organization who purchase, operate, and service your products. Customers can also be internal, i.e., the end users of your work within the organization. • The vertical arrangement is referred to as the Technical Axis Customer Wants into technical metrics. • The intersection of the axes (referred to as the Relationship Matrix) identifies how well engineering metrics correlate to customer satisfaction. • A closer look reveals that the interrelated matrices build upon one another beginning with a validated list of Customer Wants.

DEVELOPING

A

“QFD” PROJECT PLAN

Perhaps one of the most important issues in QFD is the selection of appropriate teams. Teams must share a common vision and mission to accomplish their objectives. Some of the reasons are:

SL3151Ch02Frame Page 77 Thursday, September 12, 2002 6:13 PM

Customer Understanding

• Building a project plan is the first critical team-building exercise • The project plan has been standardized in QFD, so all teams follow a basic strategy that includes the following steps: • Develop Project Plan to include safety standards and any governmental regulations, as well as timing. • Review Project Plan with program management for buy-in. • Complete the Customer Axis. • Review Customer Axis interim report with program management. • Complete Technical Axis. • Develop corporate strategy. • Develop final report. • Develop Deployment Plan for integrating into business cycle. • Communicate results to all programs and affected activities. The Customer Axis The steps necessary for completion of the customer axis include the following: Determining Customer Wants a. Obtain Customer Wants. b. Select relevant Customer Wants — about 30% of total Wants. c. Add applicable Wants. d. Set up focus groups, interviews, surveys, etc. e. Refine Customer Wants list. f. Enter Customer Wants into QFD net. g. Give Customer Wants to strategic standardization organization (SSO). Obtaining customer competitive evaluations a. Submit Customer Wants to market research (team). b. Develop mail-out questionnaire and/or clinic (market research). c. Send mail-out questionnaire and/or conduct clinic (market research). d. Report results to project team (market research). e. Enter customer competitive evaluation data into the internal team base. Setting customer targets a. Identify Customer Want (team). b. Review its Customer Desirability Index (CDI) rating and rank (team). c. Identify baseline product (team). d. Review customer competitive evaluations (team). e. Identify corporate strategy (team). Calculate image ratio for each Customer Want: customer target/baseline product. Calculate strategic CDI for each Customer Want: CDI × image ratio × sales point. f. Enter corporate strategy into customer targets matrix (team). g. Set customer targets — either opportunity to copy or sales point. h. If opportunity to copy, enter symbol into customer targets matrix. i. If sales point, enter values into customer targets matrix (team). j. End.

77

SL3151Ch02Frame Page 78 Thursday, September 12, 2002 6:13 PM

78

Six Sigma and Beyond

Determining Technical System Expectations (TSE) a. Review and adapt TSE template (team). b. Review past and current projects for additional TSEs (team). c. Identify and define new TSEs (team). d. Organize adapted list of TSEs (team). e. Enter TSEs into internal base (team). Determining relationships a. Review the relationship (team). b. Confirm/establish relationships (team and subject matter experts [SMEs]). c. Seek team consensus (team). d. Collect data and/or conduct experiments (team and SMEs) to find out whether disagreements exist. e. Check that each Want is satisfied by at least one TSE (team and SMEs). f. Enter into internal base. Technical competitive benchmarking • Buy, rent, lease or borrow competitive products (team). • Select TSEs to be benchmarked (team). • Establish inventory of benchmarking tests and data (team and SMEs). • Identify additional benchmarking tests required (team and SMEs). • Develop new tests (team and SMEs). • Conduct benchmark tests (team and SMEs). • Enter data into QFDNET (team). • Establish customer/engineer correlations (team and SMEs). Setting technical targets a. Develop technical targets (team and SMEs). b. Review existing program targets for existing TSEs (team). c. Recommend technical targets to program office (team and SMEs). d. Reconcile program targets and technical targets for existing TSEs (program office). e. Enter technical targets into QFDNET (team). The steps listed above will result in the following QFD deliverables for the Customer Axis: • Validated list of Customer Wants for the product, system, subsystem, or component • Customer Wants prioritized to focus engineering attention a. Customer Desirability Index of the most to least desirable Customer Wants b. Customer satisfaction targets for all Customer Wants, expressed as a percent over/under satisfaction of base product, system, subsystem, or component c. A final rank ordered strategic index of Customer Wants based on corporate strategies and competitive opportunities

SL3151Ch02Frame Page 79 Thursday, September 12, 2002 6:13 PM

Customer Understanding

79

Technical Axis On the Technical Axis, the following items will need to be produced: • Rank ordered list of key Technical System Expectations that when correctly targeted will satisfy Customer Wants at a strategically competitive level • Target values for key TSEs derived from technical competitive benchmarking that correlate with customer’s competitive evaluations. These target values aid program management two ways: a. By driving the product and engineering program toward integrated business and technical propositions that program management can prove b. With managing the program team’s performance at program completion Internal Standards and Tests • New or modified tests or other verification methods that make certain basic and product performance wants achieved Institutionalizing revised tests and standards into real world usage — customer dependent, of course — customer requirements, corporate engineering test procedures, and other documents both generic and program specific that support the organization’s design verification system.

THE QFD APPROACH The first concern of QFD is the customer. Therefore, in planning a new product we start with customer requirements, defined through market research. Generally, we call this the product development process, and it includes the program planning, conceptualization, optimization, development, prototyping, testing, and manufacturing functions. One can see that this development process is indeed very complex. Quite often, it cannot be performed by one individual. This is because it consists of several tradeoffs, such as: • • • • • • • •

Shared responsibilities Interpretations Priorities Technical knowledge Long time experience Resource changes Communication Lots of work

SL3151Ch02Frame Page 80 Thursday, September 12, 2002 6:13 PM

80

Six Sigma and Beyond

It is precisely this complexity that all too often causes the product development process to create a product that fails to meet the customer requirements. For example: Customer requirement → Design requirements → Part characteristics → Manufacturing operations → Production requirements Note: It is of paramount importance that the communication process within an organization does not fall victim to the use of jargon.

QFD METHODOLOGY QFD is accomplished through a series of charts that appear to be very complex. They do contain a great deal of information, however. That information is both an asset and a liability. All the charts are interconnected to what is called the House of Quality because of the roof-like structure at its top. Since this house is made up of distinct parts or “rooms,” let us find the function of each part, so that we can comprehend what QFD is all about — see Figure 2.10. QFD begins with a list of objectives or the “what” that we want to accomplish — see Figure 2.11. This is usually the voice of the customer and as such is very general, vague, and difficult to implement directly. It is given to us in raw form, that is, in the everyday language of the customer. (Example: “I don’t want a leaky window when it rains.”) For each what, we refine the list into the next level of detail by listing one or more “hows” for each what. The hows are an engineering task. Figure 2.11 shows the relationship between the what and the how. Figure 2.12 shows that it is possible to have an iterative process between the what and the how, with a possible refinement of the “old how” into the “new what” and ultimately to generate a very good “new how.” Even though this step shows greater detail than the original what list, it is by itself often not directly actionable and requires further definition. This is accomplished by further refinement until every item on the list is actionable. This level is important because there is no way of ensuring successful realization of a requirement that no one knows how to accomplish. (Note: Remember that our level of refinement within the how list may affect more than one how or what and can in fact adversely affect one another. That is why the arrows in Figure 2.11 are going in multiple directions.) To reduce possible confusion we represent the what and how in the following manner. The enclosed matrix becomes the relationships. The relationships are shown at the intersections of the what and how. Some common symbols are: □ Medium relationship  Weak relationship  Very strong relationship

SL3151Ch02Frame Page 81 Thursday, September 12, 2002 6:13 PM

Customer Understanding

81

What

How

FIGURE 2.11 The initial “what” of the customer.

What

How/What

How

FIGURE 2.12 The iterative process of “what” to “how.”

The method of using symbols allows very complex relationships to be shown, and the interpretation is easy and is not dependent on experience. There are many variations of this, and readers are encouraged to use what is comfortable for them. Figure 2.13 presents a typical matrix. Once the what, how, and relationships have been identified, the next step is to establish a “how much” for each how — see Figure 2.14. The intent here is to provide specific objectives that guide the subsequent design and provide a means of objectivity to the process. The result is minimum interference from opinion. (Note: This how much is another cross check on our thinking process. It forces us to think in a very detailed, measurable fashion.) To summarize: The what identifies the customer’s requirements in everyday language. The how refines the customer’s requirements (from an engineering perspective). The relationship defines the relationship between what and how via a symbolic language. The how much provides an objective means of assessing that requirements have been met and provides targets for further detail development. Pictorially, the flow is shown in Figure 2.14.

SL3151Ch02Frame Page 82 Thursday, September 12, 2002 6:13 PM

82

Six Sigma and Beyond

How What

Importance 4

•

5 •

1 3 •

2 How much Importance ratings

42

21

33

28

24

Where = 3 •

= 9 = 1

Therefore: (4x9) + (2x3) = 42 and so on. Make sure that the ratings differentiate to the point of discrimination between each other. Remember, you are interested in great differentiation rather than a simple priority.

FIGURE 2.13 The relationship matrix.

HOW What

How much

FIGURE 2.14 The conversion of “how” to “how much.”

At this point, even though a lot of information is at hand, it is not unusual to refine the hows even further until an actionable level of detail is achieved. This is done by creating a new chart in which the hows of the previous chart become the whats of the new chart. The “how much” information as a general rule is carried along to the next chart to facilitate communication. This is done to ensure that the objectives are not lost. The process is repeated as necessary. In the product development process, this means taking the customer requirements and defining design requirements that are carried on to the next chart to establish the basis for the part characteristic. This is continued to define the manufacturing operations and the production requirements — see Figure 2.15. (Note: The greatest gains using QFD can be realized only when

SL3151Ch02Frame Page 83 Thursday, September 12, 2002 6:13 PM

Customer Understanding

83

Functional spec VOC

Requirements analysis Design System design

Where: VOC = Voice of the customer

Methods, tools, procedures Technical assessment

Resource plan Implementation plan

FIGURE 2.15 The flow of information in the process of developing the final “House of Quality.”

taken down to the work detail level of production requirements. The QFD process is well suited to simultaneous engineering in which product and process engineers participate in a team effort.) For more information on the cascading process of the QFD methodology, see the Appendix. So far, we have talked about the basic charts in the House of Quality, and as a result we have gained much information about the problem at hand. However, there are several useful extensions to the basic QFD charts that enhance their usefulness. These are used as required based on the content and purpose of each particular chart. One such extension is the correlation matrix. The correlation matrix — see Figure 2.10 — is a triangular table often attached to the “hows.” The purpose of such placement is to establish the correlation between each “how” item, i.e., to indicate the strength of the relationship and to describe the direction of the relationship. To do that, symbols are used, most commonly:  Positive  Strong positive

X Negative # Strong negative

A second extension is the competitive assessment — see Figure 2.10. This is a pair of graphs that shows item for item how competitive products compare with current company products. Its strength is the fact that it can be done for the whats, hows, and how muchs. The competitive assessment may also be used to uncover gaps in engineering judgment. What and how items that are strongly related should also exhibit a relationship in the competitive assessment. For example, if we believe superior dampening will result in an improved ride, the competitive assessment would be expected to show that products with superior dampening also have a superior ride. If this does not occur, it calls attention to the possibility that something significant may have been overlooked. If not acted upon, we may achieve superior performance

SL3151Ch02Frame Page 84 Thursday, September 12, 2002 6:13 PM

84

Six Sigma and Beyond

to our “in house” tests and standards but fail to achieve expected results in the hands of our customers. Why are we doing this? Basically, for two reasons: 1. To establish the values of the objectives to be achieved 2. To uncover engineering judgment errors Remember that the correlation must be related to real world usage from the customer’s perspective. What and how items that are strongly related should also be shown to relate to one another in the competitive assessment. If the correlation does not agree, it may mean that we overlooked something very significant. A third extension is the importance rating — see Figure 2.10. This is a mechanism for prioritizing efforts and making trade-off decisions for each of the whats and hows. It is important to keep in mind that the values by themselves have no direct meaning; rather, their meaning surfaces only when they are interpreted by comparing their magnitudes. The importance rating is useful for prioritizing efforts and making trade-off decisions. (Some of the trade-offs may require high level decisions because they cross engineering group, department, divisional, or company lines. Early resolution of trade-offs is essential to shorten program timing and avoid non-productive internal iterations while seeking a nonexistent solution.) The rating itself may take the form of numerical tables or graphs that depict the relative importance of each what or how to the desired end result. Any rating scale will work, provided that the scale is a weighted one. A common method is to assign weights to each relationship matrix symbol and sum the weights, just as we did in Figure 2.13. Another more technical way is the following: w ′functioni =

∑w r

yj ij

j

wfunction i =

5(w ′functioni ) maxi (w ′functioni )

where w ′functioni = unnormalized function importance; wyj = importance rating; rij = individual rating of functions; and wfunction i = weighted function importance. Applying this methodology to Figure 2.13 yields Figure 2.16.

QFD

AND

PLANNING

Contrary to what the name implies, quality function deployment (QFD) is not just a quality tool. QFD was developed in Japan, growing out of the need to simultaneously achieve a competitive advantage in quality, cost, and timing. To better comprehend QFD, it is important to understand what the Japanese mean by the word “quality.” The word “quality,” which we generally define as conformance to requirements, fitness for use, or some other measure of goodness, takes on a much broader meaning

SL3151Ch02Frame Page 85 Thursday, September 12, 2002 6:13 PM

Customer Understanding

85

HOW What

Importance 4

•

5

•

1 3

•

2 How much Importance 42 ratings – unnormalized Importance of 5 how

21

33

28

24

2.5 or 3

3.9 or 4

3.3 or 3

2.9 or 3

W’ function i = (4x9) + (2x3) = 42 and so on W function i =

5 (42) 42

= 5 and so on

Keep in mind that when you are addressing the “hows” in essence you are dealing with customer functionalities. Therefore, it is recommended to design for the average, based on each function’s importance according to its capability to supply each original Y.

FIGURE 2.16 Alternative method of calculating importance.

in Japan (there is probably no exact English translation of the Japanese version). However, according to Japanese industrial standard Z8101–1981, “quality control” is “a system of means to economically produce goods or services which satisfy customer requirements.” (Italics added.) Thus to the Japanese, “quality” means conducting the business effectively, not just producing a good product. In this context, QFD really becomes a planning tool for implementing business objectives, of which the most widely known application is to product development. In planning a new product, we start with customer needs, wants, and expectations, often defined through market research. We wish to design and manufacture a product that satisfies the customer’s perception of intended function, as well as or better than our competitors (subject to certain internal company constraints). In other words: CUSTOMER REQUIREMENTS ⇓ ⇓ ⇓ PRODUCT

SL3151Ch02Frame Page 86 Thursday, September 12, 2002 6:13 PM

86

Six Sigma and Beyond

Let us call the process of translating these requirements into a viable product the “product development process.” This process includes program planning, concepting, optimization, development, prototyping, and testing, as well as the corresponding manufacturing functions. Thus: CUSTOMER REQUIREMENTS ⇓ ⇓ ⇓ PRODUCT DEVELOPMENT PROCESS ⇓ ⇓ ⇓ PRODUCT In a large organization, the product development process is so detailed that often no one individual can comprehend it all. For some, the process looks like a maze or a mysterious “black box.” For others the process is an intricate network of activities. Regardless of how it is represented, the product development process is exceedingly complex, consisting of numerous trade-offs. Shared responsibilities and interpretation differences often result in conflicting priorities. That is the reason the team must have ownership of the projects and must have a substantial body of technical knowledge over a relatively long time frame while enduring resource changes. This, of course, requires a great deal of communication and a substantial work effort.

PRODUCT DEVELOPMENT PROCESS The complexity of the product development process makes it a natural haven for Murphy’s law, with nearly an infinite number of opportunities for problems to occur. Despite the best of intentions and efforts, all too often the product development process creates a product that fails to meet the customer requirements. Such failures may occur due to: • • • • • • • •

Trade-offs Shared responsibilities Interpretations Priorities Technical knowledge Long time frame Resource changes Communication — lots of work

The QFD approach focuses on customer requirements in a manner that directs efforts toward achieving those requirements — see Figure 2.17. In Figure 2.17, for

SL3151Ch02Frame Page 87 Thursday, September 12, 2002 6:13 PM

Customer Understanding

87

Design requirements

Product planning

Part characteristic Part deployment

Manufacturing operations

Process planning

Production requirements

Production planning

FIGURE 2.17 The development of QFD.

each of the customer requirements, a set of design requirements is determined, which if satisfied will result in achieving the customer requirements. In like manner, each design requirement is evolved into part characteristics, which in turn are used to determine manufacturing operations and specific production requirements. The flow is as follows: CUSTOMER REQUIREMENTS ⇓ ⇓ ⇓ DESIGN REQUIREMENTS ⇓ ⇓ ⇓

SL3151Ch02Frame Page 88 Thursday, September 12, 2002 6:13 PM

88

Six Sigma and Beyond

PART CHARACTERISTICS ⇓ ⇓ ⇓ MANUFACTURING OPERATIONS ⇓ ⇓ ⇓ PRODUCTION REQUIREMENTS ⇓ ⇓ ⇓ So, for example: The customer requirement of “years of durability” may be achieved in part by the design requirement of no visible rust in three years. This in turn may be achieved in part by ensuring part characteristics that include a minimum paint film build and maximum surface treatment crystal size. The manufacturing process that provides these part characteristics consists of a three-coat process that includes a dip tank. The production requirements are the process parameters within the manufacturing process that must be controlled in order to achieve the required part characteristics (and ultimately the customer requirements). Therefore, we can present this in a summary form as: CUSTOMER REQUIREMENT: Years of durability DESIGN REQUIREMENT: No visible exterior rust in 3 years PART CHARACTERISTICS: Paint weight — 2–2.5 gm/m2; Crystal size — 3 max MANUFACTURING OPERATIONS: Dip tank; 3 coats PRODUCTION REQUIREMENTS: Time = 2.0 minutes; Acidity = 1.5 to 2.0; Temperature = 45–55ο C

CONJOINT ANALYSIS WHAT IS CONJOINT ANALYSIS? We introduced conjoint analysis in Volume III of this series. Recall that conjoint analysis is a multivariate technique used specifically to understand how respondents develop preferences for products or services. It is based on the simple premise that consumers evaluate the value of a product/service/idea (real or hypothetical) by combining the separate amounts of value provided by each attribute. It is this characteristic that is of interest in the DFSS methodology. After all, we want to know the bundle of utility from the customer’s perspective. (The reader is encouraged to review Volume III, Chapter 11.) So in this section, rather than dwelling on theoretical statistical explanations, we will apply conjoint analysis in a couple of hypothetical examples. The examples are based on the work of Hair et al. (1998) and are used here with the publisher’s permission.

SL3151Ch02Frame Page 89 Thursday, September 12, 2002 6:13 PM

Customer Understanding

A HYPOTHETICAL EXAMPLE

89 OF

CONJOINT ANALYSIS

As an illustration of conjoint analysis, let us assume that HATCO is trying to develop a new industrial cleanser. After discussion with sales representatives and focus groups, management decides that three attributes are important: cleaning ingredients, convenience of use, and brand name. To operationalize these attributes, the researchers create three factors with two levels each: Factor Ingredients Form Brand name

Level Phosphate-free Liquid HATCO

Phosphate-based Powder Generic brand

A hypothetical cleaning product can be constructed by selecting one level of each attribute. For the three attributes (factors) with two values (levels), eight (2 × 2 × 2) combinations can be formed. Three examples of the eight possible combinations (stimuli) are: • HATCO phosphate-free powder • Generic phosphate-based liquid • Generic phosphate-free liquid HATCO customers are then asked either to rank-order the eight stimuli in terms of preference or to rate each combination on a preference scale (perhaps a 1-to-10 scale). We can see why conjoint analysis is also called “trade-off analysis,” because in making a judgment on a hypothetical product, respondents must consider both the “good” and “bad” characteristics of the product in forming a preference. Thus, respondents must weigh all attributes simultaneously in making their judgments. By constructing specific combinations (stimuli), the researcher is attempting to understand a respondent’s preference structure. The preference structure “explains” not only how important each factor is in the overall decision, but also how the differing levels within a factor influence the formation of an overall preference (utility). In our example, conjoint analysis would assess the relative impact of each brand name (HATCO versus generic), each form (powder versus liquid), and the different cleaning ingredients (phosphate-free versus phosphate-based) in determining the utility to a person. This utility, which represents the total “worth” or overall preference of an object, can be thought of as based on the part-worths for each level. The general form of a conjoint model can be shown as (Total worth for product)ij…,n = Part-worth of level i for factor 1 + Part-worth of level j for factor 2 +... + Part-worth of level n for factor m where the product or service has m attributes, each having n levels. The product consists of level i of factor 2, level j of factor 2, and so forth, up to level n for factor m.

SL3151Ch02Frame Page 90 Thursday, September 12, 2002 6:13 PM

90

Six Sigma and Beyond

TABLE 2.3 Stimuli Descriptions and Respondent Rankings for Conjoint Analysis of Industrial Cleanser Stimuli Descriptions

1 2 3 4 5 6 7 8

Respondent Rankings

Form

Ingredients

Brand

Liquid Liquid Liquid Liquid Powder Powder Powder Powder

Phosphate-free Phosphate-free Phosphate-based Phosphate-based Phosphate-free Phosphate-free Phosphate-based Phosphate-based

HATCO Generic HATCO Generic HATCO Generic HATCO Generic

Respondent 1

Respondent 2

1 2 5 6 3 4 7 8

1 2 3 4 7 5 8 6

In our example, a simple additive model would represent the preference structure for the industrial cleanser as based on the three factors (utility = brand effect + ingredient effect + form effect). The preference for a specific cleanser product can be directly calculated from the part-worth values. For example, the preference for HATCO phosphate-free powder is: Utility = Part-worth of HATCO brand + Part-worth of phosphate-free cleaning ingredient + Part-worth of powder With the part-worth estimates, the preference of an individual can be estimated for any combination of factors. Moreover, the preference structure would reveal the factor(s) most important in determining overall utility and product choice. The choices of multiple respondents could also be combined to represent the competitive environment faced in the “real world.”

AN EMPIRICAL EXAMPLE To illustrate a simple conjoint analysis, assume that the industrial cleanser experiment was conducted with respondents who purchased industrial supplies. Each respondent was presented with eight descriptions of cleanser products (stimuli) and asked to rank them in order of preference for purchase (1 = most preferred; 8 = least preferred). The eight stimuli are described in Table 2.3, along with the rank orders given by two respondents. As we examine the responses for respondent 1, we see that the ranks for the stimuli with the phosphate-free ingredients are the highest possible (1, 2, 3, and 4), whereas the phosphate-based product has the four lowest ranks (5, 6, 7, and 8). Thus, the phosphate-free product is much more preferred than the phosphate-based cleanser. This can be contrasted to the ranks for the two brands, which show a mixture of high and low ranks for each brand. Assuming that the basic model (an

SL3151Ch02Frame Page 91 Thursday, September 12, 2002 6:13 PM

Customer Understanding

91

additive model) applies, we can calculate the impact of each level as differences (deviations) from the overall mean ranking. (Readers may note that this is analogous to multiple regression with dummy variables or ANOVA.) For example, the average ranks for the two cleanser ingredients (phosphate-free versus phosphate-based) for respondent 1 are: Phosphate-free: (1 + 2 + 3 + 4)/4 = 2.5 Phosphate-based: (5 + 6 + 7 + 8)/4 = 6.5 With the average rank of the eight stimuli of 4.5 [(1 + 2 + 3 + 4 + 5 + 6 + 7 + 8)/8 = 36/8 = 4.5], the phosphate-free level would then have a deviation of –2.0 (2.5 – 4.5) from the overall average, whereas the phosphate-based level would have a deviation of +2.0 (6.5 – 4.5). The average ranks and deviations for each factor from the overall average rank (4.5) for respondents 1 and 2 are given in Table 2.4. In our example, we use smaller numbers to indicate higher ranks and a more preferred stimulus (e.g., 1 = most preferred). When the preference measure is inversely related to preference, such as here, we reverse the signs of the deviations in the part-worth calculations so that positive deviations will be associated with part-worths indicating greater preference. Deviation is calculated as: deviation = average rank of level – overall average rank (4.5). Note that negative deviations imply more preferred rankings. The part-worths of each level are calculated in four steps: • Step 1: Square the deviations and find their sum across all levels. • Step 2: Calculate a standardizing value that is equal to the total number of levels divided by the sum of squared deviations. • Step 3: Standardize each squared deviation by multiplying it by the standardizing value. • Step 4: Estimate the part-worth by taking the square root of the standardized squared deviation. Let us examine how we would calculate the part-worth of the first level of ingredients (phosphate-free) for respondent 1. The deviations from 2.5 are squared. The squared deviations are summed (10.5). The number of levels is six (three factors with two levels apiece). Thus, the standardizing value is calculated as .571 (6/10.5 = .571). The squared deviation for phosphate-free (22; remember that we reverse signs) is then multiplied by .571 to get 2.284 (22 × .571 = 2.284). Finally, to calculate the part-worth for this level, we then take the square root of 2.284, for a value of 1.1511. This process yields part-worths for each level for respondents 1 and 2, as shown in Table 2.5. Because the part-worth estimates are on a common scale, we can compute the relative importance of each factor directly. The importance of a factor is represented by the range of its levels (i.e., the difference between the highest and lowest values) divided by the sum of the ranges across all factors. For example, for respondent 1, the ranges are 1.512 [.756 – (–.756)], 3.022 [1.511 – (–1.511)], and .756 [.378 – (–.378)]. The sum total of ranges is 5.290. The relative importance for form, ingredients,

SL3151Ch02Frame Page 92 Thursday, September 12, 2002 6:13 PM

92

Six Sigma and Beyond

TABLE 2.4 Average Ranks and Deviations for Respondents 1 and 2 Factor Level

Ranks Across Stimuli

Average Rank of Level

Deviation from Overall Average Rank

Respondent l Form Liquid Powder

1, 2, 5, 6 3, 4, 7, 8

3.5 5.5

–1.0 +1.0

Ingredients Phosphate-free Phosphate-based

1, 2, 3, 4 5, 6, 7, 8

2.5 6.5

–2.0 +2.0

Brand HATCO Generic

1, 3, 5, 7 2, 4, 6, 8

4.0 5.0

–.5 +.5

Respondent 2 Form Liquid Powder

1, 2, 3, 4 5, 6, 7, 8

2.5 6.5

–2.0 +2.0

Ingredients Phosphate-free Phosphate-based

1, 2, 5, 7 3, 4, 6, 8

3.75 5.25

–.75 +.75

Brand HATCO Generic

1, 3, 7, 8 2, 4, 5, 6

4.75 4.25

+.25 –.25

and brand is calculated as 1.512/5.290, 3.022/5.290, and .756/5.290, or 28.6, 57.1, and 14.3 percent, respectively. We can follow the same procedure for the second respondent and calculate the importance of each factor, with the results of form (66.7 percent), ingredients (25 percent), and brand (8.3 percent). These calculations for respondents 1 and 2 are also shown in Table 2.5. To examine the ability of this model to predict the actual choices of the respondents, we predict preference order by summing the part-worths for the different combinations of factor levels and then rank ordering the resulting scores. The calculations for both respondents for all eight stimuli are shown in Table 2.4. Comparing the predicted preference order to the respondent’s actual preference order assesses predictive accuracy. Note that the total part-worth values have no real meaning except as a means of developing the preference order and, as such, are not compared across respondents. The predicted and actual preference orders for both respondents are given in Table 2.6.

SL3151Ch02Frame Page 93 Thursday, September 12, 2002 6:13 PM

Customer Understanding

93

TABLE 2.5 Estimated Part-Worths and Factor Importance for Respondents 1 and 2 Estimated Part-Worths Factor Level

Reversed Squared Deviationa Deviation

Standardized Deviationb

Calculating Factor Importance Estimated Range of Factor Part-Worthc Part-Worths Importanced

Respondent 1 Form Liquid Powder

+1.0 –1.0

1.0 1.0

+.571 –.571

+.756 –.756

1.512

28.6%

Ingredients Phosphate-free Phosphate-based

+2.0 –2.0

4.0 4.0

+2.284 –2.284

+1.511 –1.511

3.022

57.1%

+.5 –.5

.25 .25 10.5

+.143 –.143

+.378 –.378

.756

14.3%

Brand HATCO Generic Sum of squared deviations Standardizing valuee Sum of part-worth ranges

.571 5.290

Respondent 2 Form Liquid Powder Ingredients Phosphate-free Phosphate-based Brand HATCO Generic Sum of squared deviations Standardizing value Sum of part-worth ranges a

+2.0 –2.0

4.0 4.0

+2.60 –2.60

+1.612 –1.612

3.224

66.7%

+.75 –.75

.5625 .5625

+.365 –.365

+.604 –.604

1.208

25.0%

–.25 +.25

.0625 .0625 9.25

–.04 +.04

–.20 +.20

.400

8.3%

.649 4.832

Deviations are reversed to indicate higher preference for lower ranks. Sign of deviation used to indicate sign of estimated part-worth. b Standardized deviation equal to the squared deviation times the standardizing value. c Estimated part-worth equal to the square root of the standardized deviation. d Factor importance equal to the range of a factor divided by the sum of the ranges across all factors, multiplied by 100 to yield a percentage. e Standardizing value equal to the number of levels (2 + 2 + 2 = 6) divided by the sum of the squared deviations.

SL3151Ch02Frame Page 94 Thursday, September 12, 2002 6:13 PM

94

Six Sigma and Beyond

TABLE 2.6 Predicted Part-Worth Totals and Comparison of Actual and Estimated Preference Rankings Stimuli Description Size

Part-Worth Estimates

Preference Rankings

Ingredients

Estimated

Size

Ingredients

Brand

Brand Total

Actual

Liquid Liquid Liquid Liquid Powder Powder Powder Powder

Phosphate-free Phosphate-free Phosphate-based Phosphate-based Phosphate-free Phosphate-free Phosphate-based Phosphate-based

HATCO Generic HATCO Generic HATCO Generic HATCO Generic

Respondent 1 .756 1.511 .756 1.511 .756 –1.511 .756 –1.511 –.756 1.511 –.756 1.511 –.756 –1.511 –.756 –1.511

.378 –.378 .378 –.378 .378 –.378 .378 –.378

2.645 1.889 –.377 –1.133 1.133 .377 –1.889 –2.645

1 2 5 6 3 4 7 8

1 2 5 6 3 4 7 8

Liquid Liquid Liquid Liquid Powder Powder Powder Powder

Phosphate-free Phosphate-free Phosphate-based Phosphate-based Phosphate-free Phosphate-free Phosphate-based Phosphate-based

HATCO Generic HATCO Generic HATCO Generic HATCO Generic

Respondent 2 1.612 .604 1.612 .604 1.612 –.604 1.612 –.604 –1.612 .604 –1.612 .604 –1.612 –.604 –1.612 –.604

–.20 .20 –.20 .20 –.20 .20 –.20 .20

2.016 2.416 .808 1.208 –1.208 –.808 –2.416 –2.016

2 1 4 3 6 5 8 7

1 2 3 4 7 5 8 6

The estimated part-worths predict the preference order perfectly for respondent 1. This indicates that the preference structure was successfully represented in the part-worth estimates and that the respondent made choices consistent with the preference structure. The need for consistency is seen when the rankings for respondent 2 are examined. For example, the average rank for the generic brand is lower than that for the HATCO brand (refer to Table 2.4), meaning that, all things being equal, the stimuli with the generic brand will be more preferred. Yet, examining the actual rank orders, this is not always seen. Stimuli 1 and 2 are equal except for brand name, yet HATCO is preferred. This also occurs for stimuli 3 and 4. However, the correct ordering (generic preferred over HATCO) is seen for the stimuli pairs of 5–6 and 7–8. Thus, the preference structure of the part-worths will have a difficult time predicting this choice pattern. When we compare the actual and predicted rank orders (see Table 2.6), we see that respondent 2’s choices are many times mispredicted but most often just miss by one position due to the brand effect. Thus, we would conclude that the preference structure is an adequate representation of the choice process for the more important factors, but that it does not predict choice perfectly for respondent 2, as it does for respondent 1.

SL3151Ch02Frame Page 95 Thursday, September 12, 2002 6:13 PM

Customer Understanding

THE MANAGERIAL USES

OF

95

CONJOINT ANALYSIS

It is beyond the scope of this section to discuss the statistical basis of conjoint analysis. However, in DFSS, we should understand the technique in terms of its role in decision making and strategy development. The simple example we have just discussed presents some of the basic benefits of conjoint analysis. The flexibility of conjoint analysis gives rise to its application in almost any area in which decisions are studied. Conjoint analysis assumes that any set of objects (e.g., brands, companies) or concepts (e.g., positioning, benefits, images) is evaluated as a bundle of attributes. Having determined the contribution of each factor to the consumer’s overall evaluation, the marketing researcher could then: 1. Define the object or concept with the optimum combination of features 2. Show the relative contributions of each attribute and each level to the overall evaluation of the object 3. Use estimates of purchaser or customer judgments to predict preferences among objects with differing sets of features (other things held constant) 4. Isolate groups of potential customers who place differing importance on the features to define high and low potential segments 5. Identify marketing opportunities by exploring the market potential for feature combinations not currently available The knowledge of the preference structure for each individual allows the researcher almost unlimited flexibility in examining both individual and aggregate reactions to a wide range of product- or service-related issues.

REFERENCES Fowler, T.C., Value Analysis in Design, Van Nostrand Reinhold, New York, 1990. Hair, J.F., Multivariate Data Analysis, 5th ed., Prentice Hall, Upper Saddle River, NJ, 1998. Harry, M.,The Vision of Six Sigma: A Roadmap for Breakthrough, 5th ed., Vol. 1, TriStar Publishing, Phoenix, 1997. Porter, M., Competitive Advantage, Free Press, New York, 1985. Rechtin, E. and Maier M., The Art of Systems Architecting, CRC, Boca Raton, FL, 1997. Shoji, S., A New American TQM, Productivity Press, Portland, OR, 1993.

SELECTED BIBLIOGRAPHY Afors, C. and Michaels, M.Z., A Quick, Accurate Way to Determine Customer Needs, Quality Progress, July 2001, pp. 82–88. Anon., Quality Function Deployment, American Supplier Institute, Inc., Dearborn, MI, 1988. Bialowas, P. and Tabaszewska E., How to Evaluate the Internal Customer Supplier Relationship, Quality Progress, July 2001, pp. 63–67. Carlzon, J., Moments of Truth, HarperCollins, New York, 1989. Fredericks, J. O. and Salter, J.M., What Does Your Customer Really Want? Quality Progress, Jan. 1998, pp. 63–70.

SL3151Ch02Frame Page 96 Thursday, September 12, 2002 6:13 PM

96

Six Sigma and Beyond

Gale, B.T., Managing Customer Value: Creating Quality and Service that Customers Can See, Free Press, New York, 1994. Gobits, R., The Measurement of Insight, unpublished paper presented at the 2nd International Symposium on Educational Testing, Montreux, 1975. Goncalves, K.P. and Goncalves, M.P., Use of the Kano Method Keeps Honeywell Attuned to the Voice of the Customer, Quirk’s Marketing Research Review, Apr. 2001, pp. 18–25. Gutman, J. and Miaoulis, G., Past Experience Drives Future CS Behavior, Marketing News, Oct. 22, 2001, pp. 45–46. Harry, M., The Vision of Six Sigma: A Roadmap for Breakthrough, 5th ed., Vol. 2, TriStar Publishing, Phoenix, 1997. James, H.L., Sasser, W.E., and Schlesinger, L.A., The Service Profit Chain: How Leading Companies Link Profit and Growth to Loyalty, Satisfaction and Value, Free Press, New York, 1997. Mariampolski. H, Qualitative Market Research, Sage Publications, Newbury Park, CA, 2001. Morais, R., The End of Focus Groups, Quirk’s Marketing Research Review, pp. 153–154, May 2001. Mudge, A.E., Numerical Evaluation of Functional Relationships, Proceedings, Society of American Value Engineers, 1967. Murphy, B., Methodological Pitfalls in Linking Customer Satisfaction with Profitability, Quirk’s Marketing Research Review, Oct. 2001, pp. 22–27. Murphy, B., Qualitatively Speaking: Of Bullies, Friends and Mice, Quirk’s Marketing Research Review, Oct. 2001, pp. 16, 61. Saliba, M.T. and Fisher, C.M., Managing Customer Value, Quality Progress, June 2000, pp. 63–70. Shillito, M.L., Pareto Voting. Proceedings, Society of American Value Engineers, 1973. Stamatis, D.H., Total Quality Management: Engineering Handbook, Marcel Dekker, New York, 1997. Stamatis, D.H., Total Quality Service, St. Lucie Press, Delray Beach, FL, 1996. Sullivan, L.P., The Seven Stages in Company Wide Quality Control, Quality Progress, May 1986, pp. 77–83. Sullivan, L.P., Quality Function Deployment, Quality Progress, June 1986, 1986, pp. 39–50. Thomas, J. and Sasser, W.E., Why Satisfied Customers Defect, Harvard Business Review, Nov.-Dec. 1995, pp. 88–89. VanVierah, S. and Olosky, M., Achieving Customer Satisfaction: Registrar Satisfaction Survey Counterbalances the Myth About Registrars, Automotive Excellence, Winter 1999, pp. 10–15. Veins, M., Wedel, M., and Wilms, T., Metric conjoint segmentation methods: a Monte Carlo comparison, Journal of Marketing Research, 33, 73–85, 1996. Wittink, D.R. et al., Commercial use of conjoint analysis: an update, Journal of Marketing, 53, 91–96, 1989.

SL3151Ch03Frame Page 97 Thursday, September 12, 2002 6:12 PM

3

Benchmarking

Benchmarking is a tool, a technique or process, a philosophy, and a new name for old practices. It involves operations research and management science for determining (a) what to do or “goal setting” and (b) how to do it or “action plan identification.” Benchmarking can be applied (a) systematically and comprehensively or (b) ad hoc project by project. In both cases it can require (a) sophisticated statistical analysis, (b) utilization of a wide variety of analytical tools, and (c) a wide range of data sources. The basic requirements for success are: • • • • • •

Time, effort, and resources A willingness to learn and to change Continuing, long-term top management support An external focus on customers and competitors A common-sense approach and active listening The ability to look at the old in a new way

GENERAL INTRODUCTION TO BENCHMARKING A BRIEF HISTORY

OF

BENCHMARKING

The term “benchmarking” was coined by Xerox in 1979. Xerox has now performed over 400 benchmark studies, and the process is totally integrated at all levels as part of the business planning process. The approach has actually been in use for a number of years — although it was often called by different names. (Reverse engineering is an approach used to study the design and manufacturing characteristics of competitive products. Benchmarking of computer hardware and software is a very common practice.) Benchmarking extends the concept to consider administrative and all management processes. There is a conscious attempt to compare with the “best of the best” even — especially — if that is not a direct competitor. The fundamental process in starting benchmarking is to think about the area to be benchmarked, which can be just about anything, and ask yourself, “Who is especially good at that? What can I creatively imitate?” A typical process for doing a benchmarking is shown in Figure 3.1.

POTENTIAL AREAS

OF

APPLICATION

OF

BENCHMARKING

Benchmarking is a methodology that can be used along with other systematic, comprehensive management approaches to improve performance. It is not an end unto itself. Some examples of applications of benchmarking include: 97

SL3151Ch03Frame Page 98 Thursday, September 12, 2002 6:12 PM

98

Six Sigma and Beyond

Prepare and respond to surveys

Agree to site visits

Builds customer goodwill Builds network of benchmarking partners May provide “in” to target companies but time consuming Must be managed to gain long term benefit Must be viewed as an investment

Two-way site visits

Informal search for the best

Define a process A benchmarking study IS a project Make sure you have clearance with legal department Involve the process owner Avoid the following mentality: We are unique We know it all It was not invented here It is too complex We already tried it and it does not work here

Follow a model

Form consortium group

Provides true improvement opportunities Answers “How do the best do it? Provides actionable data But Time consuming, must be focused Disciplined approach builds results Must be treated as an ongoing way of doing business

FIGURE 3.1 The benchmarking continuum process.

Broad management focus • Cost reduction • Profit improvement • Business strategy development • Total quality management Individual management processes • Improving customer service • Reducing product development time • Market planning • Product distribution Highly specific focus • Invoice design • Sales force compensation • Fork lift truck maintenance The critical questions to ask are: • What are the areas that potentially could be benchmarked? • How do you prioritize and focus the efforts?

SL3151Ch03Frame Page 99 Thursday, September 12, 2002 6:12 PM

Benchmarking

99

BENCHMARKING AND BUSINESS STRATEGY DEVELOPMENT Hall (1980) observed that certain industry leaders had exceptional performance even in the bad times of 1979–1980. For example:

Goodyear Inland Steel Paccar Caterpillar General Motors Maytag G. Heilman Brewing Philip Morris Average

Company ROE

Industry ROE

9.2 10.9 22.8 23.5 19.8 27.8 25.8 22.7 20.2

7.4 7.1 15.4 15.4 15.4 10.1 14.1 18.2 12.9

How can this be so? What strategy did the more successful competitors follow?

LEAST COST

AND

DIFFERENTIATION

Hall’s study itself is an early example of successful benchmarking. By extensive interviewing and data analysis, Hall reached conclusions based on the performance and the experience of a group of highly successful companies. As determined by Hall and also described in the book Competitive Strategy by Michael Porter, the successful competitor tends to follow one of two strategies: • Least cost • Differentiation Those competitors who do not explicitly follow one strategy or the other tend to get “stuck in the middle” and do not have the highest return on investment. Hall’s findings do, however, indicate that some firms can successfully manage both strategy options. The generic strategies identified by Hall and Porter have been supported by a number of research studies (see Higgins and Vincze, 1989). For a successful business strategy to be developed, a company must decide what course it will follow. It must also be certain that it is, in fact, realistically able to pursue that alternative. Some questions to be asked include the following: • Does a company really have the least cost? How do they know? What is the basis for the claim? • Is the company really differentiated in the eyes of the customer? How do they know? What is the basis for the claim? • How might competitive conditions change in the future? Benchmarking can provide — in part — the information necessary to answer these questions by providing focus and insight on what the best companies are doing.

SL3151Ch03Frame Page 100 Thursday, September 12, 2002 6:12 PM

100

Six Sigma and Beyond

In addition to making a choice relative to least cost versus differentiation, an important strategy choice is that of being a mass marketer versus supplying the needs of a specific market segment. Therefore, when benchmarking is performed the following must always be present: • • • • • •

Build a relationship with your benchmarking partner. Establish trust and mutual interest. Be worthy of trust. Make it last. Be open to reciprocity. Follow a code of conduct. • Principle of confidentiality • Principle of first party contact • Principle of preparation • Principle of third party contact

CHARACTERISTICS

OF A

LEAST COST STRATEGY

A firm following the least cost strategy must be able to deliver a product or service with acceptable quality at a lower total cost than any of its competitors. Note that total cost is the critical concern. The company does not have to be least cost in every aspect of the business. The fact that the total cost is the lowest does not necessarily mean that the price that is charged is the lowest. To determine if the least cost strategy is viable, it is necessary to perform competitive benchmarking and gain information relative to the following: • What is the relative market share of the company? Does the experience curve have a significant effect on cost reduction? • Is the industry one that can be affected by automation possibilities, conveyorized assembly, or new production technology? Is the capital available for investment in efficient scale facilities and product and process engineering innovation? • Do competitors have a different mix of fixed and variable costs? • What is the percent capacity utilization by competitive firms? • Are the competitive firms using activity-based accounting? • How critical is raw material supply? Does the firm have preemptive sources of supply? • Does the firm have a tight system of budgeting and cost control for all functions? • Are productions designed for low cost productions? Are products simplified and product lines reduced in number? Are bills of material standardized? • What is the level of product/service quality versus competition? • How labor intensive is the process? How effective are labor/management relations? • Are marginal accounts minimized?

SL3151Ch03Frame Page 101 Thursday, September 12, 2002 6:12 PM

Benchmarking

101

Improved quality through benchmarking can lead to lower costs. The cost of quality — really the cost of non-quality — consists of the costs of prevention, appraisal (inspection), internal quality failures, and external quality failures. This cost can amount to as much as 30–40% of the cost of goods sold. Costs include the following: Costs of prevention Training Equipment Costs of appraisal (inspection) Inspectors Equipment Cost of internal quality failures Scrap Rework Machine downtime Missed schedules Excess inventory Cost of external quality failures Warranty expense Customer dissatisfaction Studies have shown that the average quality improvement project results in $100,000 of cost reduction. The associated cost to diagnose and remedy the problem has averaged $15,000. Consequently, the payout from benchmarking in this area can be significant. Velcro reported a 50% reduction in waste as a percentage of total manufacturing cost in the first year and an additional 45% decrease in the second year of its quality program. Motorola achieved a quality level in 1991 that was 100 times better than it was in 1987. By 1992, this company was striving for six sigma quality. That means three defects per million or 99.9997 percent perfection. Motorola believes that super quality is the lowest cost way of doing things, if you do things right the first time. Their director of manufacturing — at that time — pointed out that each piece of equipment has 17,000 parts and 144,000 opportunities for mistakes. A 99 percent quality rate is equivalent to 1,440 mistakes per piece. The cost to hire and train people to fix those mistakes would put the company out of business.

CHARACTERISTICS

OF A

DIFFERENTIATED STRATEGY

A firm following the differentiation strategy must be able to provide a unique product or service to meet the customer’s expectations. The challenge of being unique is that of providing a sustainable source of differentiation. It is very difficult to create something that is totally sustainable. This may depend upon a corporate culture producing a positive attitude toward quality and customer service or perhaps the value of information or computer-to-computer linkages.

SL3151Ch03Frame Page 102 Thursday, September 12, 2002 6:12 PM

102

Six Sigma and Beyond

Following a differentiation strategy does not mean that a company can be inefficient relative to costs. Although cost is not the primary driving force, costs still must be minimized for the degree of differentiation provided. To determine if the differentiation strategy is viable it is necessary to perform competitive benchmarking and gain information relative to segmentation. When developing corporate or marketing strategy, it is important to identify the different market segments that make up the total market. A market segment is a group of customers with similar or related buying motives. The members of the segment have similar needs, wants, and expectations. A focus on market segments allows a company to tailor its products, services, pricing, distribution, and communication message to meet the specific needs of a market. The opposite of market segmentation is mass marketing. Segmentation allows a smaller company to successfully compete with a larger company by concentrating resources at the specific point of competition. Any market can be segmented. The toothpaste market, for example, can be segmented into the sensory segment (principal benefit sought is flavor or product appearance), the sociable segment (brightness of teeth), the worriers (decay prevention), and the least cost buyer. To segment a market you need to know who the customers are, what they buy, how they buy, when they buy, why they buy, and where they buy. Some typical questions in this area are: • • • •

How do you segment your market? What do you do differently for each of these segments? How does the competition segment the market? What new segments are likely to develop due to changes in sociological factors, technology, legislation, economic conditions, or growing internationalism?

BENCHMARKING AND STRATEGIC QUALITY MANAGEMENT Strategic Quality Management (SQM) or Total Quality Management (TQM), as defined by J.M. Juran, W. Edwards Deming and others, consists of a systematic approach for setting and meeting quality goals throughout a company. Just as companies have set out to achieve financial goals through a process of corporate business planning, so also can companies achieve quality goals by SQM or TQM or six sigma. An overly simplified definition of TQM is “Doing the right thing, right the first time, on time, all the time; always striving for improvement, and always satisfying the customer.” This requires a focus on customer needs, people, systems and process, and a supportive cultural environment. But this really is not any different from what the six sigma methodology proposes. The essential steps of the quality management process consist of: Quality planning • Identifying target market segments • Determining specific customers’ needs, wants, and expectations

SL3151Ch03Frame Page 103 Thursday, September 12, 2002 6:12 PM

Benchmarking

103

• Translating the customer needs into product and process requirements • Designing products and processes with the required characteristics (Competitive benchmarking can assist in this part of the process.) Quality control • Measuring actual quality performance versus the design goals • Diagnosing the causes of poor quality and initiating the required corrective steps • Establishing controls to maintain the gains Quality improvement • Establishing a benchmarking process • Providing the necessary resources It is important to note that the process: • Is strategic in nature, proactive • Is competitively focused on meeting customer needs as opposed to techniques of analysis • Is goal oriented • Is comprehensive in terms of level and functions • Manages in quality, not simply defect reduction The following are very closely linked: • • • • • •

Six sigma Business strategic planning Strategy development (least cost versus differentiation) TQM Pricing strategy Benchmarking

The classical approach to benchmarking viewed as process — which has become the de facto process — has the following characteristics: • • • • •

Inspection to control defects is primary tool. Better quality means higher costs. Significant scrap and rework activity takes place. Quality control is found only in manufacturing. SPC is used as an example; other tools are used occasionally.

Top management commitment • Level 5: Continuous improvement is a natural behavior even for routine tasks. • Level 4: Focus is on improving the system. • Level 3: Adequate money and time are allocated to continuous improvement and training. • Level 2: There is a balance of long-term goals with short-term objectives.

SL3151Ch03Frame Page 104 Thursday, September 12, 2002 6:12 PM

104

Six Sigma and Beyond

• Level 1: The traditional approach is in place. Note that the Level 1 commitment is the status quo, and not much is happening. It is the least effective way of demonstrating to the organization at large that management commitment is a way of life. On the other hand, Level 5 is the most effective and demands change of some kind. Obsession with excellence • Level 5: Constant improvement in quality, cost, and productivity • Level 4: Use of cross-functional improvement teams • Level 3: TQM and six sigma support system set up and in use • Level 2: Executive steering committee set up • Level 1: Traditional approach Organization is customer satisfaction driven • Level 5: Customer satisfaction is the primary goal. More customers desire a long-term relationship. • Level 4: Striving to improve value to customers is a routine behavior. • Level 3: Customer feedback is used in decision making. • Level 2: Customer rating of company is known. • Level 1: The traditional approach is in place. Supplier involvement • Level 5: Suppliers fully qualified in all benchmark areas • Level 4: Suppliers actively implementing TQM and aware of the six sigma demands • Level 3: Direct involvement in supplier awareness training; supplier criteria in place • Level 2: Suppliers knowledgeable about your TQM as well as the six sigma direction; supplier number reduction started • Level 1: Traditional approach Continuous learning • Level 5: Training in TQM and six sigma tools is common among all employees. • Level 4: Top management understands and applies TQM and the six sigma philosophy. • Level 3: Ongoing training programs are in place. • Level 2: A training plan has been developed. • Level 1: The traditional approach is in place. Employee involvement • Level 5: People involvement; self-directed work groups. • Level 4: Manager defines limits, asks group to make decisions. • Level 3: Manager presents problem, gets suggestions, makes decision. • Level 2: Manager presents ideas and invites questions, makes decision. • Level 1: The traditional approach is used. Use of incentives • Level 5: Gainsharing • Level 4: More team than individual incentives and rewards • Level 3: Quality-related employee selection and promotion criteria

SL3151Ch03Frame Page 105 Thursday, September 12, 2002 6:12 PM

Benchmarking

• Level • Level Use of tools • Level • Level • Level • Level • Level

105

2: Effective employee suggestion program used 1: Traditional approach 5: Statistics a common language among all employees 4: More team than individual incentives and rewards 3: SPC used for variation reduction 2: SPC used in manufacturing 1: Traditional approach

The Malcolm Baldrige National Quality Award encapsulates the essential elements of Strategic Quality Management. The key attributes considered when making this award are listed below. Many agree that the criteria provide the blueprint for a better company. The urgency to win the award can accelerate change within an organization. Some companies have told their suppliers to compete or else. These are the criteria: • Quality is defined by the customer. • The senior management of a business needs to have clear quality values and build the values into the way the company operates on a day-to-day basis. • Quality excellence derives from well-designed and well-executed systems and processes. • Continuous improvement must be part of the management of all systems and processes. • Companies need to develop goals, as well as strategic and operational plans, to achieve quality leadership. • Shortening the response time of all company operations and processes needs to be part of the quality improvement effort. • Operations and decisions of the company need to be based on facts and data. • All employees must be suitably trained and involved in quality activities. • Design quality and defect and error prevention should be major elements of the quality system. • Companies need to communicate quality requirements to suppliers and work with suppliers to elevate supplier quality performance. Achievement of the award requires extensive top management effort and support. All of the Quality Award winners have been in highly competitive industries and either had to improve or get out of the business. On a scale of 10 (best) to 1 (poor), how would you rate your company on each of these attributes? If you find yourself on the low end, there may be a need for benchmarking.

BENCHMARKING

AND

SIX SIGMA

Within the information and analysis part of the examination or survey, the practitioners of benchmarking look specifically at competitive comparisons and benchmarks. It has been reported in the literature that many companies do not do enough

SL3151Ch03Frame Page 106 Thursday, September 12, 2002 6:12 PM

106

Six Sigma and Beyond

in the way of benchmarking. They compare themselves against other manufacturers but do not make comparisons with outside businesses or even “true” best-in-class companies. A six sigma company is expected to describe the company’s approach to selecting quality-related competitive comparisons and world-class benchmarks to support quality planning, evaluation, and improvement. The specific areas to address are: • Criteria and rationale the company uses for making competitive comparisons and benchmarks. These include: • The relationship to company goals and priorities for the improvement of product and service quality and/or company operations • The companies for comparison within or outside the industry • Current scope of competitive and benchmark involvement and data collection relative to: • Product and service quality • Customer satisfaction and other customer needs • Supplier performance • Employee data • Internal operations, business processes, and support services • Other • For each, the company is directed to list sources of comparisons and benchmarks, including companies benchmarked and independent testing or evaluation, and: • How each type of data is used • How the company evaluates and improves the scope, sources, and uses of competitive and benchmark data • The company must also indicate how this data is used to support: • Company planning • Setting of priorities • Quality performance review • Improvement of internal operations • Determination of product or service features that best predict customer satisfaction • Quality improvement projects Specific uses of benchmarking are to assist in: • • • •

Developing plans Goal setting Continuous process improvement Determining trends and levels of product and service quality, the effectiveness of business practices, and supplier quality • Determining customer satisfaction levels

A closer review of the criteria indicates several factors that are essential for effective quality excellence and benchmarking activities within a company, including:

SL3151Ch03Frame Page 107 Thursday, September 12, 2002 6:12 PM

Benchmarking

107

Customer-driven quality • Quality is judged by the customer. The customer’s expectations of quality dictate product design and this, in turn, drives manufacturing. • All product and service attributes that lead to customer satisfaction and preference must be taken into consideration. • Customer driven quality is a strategic concept. Why do people buy your product? How do you know? • Leadership is crucial. A company’s senior management must create clear quality values, specific goals, and well-defined systems and methods for achieving the goals. • Ongoing personal involvement is essential. The attitude must be changed from a “management control” focus to a “management committed to help you” focus. Continual improvement • Constant improvement in many directions is required: improved products and services, reduced errors and defects, improved responsiveness, and improved efficiency and effectiveness in the use of resources. All of this takes time. If you do not have the time, do not start. Fast response • An increasing need exists for shorter new product and service development and introduction cycles and a more rapid response to customers. Actions based on facts, data and analysis • A wide range of facts and data is required, e.g., customer satisfaction, competitive evaluations, supplier data, and data relative to internal operations. • Performance indicators to track operational and competitive performance are critical. These performance indicators or goals can act as the cohesive or unifying force within an organization. They can also provide the basis for recognition and reward. • Participation by all employees is important. Reward and recognition systems need to reinforce total participation and the emphasis on quality. • Factors bearing on the safety, health, and well being of employees need to be included in the improvement objectives. • Effective training is required. The emphasis must be on preventing mistakes, not merely correcting them. Employees must be trained to inspect their own work on a continuous basis. • Participation with suppliers is essential. It is important to get suppliers to improve their quality standards.

NATIONAL QUALITY AWARD WINNERS

AND

BENCHMARKING

Example — Cadillac To show the strong relationship between National Quality Award winners and benchmarking, we provide a historical perspective. The first example comes to us from

SL3151Ch03Frame Page 108 Thursday, September 12, 2002 6:12 PM

108

Six Sigma and Beyond

Cadillac’s approach to excellence. (Cadillac was the 1990 winner of the National Quality Award.) The brief case study that follows indicates the integration of business planning, excellent quality management, and benchmarking. The Business Plan was the Quality Plan. The plan was designed to ensure that Cadillac is the “Standard of the World” in all measurable areas of quality and customer service. The major components of the plan were: • Mission • Objectives • Quality — Emphasis on six major vehicle systems: • Exterior component and body mechanical • Chassis/powertrain • Seats and interior trim • Electrical/electronics • Body in white • Instrument panel • Competitiveness • Disciplined planning and execution • Leadership and people • Goals For each objective, the following issues were addressed: • What are the measurable performance indicators of quality and customer service? When answering, consider both the product itself and the management process that led to the improved product or service. • What does the customer need or want? • What levels are achieved by the best-of-class companies considering both direct competitors and any other company? • What are the time-phased quality improvement goals? Action plans • Took appropriate and applicable action to fulfill all the requirements so that the customer could be satisfied. A Second Example — Xerox In the early 1980s, Xerox realized that Japanese competition was selling products for less than the Xerox cost. Many of the required reforms focused on Xerox suppliers because the cost of purchases amounted to 80% of the copiers cost of goods sold. Xerox asked suppliers to restate their company performance data so that the supplier could be compared with the best of class Xerox could find anywhere in the world. Some of the benchmarks Xerox used to measure operations proficiency included: • Ratio of functional cost to revenue (percent) • Headcount per unit of output • Overhead rate (dollars/hour)

SL3151Ch03Frame Page 109 Thursday, September 12, 2002 6:12 PM

Benchmarking

• • • • • • • • • • •

109

Cost per order entered Cost per engineering drawing Customer satisfaction rating (index value) Internal and external defect rates (parts per million) Service response time (hours) Billing error rates Days inventory on hand Total manufacturing lead time (days) New product development time (weeks) Percent of parts delivered on time New ideas per employee

Xerox reduced its number of suppliers from 5,000 in 1980 to 300 by 1986 based on performance data and attitude. Suppliers were classified as: (a) does not think improvement is necessary, (b) slow to accept or manage change, and (c) willing to go for it and strong enough to be a survivor. Xerox reallocated its internal efforts to concentrate on the companies in the third group. Xerox provided extensive training to these companies, and defect rates in incoming materials dropped 90 percent in three years. In addition to performance improvement, the suppliers were asked to participate in copier design, as early in the concept phase as possible, and to make suggestions so that overall quality could be improved and costs reduced. When this information was used, the cost of purchased material dropped by 50 percent. Third Example — IBM Rochester IBM Rochester describes its quality journey as follows: 1981 1984 1986

Vision Goal Vision Goal Vision Goals

1989

Vision

1990–1994

Goal Vision

Goal

Product reliability Zero defects Process effectiveness and efficiency All process rated Customer and supplier partnerships Competitive and functional benchmarks Best of competition Over 350 benchmarking teams are in place; scores of benchmarking studies have been completed; strategic targets are derived from the comprehensive benchmarking process Market-driven customer satisfaction Total business process focus Closed loop quality management system Total customer satisfaction Customer — the final arbiter Quality — excellence in execution Products and services — first with the best People — enabled, empowered, excited, rewarded Undisputed leadership in customer satisfaction

SL3151Ch03Frame Page 110 Thursday, September 12, 2002 6:12 PM

110

Six Sigma and Beyond

Results: A 30 percent improvement in productivity occurred between 1986 and 1989. This was a period of extensive benchmarking activity. Product development time has been reduced by more than half, and manufacturing cycle time has been trimmed by 60 percent since 1983. Fourth Example — Motorola Each of the firm’s six major groups and sectors have “benchmarking” programs that analyze all aspects of a competitor’s products to assess their manufacturability, reliability, manufacturing cost, and performance. Motorola has measured the products of some 125 companies against its own standards, verifying that many Motorola products rank as “best in their class.” (It is imperative for the reader to understand that the result of a benchmarking study may indeed provide the researcher with data to support the assertion that the current practices of your own organization are the “best in class.”)

BENCHMARKING

AND THE

DEMING MANAGEMENT METHOD

There is a very close relationship between the approach of W. Edwards Deming and that specified by the requirements of the National Quality Award. The potential role of benchmarking to implement certain aspects of the Deming approach is apparent. Deming’s fourteen points are summarized below: 1. Create constancy of purpose for the improvement of product and services. 2. Adopt the new philosophy that quality is critical for the competitive survival of a company. 3. Cease dependence on mass inspection, and create the processes that build a quality product from the start. 4. End the practice of awarding business based on price alone, and take into consideration the quality of products and services received. 5. Improve constantly and forever the system of production and service. This begins with product design and goes through every phase of business operations. 6. Institute training and retraining. 7. Provide leadership and the resources required to get the job done. 8. Drive out the fear of admitting problems and suggesting new and different ways of doing things. Get around the not invented here syndrome. 9. Break down interdepartmental barriers so that all departments can work toward the common objective of satisfying the customer. 10. Eliminate slogans, exhortations, and targets for the workforce without providing the ways and means for accomplishment. Do not tell people what to do without telling them how to do it and providing the systems and support necessary.

SL3151Ch03Frame Page 111 Thursday, September 12, 2002 6:12 PM

Benchmarking

111

11. Eliminate numerical quotas. These often promote poor quality. Instead analyze the process to determine the systemic changes required to enable superior performance. 12. Remove barriers to pride in workmanship by providing the training, communication, and facilities required. 13. Institute a vigorous program of education and retraining. Help people to improve every day. 14. Take action to accomplish the transformation required.

BENCHMARKING

AND THE

SHEWHART CYCLE

OR

DEMING WHEEL

Plan Study a process to determine what changes might be made to improve it. What type of performance is achieved by the best of the best? What do they do that we are not doing? What results do they achieve? What changes would we have to make? What does the customer expect? What is the customer level of satisfaction? Is the change economically justified? Do Determine the specific plan for improvement and implement it. This involves the development of creative alternatives by work teams and the conscious choice of a strategy to be followed. This may require internal or external benchmarking. Study — Observe the Effects Was the root cause of the problem identified and corrected? Will the problem recur? Are the expected results being achieved? Act Study the results and repeat the process. Was the plan a good one? What was learned? This approach amounts to the application of the scientific method to the solution of business problems. It is the basis of organizational learning.

WHY DO PEOPLE BUY? Differentiation and quality management both focus on the need to meet customer needs, wants, and expectations. Why does a person buy a particular product? • One view: (marketing based) • A second view: (psychologically based) How can we define quality? This is a very critical question and may indeed prove the most important question in pursuing benchmarking. The importance of this question is that it will focus the research on “best” in a very customized fashion

SL3151Ch03Frame Page 112 Thursday, September 12, 2002 6:12 PM

112

Six Sigma and Beyond

from the organization’s perspective. This is a question that must be addressed as early as possible.

ALTERNATIVE DEFINITIONS

OF

QUALITY

People buy a combination of products and services for a price that depends upon the perception of the value received. In order to conduct benchmarking studies relative to quality, it is important to define the elusive term “quality.” Garvin (1988) and Stamatis (1996, 1997) provide various definitions of quality as follows: Garvin’s eight dimensions of product quality are: Performance — Performance refers to the ability of the product to perform up to expectations relative to its primary operating characteristics. For example, a camera can be self-focusing and automatically adjust the lens opening. Products can often be ranked in terms of levels of performance, i.e., good, better, best. People’s expectations differ depending upon the task to be performed. Products are designed for different uses. Therefore, a failure to perform might simply indicate another product class or market segment focus and not inferior quality. Features — Features are secondary attributes that affect a product’s performance. For example, the camera mentioned above can weigh less than two pounds. A car can have power steering as a feature. Features can often be bundled or unbundled. The distinction between performance and features is arbitrary. One person’s performance characteristics can be another person’s features. Reliability — Reliability reflects the ability of a product to perform properly over a period of time. A car, for example, might perform without major repairs for 50,000 miles. Measures used to evaluate reliability are factors such as the mean time between failures, the mean time to first failure, and the failure rate per 1000 items. Conformance — Conformance measures whether product quality specifications have been met. Is a shaft the required diameter? Are the parts of impurity per million within the specified limits? Individual parts can be within tolerance; however, there can be a problem of tolerance stackup. Four parts, each 1.000 inch wide plus or minus .0005 inch, when stacked up will not be 4.000 inches tall plus or minus .0005 inch. Durability — Durability measures a product’s expected operating life. Product life can be limited due to technical failure (mechanical, electrical, hydraulic, pneumatic), technical obsolescence, or the economics of continued repair. For example, a light bulb has technical failure when the filament burns out. An automobile has economic failure when the owner decides it is no longer economically advantageous to repair it. Serviceability — Serviceability refers to the speed, ease, cost, certainty, and effectiveness of repair. Of critical concern are the courtesy of the repair people, the speed of getting the product back, and whether or not it is really fixed.

SL3151Ch03Frame Page 113 Thursday, September 12, 2002 6:12 PM

Benchmarking

113

Aesthetics — Aesthetics are concerned with the look, taste, feel, sound, and smell of an item. This can be critical for products such as food, paint, works of art, fashion, and decorative items. Perceived quality — Perceived quality is determined by factors such as image, advertising, brand identity, and word of mouth reputation. Stamatis, on the other hand, has introduced a modified version of the above points with some additional points — especially for service organizations. They are: Function — The primary required performance of the service Features — The expected performance (bells and whistles) of the service Conformance — The satisfaction based on requirements that have been set Reliability — The confidence of the service in relationship to time Serviceability — The ability to service if something goes wrong Aesthetics — The experience itself as it relates to the senses Perception — The reputation of the quality To be effective and efficient, the following characteristics must be present: • • • • • •

Be accessible Provide prompt personal attention Offer expertise Provide leading technology Depend — quite often — on subjective satisfaction Provide for cost effectiveness

What is interesting about these two lists is the fact that both Garvin and Stamatis recognize that design for optimum customer satisfaction is a design issue. Design, indeed, is the integrating factor. The designer has to make the tough trade-offs. Concurrent engineering and Quality Function Deployment suggest that the product designer, the manufacturing engineer, and the purchasing specialist work jointly during the product design phase to build quality in from the start. The focus, of course, is to design all the above characteristics as a bundle of utility for the customer. That bundle must address in holistic approach the following: Image Transcendent view — This view defines quality as that property that you will innately recognize as such once you have been exposed to it. Something about the product or service or the way it has been promoted/communicated to you causes you to recognize it as a quality offering — perhaps an excellent one. Performance Product-based view — This view defines quality in terms of a desirable attribute or series of attributes that a product contains. A high-quality fuel product could have a high BTU content and a low percentage of sulfur.

SL3151Ch03Frame Page 114 Thursday, September 12, 2002 6:12 PM

114

Six Sigma and Beyond

User-based view — This view defines quality in terms of how well a product or service meets the expectation of the customer. If the product meets expectations, it is considered to be of high quality. Expectations vary widely, and meeting expectations may not lead to the best product. For example, a bestseller may not be the best literature. Manufacturing-based view — This view defines quality in terms of conformance to manufacturing specifications. This view may, however, promote manufacturing efficiencies at the expense of suitability to the user. For example, problems of tolerance stackup are particularly noteworthy. Value Value-based view — This view, which is gaining in popularity, looks at value as the trade-off between quality and price. From this perspective, quality consists of all of the non-price reasons to buy a product or service. To come up with reasonable definitions and actions for the above characteristics, a team must be in place and team dynamics at work. A very good approach for this portion of benchmarking that we recommend is the nominal group process: The process features are as follows: Group size: five to nine core individuals Group composition: multidisciplinary and cross-functional Reflection — 20 minutes: all participants allowed to express their views as to what the problem is and how the team should progress Sharing of ideas: Discussion of the presented ideas Voting: evaluation of ideas and selection of the “best” Tabulation: Final resolution of what is at stake and how to proceed so that success will result The discussion and direction of the nominal process must not focus on price alone because that is a very narrow point of view. Some examples of non-price reasons to buy are: Product non-price reasons to buy • Ease of product use • Performance • Features • Reliability • Conformance • Durability • Serviceability • Aesthetics/style • Perceived quality • Ability to provide a bundled package Service and image non-price reasons to buy • Speed of delivery • Dependability of delivery

SL3151Ch03Frame Page 115 Thursday, September 12, 2002 6:12 PM

Benchmarking

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Fill rate Fun to deal with Number/location of stocking warehouses Repair facilities and location Technical assistance Service — before, during, and after sale Willingness to hold inventory Flexibility Access to salespeople Access to multiple supply sources Reputation Life cycle cost Financing terms Turnkey operations Consulting/training Warehousing Guarantees/warranty Services provided by salespeople Ease of resale Computer placement of orders Professional endorsement Packaging Up front engineering Vendor financial stability Confidence in salespeople Backup facilities Courtesy Credibility Understandability Responsiveness Accessibility of key players Flexibility Confidentiality Safety Delivery Ease of installation Ability to upgrade Customer training Provision of ancillary services Product briefing seminars Repair service and parts availability Warranty Image Brand recognition Atmosphere of facilities Sponsor of special events

115

SL3151Ch03Frame Page 116 Thursday, September 12, 2002 6:12 PM

116

Six Sigma and Beyond

The service and image features define the “augmented product.” They answer the questions: • What does your customer want in addition to the product itself? (the unspoken requirements) • What does your customer perceive to have value? • What does your customer view as “quality”? In order to focus benchmarking efforts, it is critical to define the unique selling proposition or the product concept. A statement of product concept requires the definition of both attribute(s) and benefit(s). Attributes consist of both form and features (specific product or service characteristics) and technology (how they are to be provided). For example, a new brewing technique brings a double-strength beer to add to your enjoyment by capturing the taste of the 1800s (technology, form, benefit). So what do you expect to get out of this team effort and integration? Simply put, you should get the answers to some very fundamental questions about your organization and the product/service you offer. Some typical questions are: • What are the non-price reasons to buy your product? How do they compare with the product and service attributes listed above? • How do your customers define quality? How does your company define quality? • What is more important? Product or service? • Can specific, measurable attributes be defined? • How does your competitor define quality? • How do you compare with your competitor? • What other companies or industries influence your customer as to what should be expected relative to each of these characteristics? • What does this suggest in the way of benchmarking opportunities? For example, here are some non-price reasons to buy that might apply to a supermarket: • • • • • • • • • • • •

Large parking lot Zoo in parking lot Lots of giveaways Makes shopping fun for the entire family Clowns Disneyland figures Well-stocked, attractive displays Rock hard containers of ice cream Complaint box (policy to respond the next day to the customer) Fast cash out “Forget your checkbook? Pay next time.” Trains all associates

SL3151Ch03Frame Page 117 Thursday, September 12, 2002 6:12 PM

Benchmarking

• • • • • • • •

117

Uses Dale Carnegie courses Walt Disney people management One aisle that rambles through the store No question return policy Bus for senior citizens Customer focus groups every three weeks Associates who take the initiative to please customers In-store dairy and bakery

None of these, in itself, is earth-shaking. But they could make the difference in an industry that operates with a profit margin of less than 1%. We cannot pass up the opportunity to address non-price issues for the WALMart corporation, which allegedly spends 1% of 100 details in the following items: • • • • • • • • • • • • • • • • • •

Aggressive hospitality People greeters Associates not employees Tough people to sell to Weekly top management meetings Low cost, no frills environment Good computerized database Rapid communications by phone Managers in the field Monday through Thursday High-efficiency distribution centers Emphasis on training of people Department managers having cost and margin data Profit sharing if store meets goals Bonus if shrinkage goal is met Open door policy Grass-roots meetings Constant improvements Competitive ads shown in store

DETERMINING

THE

CUSTOMER’S PERCEPTION

OF

QUALITY

Differentiation is uniqueness in the eyes of the customer. Quality is meeting the unique needs, wants, and expectations of the customer in terms of the non-price reasons to buy. But who is the customer? Depending on (a) defining the customer for multiple channels of distribution or (b) identifying the multiple buying influences in a business-to-business sale, the customer may be: • • • •

User Technical buyer Economic buyer Corporate general interest buyer

SL3151Ch03Frame Page 118 Thursday, September 12, 2002 6:12 PM

118

Six Sigma and Beyond

Who is the competitor? Assume for example a recreation environment. Here are some questions you might ask that would help you to determine who the competitors are: What • • What • • What • • What • •

is the desire I want to satisfy? (Desire competitor) Recreation Education kind of recreation do I enjoy? (Generic competitor) Baseball Boating kind of boating? (Form competitor) Power boat Sailboat brand boat? (Brand competitor) Bayliner Boston Whaler

Once these questions have been addressed, now we are ready to do the competitive evaluation in the following stages: Survey design • Attributes considered • Relative weight given to each • Direct competitors • Performance versus competition Approaches to making the survey Internal • Sales force • Sales management (Remember, the more accurate data you have, the better the survey. For example: Colgate Palmolive audits 75,000 customers for all products. “People know what they want and will not settle for happy mediums.”) External • Market research firms/universities • Attribution/non-attribution • Use of customer service hot line — GE progressed from receiving 1000 calls per week in 1982 to receiving 65,000 calls per week with the installation of an 800 number answer center. The 150 phone reps need a college degree and sales experience. They have been effective in spotting trends in complaints as well as increasing sales. The increase in sales has been estimated at more than twice the operating cost of the center. (Did this trigger off a benchmarking candidate for you?) Groups to be surveyed • Current customers • Lost customers • Prospects

SL3151Ch03Frame Page 119 Thursday, September 12, 2002 6:12 PM

Benchmarking

119

Survey frequency Comparison of company internal view versus the customer view

QUALITY, PRICING

AND

RETURN

ON INVESTMENT

(ROI) — THE PIMS RESULTS

Being perceived as being the best or having the product with the highest quality can have significant bottom line results. Buzzell and Gale (1987) introduced the PIMS (Profit Impact of Marketing Strategies) system, which is an elaborate benchmarking database developed by the Strategic Planning Institute in Cambridge, Mass. The database contains information for over 450 companies and over 3000 business experience pools in a wide variety of industries, including manufacturers, raw material producers, service companies, distributors, and durable and non-durable consumer products. Data are collected for independent business units, each with a defined served market. The objectives of the Strategic Planning Institute and benchmarking are to help organizations in the process of becoming excellent organizations. How do they do it? By: 1. Using the statistical analysis and modeling of business experience 2. Isolating the key factors that determine return on investment (ROI) ROI equals net income before interest and taxes divided by the total of working capital and fixed capital. As a result the Institute can help organizations with: • Understandability • Predictability of their own organization’s behavior and their own products and services. Of course, the choice of strategy depends upon several factors, including but not limited to: • • • • • • • •

Market growth rate and product life cycle Current market share Price/quality sensitivity by segment Competitive response profiles Current and planned capacity Cost and feasibility of quality improvements Market perception of quality improvements Financial and marketing goals — long and short term (The period described as “short term” and “long term” will differ widely among various strategies and organizations.)

BENCHMARKING AS A MANAGEMENT TOOL So far we have talked about benchmarking but we really have not defined it. A formal definition, then, is that benchmarking is a systematic, continual (ongoing)

SL3151Ch03Frame Page 120 Thursday, September 12, 2002 6:12 PM

120

Six Sigma and Beyond

management process used to improve products, services, or management processes by focusing on and analyzing the best of the best practices, by direct competitors or any other companies, to determine standards of performance and how to achieve those standards, to provide least cost, quality or differentiation, in the eye of the customer. Key words in this definition are systematic and ongoing, which imply that in order to have a successful benchmarking one must be familiar with the Kano model, Shewhart-Deming cycle, and principle of Kaizen improvement. This systematic and ongoing pursuit of excellence is applicable to all aspects of business and in all methodologies including the six sigma. It is an integral part of the strategic, operational, and quality planning process. It is not an end in itself. Benchmarking identifies the best of class and determines standards of excellence based on the market — considering both customers and competitors. It is a challenge with a solution. It provides the what and how. (A narrow focus on what you want to get done — a results orientation that controls performance with a carrot and a stick — is not effective without a broader focus on how best to do it — a process orientation that identifies the process changes that need to be made in order for the results to be achieved consistently.) Benchmarking is a creative imitation because part of its goal setting process that encourages the development of proactive plans is the action to bring about change. To do that, of course, analysis is required to determine all of the factors necessary for a solution to work, as appropriate and applicable to a given organization. In addition, it is necessary to project the future performance of the competition to set improvement goals. Otherwise, a company is always playing catch-up. Some of the key factors in this analysis are: • • • •

People/culture/compensation Process/procedure Facilities/systems Material

WHAT BENCHMARKING IS

AND IS

NOT

Benchmarking is not: • A way to cut costs or headcount, necessarily • A quick fix or a panacea • A cookbook approach Rather, it is a methodology that is an integral part of the management process and provides the organization with many benefits including but not limited to: • Identifying the specific action plans required to achieve success in company growth and profitability • Assessing objectively strengths and weaknesses versus competition and the best in class

SL3151Ch03Frame Page 121 Thursday, September 12, 2002 6:12 PM

Benchmarking

121

• Improving quality as perceived by the customer (The customer can be external to the company or the next department in the company.) • Determining goals objectively and realistically based on the actual achievements of others • Providing a vision of what can be accomplished in terms of both what and how • Providing hard, reliable data as a basis for performance improvement • Causing people to think creatively and to look at proactive alternative solutions to a problem • Promoting an opportunity for personal and corporate growth, learning, and development • Raising the company level of awareness of the outside world and of customer needs • Stimulating change — “Others are doing this, why can’t/shouldn’t we?” • Identifying all of the factors required to get a job done • Promoting an in-depth analysis and quantification of operations and management processes • Encouraging teamwork and communication within an organization • Creating an awareness of problems and stimulating change • Documenting the fact that a good job is being done and that you are the best in class • Allowing a company to leapfrog competition by looking outside of an industry • Changing the rules of the game by breaking with the traditions of an industry

THE BENCHMARKING PROCESS The benchmarking process can differ from company to company. However, the tenstep process below is generally followed. I. Benchmark planning and prioritization 1. Identification of benchmarking alternatives 2. Prioritization of the benchmarking alternatives II. Benchmark data collection 3. Identification of the benchmarking sources 4. Benchmarking performance and process analysis — company operations • What do we do? • What is the process? • What are the resource inputs? • What are the outputs? • What is the resource cost per unit of output? • What are the limitations? • What are possible changes? 5. Benchmarking performance and process analysis — partner’s operations

SL3151Ch03Frame Page 122 Thursday, September 12, 2002 6:12 PM

122

Six Sigma and Beyond

III. Benchmark implementation 6. Gap analysis 7. Goal setting 8. Action plan identification and implementation IV. Benchmark monitoring and control 9. Monitor company performance and action plan milestones 10. Identify the new “best in class”

TYPES

OF

BENCHMARKING

Benchmarking can be performed for any product, service, or process. Different classification schemes have been suggested. For example, Xerox classifies benchmarking in the following categories: • Internal benchmarking • Direct product competitor benchmarks • Functional benchmarking — This is a comparison with the best of the best even if from a different industry. • Generic benchmarking — This is an extension of functional benchmarking. It requires the ability to creatively imitate any process or activity to meet a specific need. For example, the technique used for high speed checking of paper currency (into the categories of good, mutilated, or counterfeit) by a bank could be adapted for high speed identification and sorting of packages in a warehouse. ATT, on the other hand, uses the classification indicated below. Specific examples of benchmarking studies for each are shown. These are not limited to ATT examples: Task • Invoicing • Order entry • Invoice design • Customer satisfaction • Supplier evaluation • Flow charting • Accounts payable Functional • Promotion by banks • Purchasing • Advertising by media type • Pricing strategy • Safety • Security Management process • PIMS par report • Profit margin/asset turnover

SL3151Ch03Frame Page 123 Thursday, September 12, 2002 6:12 PM

Benchmarking

123

• Strategic planning • Operational planning • Capital project approval process • Technology assessment • Research and development (R and D) project selection • Innovation • Training • Time-based competition • Benchmarking • Self-managed teams Operations process • Warehouse operations • Make versus buy Another classification of benchmarking projects is by: • • • •

Function — sales and marketing Process — missionary selling Activity — cold calling Task — preparation of target list

Still another classification is in terms of: • Overall financial performance • Department or functional benchmarking • Cost benchmarking

ORGANIZATION

FOR

BENCHMARKING

Ad hoc benchmarking studies can be helpful and productive. However, many companies are attempting to institutionalize benchmarking as part of the business planning and six sigma process. The business planning process consists of strategic planning followed by operational planning. Both phases require the development of functional area plans. However, the time periods considered, the alternatives of interest, and the level of detail are very different. The general flow of the planning process is: What should we do? • Situation analysis performed to determine critical success factors, strengths, weaknesses, opportunities, and threats • Mission development • Statement of objectives and goals How should we to do it? • Strategy determination • Tactics identified • Action plans specified

SL3151Ch03Frame Page 124 Thursday, September 12, 2002 6:12 PM

124

Six Sigma and Beyond

What are the expected results? • Budgets and financial projections How did we do? • Monitoring and control Who should get rewarded? • Performance evaluation and compensation Benchmarking is often an integral part of the situation analysis. It can also have a major impact on the mission statement, the goals, the strategy, the tactics, and the identification and determination of action plans. Benchmarking can provide major guidance when determining what to do, how to do it, and what can be expected. Benchmarking for strategic planning might concentrate on the determination of the critical success factors for an industry (based on customer and competitive inputs) and identifying what has to be done to be the success factors. This then leads to the development of detailed action plan with effort and result goals. Benchmarking for operational planning might concentrate on the cost and cost structure for each functional area relative to the outputs produced. All quality initiatives — including six sigma — have a significant influence on the mission statement and the objectives and goals of an organization. As such, they can provide an added impetus to do benchmarking to satisfy the quality goals. Benchmarking can be centralized (ATT) or decentralized (Xerox). Xerox has several functional area benchmarking specialists, including specialists for finance, administration, marketing, and manufacturing. The big advantage of a decentralized approach is a greater likelihood of organizational buying of the final results of the benchmarking study. The effort required to perform a benchmarking study can vary significantly. For example, the L.L. Beam study performed by Xerox took one person year of effort. Generally, three to six companies are included in the benchmark. However, some companies use only one or two. Also, some studies are performed in depth, while others are fairly casual. The “One Idea Club” was a simple approach with a substantial reward.

REQUIREMENTS

FOR

SUCCESS

All initiatives have requirements for success. Benchmarking is no different. Some of these requirements are: • Management vision and support to ensure the conditions necessary for the success of the strategy — people, money, time. • Goal-focused management with a customer/competitive focus on continuously improved quality • Performance- or results-based compensation • Action plan prioritization and focus • Defined roles and responsibilities for a multidisciplinary approach • Defined organizational approach — central versus decentralized • Integration with other management processes

SL3151Ch03Frame Page 125 Thursday, September 12, 2002 6:12 PM

Benchmarking

125

• Ability to maintain focus on the continuous improvement of hundreds of small items a little bit at a time • Willingness to deal with the conflict caused by a lack of goal congruity and the need to share scarce resources and to make tough decisions • Tolerance to deal with the ambiguity of results as research is conducted to determine when, where, and how to improve operations • Openness to learn and to change; results can affect organization structure, allocation of resources, corporate culture, and individual work assignments • Use of the scientific method: hypothesis formation, data collection, testing, and learning • Humility and the willingness to admit weakness and the possibility for improvement • Identification of the impediments to change and the development of a plan for change • Patience and resources to perform the analytical studies and to complete the required documentation • Long-term commitment to achieving results • Flexibility and discipline to implement the required changes • Communication of intent and approach, findings, concerns, and apprehensions • Training and total employee involvement, empowerment, and teamwork • A process that starts slow, showcases, and picks up speed as experience and confidence are gained • Market segmentation focus and a defined corporate strategy It sounds good. But does benchmarking work? Let us see what the SAS Airlines did, as an example. When Jan Carlzon took over as president of Scandinavian Airlines (SAS) in 1980, the company was losing money. For several previous years, management had dealt with this problem by cutting costs. After all, this was a commodity business. Carlzon saw this as the wrong solution. In his view, the company needed to find new ways to compete and build its revenue. SAS had been pursuing all travelers with no focus on superior advantage to offer to anyone. In fact, it was seen as one of the least punctual carriers in Europe. Competition had increased so much that Carlzon had to figure out: • Who are the customers? • What are their needs? • What must we do to win their preference? Carlzon decided that the answer was to focus SAS’s services on frequently flying business people and their needs. He recognized that other airlines were thinking the same way. They were offering business class and free drinks and other amenities. SAS had to find a way to do this better if it was to be the preferred airline of the frequent business traveler.

SL3151Ch03Frame Page 126 Thursday, September 12, 2002 6:12 PM

126

Six Sigma and Beyond

The starting point was market research to find out what frequent business travelers wanted and expected in the way of airline service. Carlzon’s goal was to be one percent better in 100 details rather than 100 percent better in only one detail. The market research showed that the number one priority was on-time arrival. Business travelers also wanted to check in fast and be able to retrieve their luggage fast. Carlzon appointed dozens of task forces to come up with ideas for improving these and other services. They came back with ideas for hundreds of projects, of which 150 were selected with an implementation cost of $40 million. One of the key projects was to train a total customer orientation into all of SAS’s personnel. Carlzon figured that the average passenger came into contact with five SAS employees on an average trip. Each interaction created a “moment of truth” about SAS. At that point of contact, the person was SAS. Given the 5 million passengers per year flying SAS, this amounted to 25 million moments of truth where the company either satisfied or dissatisfied its customer. To create the right attitudes toward customers within the company, SAS sent 10,000 front line staff to service seminars for two days and 25,000 managers to three-week courses. Carlzon taught many of these courses himself. A major emphasis was getting people to value their own self-worth so that they could, in turn, treat the customer with respect and dignity. Every person was there to serve the customer or to serve someone who was serving the customer. The results: Within four months, SAS achieved the record as the most punctual airline system in Europe, and it has maintained this record. Check-in systems are much faster, and they include a service where travelers who are staying at SAS hotels can have their luggage sent directly to the airport for loading on the plane. SAS does a much faster job of unloading luggage after landings as well. Another innovation is that SAS sells all tickets as business class unless the traveler wants economy class. The company’s improved reputation among business flyers led to an increase in its full fare traffic in Europe of 8 percent and its full fare intercontinental travel of 16 percent, quite an accomplishment considering the price cutting that was taking place and zero growth in the air travel market. Within a two-year period, the company became a profitable operation. Carlzon’s impact on SAS illustrates the customer satisfaction and profits that a corporate leader can achieve by creating a vision and target for the company that excites and gets all the personnel to swim in the same direction — namely, toward satisfying the target customers. As a leader, Carlzon created the conditions necessary to ensure the success of the strategy by implementing the projects required for the front line people to do their jobs well.

BENCHMARKING AND CHANGE MANAGEMENT Several behavioral models underscore the psychological requirements for change in a person or an organization. The classic equation for change is: D×V×F>R

SL3151Ch03Frame Page 127 Thursday, September 12, 2002 6:12 PM

Benchmarking

127

where D = dissatisfaction with current situation; V = vision of a better future; F = the first steps of a plan to convert D to V; and R = resistance to change. Typical attitudes/comments of resistance are: • • • • • • •

Perceived threat of loss — power, position Everything is OK. Why fix it? What should we change? How? What is management trying to tell me? Takes a long time to see results! We do not have time to do that “stuff.” If this is so good, why aren’t they doing it?

Benchmarking can accelerate the change process by offering the organization’s managers facts that relate to their needs and expectations, by understanding the psychology of change. For example, while the previous mathematical formula is a quantifiable entity on its own, it gives us little opportunity to explore change from the individual’s perspective. Change begins with an individual. That individual must: 1. Believe that he or she has the skill necessitated by the change. Can I do it? 2. Perceive a reasonable likelihood of personal value fulfillment as a result of making the change. What will I get out of it? 3. Perceive that the total personal cost of making the change is more than offset by the expectation of personal gain. Is it worth making the change? This model suggests that we manage change by education and communication to influence what a person thinks and that this, in turn, causes a change in behavior. Thought is affected by: • • • •

Beliefs Facts Values Feelings

Benchmarking can help implement change by providing the required facts and challenging beliefs, especially if there are data to be supported from other organizations. Other models to manage change are: • • • • •

Facilitation and support Participation and involvement Negotiation Manipulation Explicit and implicit coercion

Corporate culture is important:

SL3151Ch03Frame Page 128 Thursday, September 12, 2002 6:12 PM

128

Six Sigma and Beyond

• Reward risk taking • Encourage passionate champions • Focus on base hits versus home runs Sources of dissatisfaction that can drive change include: • Financial pressure • Quarterly earnings • Cash flow (Need: to improve operational efficiency)

STRUCTURAL PRESSURE • • • •

Cyclical business mix Customer mix Cash flow conflicts Product life cycle mix (Need: To improve business mix or effectiveness)

ASPIRATION

FOR

EXCELLENCE

The need to improve is an internal perception. “You do not have to be bad to improve.” Organization positions can be viewed as having either innovation and/or maintenance responsibilities relative to change. How does the mix change for workers, supervisors, middle management, and top management in an organization that strives for excellence? Current success can mask underlying problems and can prevent or delay action from taking place when it should, i.e., when the company has the time and resources to do something. Consider the classic story of the “boiled frog” as an example. (If you recall, the frog was boiled when the temperature was increasing at a very slow rate. The frog was adapting. The frog could not differentiate the change and ultimately, was boiled. On the other hand, the frog that was thrown into hot water jumped out right away and saved its life.)

FORCE FIELD ANALYSIS Force field analysis is a systematic way of identifying and portraying the forces (often people) for or against change in an organization. The specific forces will differ depending upon the area where benchmarking is applied. Here is how the process works: 1. 2. 3. 4. 5. 6. 7.

Define the current situation. Define the desired position based on the results of the benchmarking study. Define the worst possible situation. What are the forces for change? What is their relative strength? What are the forces against change? What is their relative strength? What forces can you influence? Define the specific action to be taken relative to each of those forces that you can influence.

SL3151Ch03Frame Page 129 Thursday, September 12, 2002 6:12 PM

Benchmarking

129

One effective way to start the benchmarking process is to select one high visibility area of concern to the influence leaders in a company and produce results that can showcase the benchmarking process. This might start with a library search to highlight the results that are possible.

IDENTIFICATION OF BENCHMARKING ALTERNATIVES As indicated earlier, benchmarking candidates can be identified in a wide variety of ways. They can be detected, for example, during the business planning process, as part of a quality initiative, during a six sigma project, or during a profit improvement campaign. Both external and internal analysis can lead to potential candidates.

EXTERNALLY IDENTIFIED BENCHMARKING CANDIDATES Industry Analysis and Critical Success Factors Based on the structure of an industry and the dynamics of the customer/supplier interface, certain factors are critical to the success of a business. An identification of the critical success factors and an evaluation of the company’s current capabilities can lead to benchmarking opportunities. The competitive rivalry among firms in an industry has a significant impact on total demand and the level and stability of prices. Competitive rivalry is a function of several interrelated factors that affect the supply and demand for products and services. The balance of supply and demand at a particular time affects the percent capacity utilization in an industry. The percent capacity utilization directly affects price levels and price elasticity. The factors affecting demand are: • The strategy of the buyer to be least cost or differentiated — How well do you meet your customers’ specific needs? • The availability of and knowledge about substitute products — What are existing and new competitors likely to do? • The ease of switching from one product to another — How can you increase the cost of switching to another supplier? • Governmental regulations — What can you do to influence these? The factors affecting supply are: • The ease of market entry — What can you do that will make it hard to enter the business? • The barriers to market exit — What can you do to make it easy for a competitor to get out of the business? • Governmental regulations — What can you do to influence these?

SL3151Ch03Frame Page 130 Thursday, September 12, 2002 6:12 PM

130

Six Sigma and Beyond

Based on an analysis of the industry as it exists now and might exist in the future, what are the factors absolutely critical for success? Five or six critical success factors can usually be identified for a company. Examples are: • • • • • • • • • • • • • • • • • • •

Customer service Distribution Technically superior product Styling Location Product mix Cost control Dealer system Product availability Supply source Production engineering Advertising and promotion Packaging Staff/skill availability Quality Convenience Personal attention Innovation Capital

Once the critical success factors have been identified, the company can assess its current position to determine whether benchmarking is required. One technique for performing this analysis is to make a tabulation showing how the major competitors in an industry rank for each critical success factor. As a cross check, there should be a correlation between the tabulated results, market share and return on equity. PIMS Par Report The PIMS par report indicates the financial results that companies in similar circumstances have been able to achieve. As such, it provides a quantitative benchmark. The PIMS report also indicates those factors that should enable you to earn greater than par and those factors that would cause you to earn less than par. Financial Comparison If PIMS data are not available, a comparison of the company’s financial performance versus that of other companies in the same industry can suggest the value of benchmarking in specific areas. Potential areas that might be identified are: • Gross margin improvement • Overhead cost reduction

SL3151Ch03Frame Page 131 Thursday, September 12, 2002 6:12 PM

Benchmarking

• • • • •

131

Fixed asset utilization Inventory or accounts receivable reduction Liquidity improvement Financial leverage Sales growth

Competitive Evaluations As discussed earlier, a competitive evaluation is a periodic assessment made to determine, objectively, what factors a buyer takes into consideration when deciding to buy from one supplier versus another, the relative weight given to each of those factors, the competing firms, and the relative performance of each firm with respect to each buying motive. Focus Groups Focus groups are used to determine what a customer segment thinks about a product or service and why it thinks that way. Participants are invited to join the group usually with some type of personal compensation. A focus group starts with a series of open-ended questions relative to a specific subject. Representatives of the sponsoring company view the entire process either through a one-way mirror or by closed circuit TV. As a second phase, the company representatives ask specific follow-up questions (through the facilitator), based on the open-ended probing. Importance/Performance Analysis The customer perception of performance versus importance can be used to identify benchmark alternatives. A list of attributes can be prepared using either a nominal group process or a focus group. The customer is asked to rate each attribute in terms of both importance and company performance. A matrix is then prepared showing high and low performance versus high and low importance. It has the following implications: • • • •

Continue with high importance, high performance. Reduce emphasis on low importance, low performance. Increase emphasis on high importance, low performance. Reduce emphasis on low importance, low performance.

In addition to determining the customer’s perception of performance versus importance, it is also valuable to determine the customer’s versus the company’s perception of importance. This can also be used to determine areas for intensification and reduction of effort and benchmarking possibilities, including: 1. Customer-oriented goals 2. Service/quality goals

SL3151Ch03Frame Page 132 Thursday, September 12, 2002 6:12 PM

132

Six Sigma and Beyond

INTERNALLY IDENTIFIED BENCHMARKING CANDIDATES — INTERNAL ASSESSMENT SURVEYS An internal assessment of strengths and weaknesses can be used to identify benchmarking candidates. This determination can be made using a Business Assessment Form followed by group discussion or by using the Nominal Group Process. The assessment can be made by the owner of a product, process, or service and/or by the department being served. An internal assessment can also be approached from the viewpoint of the generic value added chain. For each block of the chain, two questions can be asked: 1. What are the alternatives for least cost operation? 2. What are the alternatives to provide differentiation? The value added chain can also provide customer perspective by suggesting the questions: 1. How does our product or service help customers to minimize their cost? 2. How does our product or service help customers to differentiate their product? Nominal Group Process: General Areas in Greatest Need of Improvement • • • • • • •

Improving the precision of the sales forecast Reducing the cycle time to bring out new products Increasing the success rate in bidding for new business Reducing the time required to fill customer orders Reducing the errors in invoices Major problems or issues Areas of competitive disadvantage

Pareto Analysis Pareto analysis is a form of data analysis that requires that each element or possible cause of a problem be identified along with its frequency of occurrence. Items are then displayed in order of decreasing frequency of probability of occurrence. This can help to identify the most significant problem to attack first. A common expression of the Pareto Law is the 80/20 rule, which states that 20% of the problem causes 80% of the difficulties. A Pareto analysis of setup delay might include factors such as: necessary material not available, tooling not ready, lack of gages, setup personnel not available, another setup has priority, material handling equipment not available, error — incorrect setup. Develop a Pareto analysis for the production of scrap. (There is a tremendous difference between knowing the facts and guessing).

SL3151Ch03Frame Page 133 Thursday, September 12, 2002 6:12 PM

Benchmarking

133

Statistical Process Control Statistical process control is a technique for identifying random (or common) causes versus identifiable (or special) causes in a process. Both of these are potential sources for improvement. The amount of random variation affects the capability of a machine to produce within a desired range of dimensions. Hence, benchmarking could be performed to determine machine processing capabilities and how to achieve those levels. The determination and correction of recurring systematic changes is also a benchmarking possibility. The reduction of the random variation or the uncertainty of the process and the identification and correction of special causes are critical aspects of the total quality management process. Correction often requires a change in the total manufacturing process, tooling, the equipment being used, and/or training in setup and operations. The first step in process improvement is to control the environment and the components of the system so that variations are within natural, predictable limits. The second step is to reduce the underlying variation of the process. Both of these are candidates for benchmarking. Trend Charting Historic data can be used to develop statistical forecasts and confidence intervals that depict acceptable random variation. When data fall within the confidence intervals, you have no cause to suspect unusual behavior. However, data outside of the confidence intervals could provide an opportunity for benchmarking. It might also be informative to pursue benchmarking as a device to reduce the range of variation or the size of the confidence interval. A simple trend analysis of your own past data can also provide a basis for improvement. The following data relative to the percent scrap and rework illustrate the improvement made and could provide the basis for benchmarking: 1987

1988

1989

1990

2.1%

3.0%

1.0%

0.7%

Product and Company Life Cycle Position Products tend to go through a defined life cycle starting with an introductory phase and proceeding through growth, maturity, and decline. The management style and business tactics are very different at each stage. Anticipating and managing the transitions can be important. This could lead to opportunities for benchmarking of product life cycle management and product portfolio management. Product portfolio management can lead to the need for new product identification and introduction. These areas have both been the subjects of benchmarking studies. In addition to the changes that products go through, companies tend to go through various stages of development and crises. Again, the management of the transitions can be an important benchmarking candidate.

SL3151Ch03Frame Page 134 Thursday, September 12, 2002 6:12 PM

134

Six Sigma and Beyond

Failure Mode and Effect Analysis Failure Mode and Effect Analysis (FMEA) is a systematic way to study the operating environment for equipment or products and to determine and characterize the ways in which a product can fail. Benchmarking can be used to determine component and system design goals and alternatives (see Chapter 6). Cost/Time Analysis To evaluate its new product introduction process, a company may plot cost per unit produced versus elapsed time for each element of the process, e.g., design and engineering, production, sub-assembly, and assembly. The area under the curve represents money tied up (inventory), and smaller is better.

NEED

TO IDENTIFY

UNDERLYING CAUSES

Problem, Causes, Solutions When solving a problem, it is critical to attack the underlying cause of the problem and not the symptoms. The underlying cause can be identified by listing all possible causes and identifying the most probable cause based on data collection and a Pareto analysis. This sometimes leads to multiple benchmarking opportunities. Failure to diagnose a problem (ready, fire, aim) can lead to an inefficient use of resources and frustration. The Five Whys When identifying underlying causes, it can also be useful to ask five sequential “whys” to get to the heart of a problem. For example: Problem: The milling machine is down. Why? The chucking mechanism is broken. Why? A piece got jammed when being loaded. Why? There was excess flash from the stamping operation. Why? The stamping die was not changed. Why? The die usage control log was not updated daily. Cause and Effect Diagram The development of a Cause and Effect Diagram or Fishbone Diagram or Ishikawa Diagram is another way to identify and display the underlying causes of a problem. Causes are usually displayed in terms of major categories such as human or personnel, machines or technology, materials and methods or procedures. Once causes are identified, an analysis is made to determine actionable solutions. The determination of cause and effect can require the use of designed experiments to measure effects and interaction. For example, to reduce conveyor belt spillage, it was necessary to determine the effects of belt wipers, belt surface, dryness of the belt and material, and particle size in various combinations of each factor.

SL3151Ch03Frame Page 135 Thursday, September 12, 2002 6:12 PM

Benchmarking

135

BUSINESS ASSESSMENT — STRENGTHS AND WEAKNESSES You will be asked to evaluate the organization relative to sales and marketing, manufacturing and operations, R & D, and general management. A typical assessment is shown in Table 3.1.

TABLE 3.1 A Typical Assessment Instrument Please indicate how you evaluate the organization using the following key: (There are many ways to use a key. This is only one example.) ++ + E – •

Extremely strong, definite leaders Better than average Average Weak, should do better Extremely weak, area of major concern Sales and marketing Customer base Market share Market research Customer knowledge Brand loyalty Company business image Response to customers Breadth of product line Product differentiation Product quality Distributors Locations Size Warehousing Transportation Communication Influencing customers Sales force People and skills Size Type Location Productivity Morale Advertising

SL3151Ch03Frame Page 136 Thursday, September 12, 2002 6:12 PM

136

TABLE 3.1 (Continued) A Typical Assessment Instrument National regional cooperative Promotion devices Prices/incentives Customer communication Service Before sale After sale Credit Long term Short term Trade allies Costs Selling Distribution Manufacturing/operations Materials management People and skills Sourcing Inventory P & C Production P & C Capability P & C Computer system Physical plant Capacity Utilization Flexibility Plant Size location Number Age Equipment Automation Maintenance Flexibility Processes Uniqueness Flexibility Degree of integration Engineering Process Tool design Cost improve

Six Sigma and Beyond

SL3151Ch03Frame Page 137 Thursday, September 12, 2002 6:12 PM

Benchmarking

TABLE 3.1 (Continued) A Typical Assessment Instrument Time standards Quality control People and skills Workforce Skills mix Utilization Availability Turnover Safety Unionization Costs Productivity Morale Direct/indirect Research and development Basic research Concepts and studies Emphasis People and skills Conversions to applications Patents Applied research Finding Emphasis People and skills Conversion to prototype Patents Basic engineering Prototypes Emphasis People and skills Convert to product Design engineering Designs Patents and copyrights Emphasis People and skills Design for production Funding Amount Consistency Sources Project selection

137

SL3151Ch03Frame Page 138 Thursday, September 12, 2002 6:12 PM

138

TABLE 3.1 (Continued) A Typical Assessment Instrument General management Leadership Vision Risk/return profile Clarity of purpose Implementation skills Turnover Experience Motivation skills Leadership style Delegation Strategic emphasis Organization Type Size Location Communication Defined responsibility Coordination Speed of reaction Fix with strategy Commitment Planning and control Early alert system Forecasting Operational budget Control MBO program Capital planning Long range planning Contingency planning Cost analysis Resource allocation Accounting and finance Financial public relations Financial relations Auditing Decision making Style Techniques used Responsiveness Position in org. criteria used

Six Sigma and Beyond

SL3151Ch03Frame Page 139 Thursday, September 12, 2002 6:12 PM

Benchmarking

139

TABLE 3.1 (Continued) A Typical Assessment Instrument Personnel Effectiveness Hourly labor Clerical labor Sales people Scientists and engineers Supervisors Middle management Top management Comp. and reward Management development Management depth Turnover Morale Information systems Decision support system Customer data Product line data Fixed/variable costs Exception reporting Culture Shared values Pluralism Conflict resolution Openness Optional Information Name: Date: Title: Dept:

PRIORITIZATION OF BENCHMARKING ALTERNATIVES — PRIORITIZATION PROCESS A variety of prioritization approaches are available. Use the one most appropriate to a specific situation.

PRIORITIZATION MATRIX The following steps are required to complete a prioritization matrix:

SL3151Ch03Frame Page 140 Thursday, September 12, 2002 6:12 PM

140

Six Sigma and Beyond

1. 2. 3. 4. 5.

List all items to be prioritized. List the goals or the prioritization criteria. Specify the goal weights. Indicate the impact score of each item relative to each goal. Determine the value index for each item by totaling the cross product of each goal weight times the impact score. 6. Sort the items from highest to lowest value index.

QUALITY FUNCTION DEPLOYMENT (HOUSE

OF

QUALITY)

Quality function deployment is an extension of the prioritization matrix described above. However, the rows and the columns are interchanged. The rows become the evaluation criteria (or goals) and the columns represent the alternative solutions to be prioritized. The following procedure is used to complete the Quality Function Deployment analysis: 1. List the items indicating “what” you want to accomplish. These are the evaluation criteria. 2. List “how” you will accomplish what you want to do. These are the alternatives to be evaluated. 3. Indicate the degree of importance for each of the “whats.” This is a number ranging from 1 to 10 (10 is most important). 4. Indicate the company and the competitive rating using a scale from 1 to 10 (10 is best). Plot the competitive comparison. 5. Specify the planned or desired future rating. 6. Calculate the improvement ratio by dividing the planned rating by the company current rating. 7. Select at most four items to indicate as “sales points.” Use a factor of 1.5 for major sales points and a factor of 1.2 for minor sales points. 8. Calculate the importance rate as the degree of importance times the improvement ratio times the sales points. 9. Calculate the relative weight for each item by dividing its importance rate by the total of the importance rates for all “whats.” 10. Indicate the relationship value between each “what” and “how.” Use values of 9, 3, and 1 to indicate a strong, moderate or light interrelationship. 11. Calculate the importance weight for each “how.” This is the total of the cross products of the relationship value and the relative weight of the “what.” 12. Indicate the technical difficulty associated with the “how.” Use a scale of 5 to 1 (5 is the most difficult). 13. Indicate the company, competitive values, and benchmark values for the “how”. 14. Specify the plan for each of the “hows.” Quality function deployment is usually applied at four different interrelated levels:

SL3151Ch03Frame Page 141 Thursday, September 12, 2002 6:12 PM

Benchmarking

141

1. Product planning What — customer requirements How — product technical requirements 2. Product design What — product technical requirements How — part characteristics 3. Process planning What — part characteristics How — process characteristics 4. Production planning What — process characteristics How — process control methods

IMPORTANCE/FEASIBILITY MATRIX Importance is a function of urgency and potential impact on corporate goals. It is expressed in terms of high, medium, and low. Feasibility takes into consideration technical requirements, resources, and the cultural and political climate. It is also expressed in terms of high, medium, and low. Paired Comparisons This approach is based on a pair-by-pair comparison of each set of alternatives to determine the most important. Count the total number of times each alternative was selected to determine the overall prioritization. Improvement Potential To determine how to prioritize cost improvement benchmarking alternatives, perform the following analysis: 1. Make a Pareto analysis of cost components 2. Assess the percent improvement possible for each of the most significant cost components. 3. Multiply the cost times the percent improvements possible to determine the improvement potential. 4. Prioritize the benchmark studies based on improvement potential. This approach can be used to prioritize other areas as well. Prioritization Factors When prioritizing benchmarking candidates, it is important to take into consideration many factors. Some of these factors are listed below. It is important to narrow projects down to the significant few and to choose a good starting project to showcase the value of the approach.

SL3151Ch03Frame Page 142 Thursday, September 12, 2002 6:12 PM

142

Six Sigma and Beyond

The first project should be a winner. It should address a chronic problem, there should be a high likelihood of completion in several weeks, and the results should be (a) correlated to customer needs and wants, (b) significant to the company, and (c) measurable. Factors to be used subsequently are: • • • • • • • • • • • •

• • • • • •

Importance of business need long term Basis for a sustainable competitive advantage Percent improvement possible Customer impact Realism of expectations Urgency Ease of implementation/degree of difficulty Time to implement Consistency with mission, values, and culture Organizational buy in Passionate champion identified Resource requirements and availability • Capital expenditures • Working capital • Time by skill category Synergy Risk versus return Measurability of result Modularity of approach Anticipated problems Potential resistance

ARE THERE ANY OTHER PROBLEMS? WHAT IS OF EACH OF THESE?

THE

RELATIVE IMPORTANCE

The Japanese approach to improvement is called “Kaizen.” This philosophy espouses an innovative, small-step-at-a-time approach that is implemented by creating an awareness of need and empowerment throughout an organization. This contrasts to the Western approach, which tends to be higher tech, capital intense, and focused on major innovative changes. (Several studies have demonstrated that the U.S. is much better at discovery and invention than Japan, but that we lag in commercial development and implementation of the ideas.) Could the low tech, people-oriented focus work in your competitive situation? What does this suggest in terms of benchmarking prioritization?

IDENTIFICATION OF BENCHMARKING SOURCES TYPES

OF

BENCHMARK SOURCES

The benchmarking process often starts with a library search to identify alternative views, issues, approaches, and possible benchmarking sources. Benchmarking

SL3151Ch03Frame Page 143 Thursday, September 12, 2002 6:12 PM

Benchmarking

143

sources can be internal best performers, competitive best performers, or best in class worldwide. Internal Best Performers Xerox used internal benchmarking when it studied Fuji-Xerox’s manufacturing methods (but not until Florida Power and Light began to emulate them). Different divisions, plants, distribution outlets, and departments tend to do things differently. Much can often be learned by looking at these company operations. Competitive Best Performers The advantage of making comparisons with direct competitors is obvious. However, it can be difficult to get competitors to share their source of competitive advantage. When working with direct competitors, it can also be difficult to get out of the industry mind-set and come up with creative ideas. It could be that the competitors in an industry are not particularly good at what they do and hence provide little stimulus for improvement. Xerox regularly benchmarks all direct competitors, all their suppliers, and all major competitors to those suppliers. Updates are important. Knowing how fast competitors are moving is just as important as knowing where they are. Best of Class There is, in general, no way to know the “best” of the best. Companies generally pick the “best” based on reputation through publications, speeches, news releases, etc. A company might start out with four to ten “best” candidates and narrow them down based on initial discussions. Xerox looked at IBM and Kodak but also L.L. Bean, the catalog sales company, known for effective and efficient warehousing and distribution of products. Additional benchmarking partners used by Xerox were: Customer satisfaction, customer retention Financial stability and growth SPC and quality Customer care and training

USAA (Insurance Co.) A.G. Edwards & Sons Florida P&L Walt Disney

Milliken & Company, winner of the 1989 National Quality Award, provided the following partial list of benchmarks: Strategy Safety Customer satisfaction Innovation Education Strategic planning Time based competition

DuPont ATT, IBM 3M, KBC IBM, Motorola Frito-Lay, IBM, ATT Lenscrafters

SL3151Ch03Frame Page 144 Thursday, September 12, 2002 6:12 PM

144

Six Sigma and Beyond

Benchmarking Self-managed teams Continuous improvement Heroic goals concept Role model evaluation Environmental practice Statistical methods Flow charting Quality process

Security Accounts payable Order handling

Quality Process Xerox Goodyear, P&G Japanese Motorola Xerox DuPont, Mobay, Ciba-Geigy Motorola Sara Lee FP&L, Westinghouse, Motorola Miscellaneous DuPont Mobay L.L. Bean

SELECTION CRITERIA How do you know who is the best? Here are some ways to get that information: • • • •

Library search Reputation Consultants Networking

Characteristics to be examined when seeking partners include: • • • • • • •

Company size Customer non-price reasons to buy Industry critical success factors Availability of data Data collection costs Innovation Receptivity

One hundred percent accuracy of information is not required. You only need enough to head you in the right direction.

SOURCES

OF

COMPETITIVE INFORMATION

Read everything and ask, “Has anyone faced this or a similar problem? What have they done?” Do not forget to ask people in your own organization, including: • • • •

Past employees of benchmark company Family members Market researchers Sales and marketing

SL3151Ch03Frame Page 145 Thursday, September 12, 2002 6:12 PM

Benchmarking

145

It is also helpful to make use of trade associations and consultants and to network. Review studies in which people have identified the characteristics of best performers. Good sources here are Clifford and Cavanagh (1988), Smith and Brown (1986), and Berle (1991). Another good source is the Encyclopedia of Business Information Sources, published frequently by Gale Research, Detroit, Michigan. This source contains references by subject to the following: • • • • • • • • • • • • •

Abstracting and indexing services Almanacs and yearbooks Bibliography Biographical sources Directories Encyclopedias Financial ratios Handbooks and manuals Online databases Periodicals and newsletters Research centers and institutes Trade associations/professional associations Other

Additional sources may also be found in the John Wiley publication entitled Where to Find Business Information, as well as the following: Books and periodicals • Trade journals • Functional journals • F.W. Dodge reports • Technical abstracts • Local newspapers, national newspapers • Nielson — Market Share • Yellow Pages • Textbooks • Special interest books • City, region, state business reviews • Standard and Poors industry surveys Directories • Trade show directory • Directory of Associations • Brands and Their Companies • Who Runs the Corporate 1000 • Corporate Technology Directory • American Firms in Foreign Countries • Corporate Affiliations • Foreign Manufacturers in U.S. • Directory of Company Histories

SL3151Ch03Frame Page 146 Thursday, September 12, 2002 6:12 PM

146

Six Sigma and Beyond

• International Trade Names • Leading Private Companies • Marketing Economics Key Plants • Directory of Advertisers • Books of business lists • Thomas Register • Wards Directory • Lists of 9 Million Businesses — ABI Computer databases — CD-ROM or online Text databases • Business dateline — articles • BusinessWire — press releases • Intelligence Tracking Service — consumer trends • Dow Jones Business and Financial Report • Newsearch • Trade and Industry Index Statistical business information • BusinessLine • Cendata • Consumer Spending Forecast • Disclosure Database • CompuServe • Retail Scan Data • Moody’s 5000 Plus Demographic data • Census Projection 1989–1993 • Donnelley demographics Directories • Dun’s Million Dollar Directory • Thomas Register Company direct • Advertising • Benchmarking partner • Company newsletters • Minority interest partners • Speeches • Direct contact Financial sources • Annual reports, 10k, proxies, 13D • Investment reports • Prospectus • Filings with regulatory agencies • Dun and Bradstreet, Robert Morris • Moody’s Manuals • S&P Reports

SL3151Ch03Frame Page 147 Thursday, September 12, 2002 6:12 PM

Benchmarking

Individuals • Company employees • Past employees/retirees • Social events • Construction contractors • Landlords, leasing agents • Salesmen • Service personnel • Focus groups Professional societies • Professional society members • Trade shows/conventions • National associations • User groups • Seminars • Rating services • Newsletters Government • Public bid openings • Proposals • National Technical Information Center • Freedom of Information Act • Occupational Safety and Health Administration (OSHA) filings • Environmental Protection Agency (EPA) filings • Commerce Business Daily • Government Printing Office Index • Federal depository libraries • Court records • Bank filings • Chamber of Commerce • Government Industrial Program reviews • Uniform Commerce Code filings • State corporate filings • County courthouse • U.S. Department of Commerce • Federal Reserve banks • Legislative summaries • The Federal Database Finder • Patents Customers • New customers • Consumer groups Industry members • Suppliers • Equipment manufacturers

147

SL3151Ch03Frame Page 148 Thursday, September 12, 2002 6:12 PM

148

Six Sigma and Beyond

• Distributors • Buying groups • Testing firms Snooping • Reverse engineering • Hire past employees • Interview current employees • Dummy purchases • Shopping • Request a proposal • Hire to do one job • Apply for a job • Mole • Site inspections • Trash • Chatting in bars • Surveillance equipment Schools and universities • Directories of case studies • Industry studies Consultants • Business schools on a consulting basis • Jointly sponsored studies • Information brokers • Industry studies • Market research studies • Seminars

GAINING

THE

COOPERATION

OF THE

BENCHMARK PARTNER

Without confidentiality, benchmarking will not work. Some items for consideration in gaining this confidentiality and cooperation are: • Use consultants or trade associations or universities to ensure confidentiality. • Make sure that there is mutual sharing — could be different areas. • Be prepared to give and receive. • Focus on mutual learning and self improvement. • Benefit of probing questions and debate • Opening up of a vision • Confirmation of good practice • Consider benchmarking a circuit of companies. • Important that all know in advance • Consider all security and legal implications of sharing data.

SL3151Ch03Frame Page 149 Thursday, September 12, 2002 6:12 PM

Benchmarking

MAKING

149

CONTACT

THE

When making the contact for benchmarking, follow these steps: • Call and express interest in meeting. • Send/receive a detailed list of questions. • Make sure that you have prepared your questions carefully. (The quality of the questions can be the signal for a worthwhile use of time.) • Follow up by telephone. • Visit — keep an open mind — document everything

BENCHMARKING — PERFORMANCE AND PROCESS ANALYSIS PREPARATION

OF THE

BENCHMARKING

PROPOSAL

Factors to be considered in the preparation of the benchmarking proposal include: • • • • • • • • • • •

Mission Objective/scope Statement of importance Information available Critical questions Ethical and legal issues Partner selection Roles and responsibilities Visit schedule Data analysis requirements Form of recommendation

ACTIVITY

BEFORE THE

VISIT

The approach that follows is very comprehensive. It might not be economical to follow all the steps in every study. Let practical common sense be the guide to action. Understanding Your Own Operations You need to understand your own operations very thoroughly before comparing them with the operations of others. Here are some steps you should take to make sure that you understand your current methods: Ask open-ended questions. For example, for “who”: • • • •

Who Who Who Who

does it per the job description? is actually doing it? else could do it? should be doing it?

SL3151Ch03Frame Page 150 Thursday, September 12, 2002 6:12 PM

150

Six Sigma and Beyond

Ask similar questions for what, where, when, why, and how. Activity Analysis Activity analysis consists of the following steps: 1. Define the Activity Activities can be defined through: • • • •

Function Process Marketing and sales Sell products

Activity: These are the major action steps of a process. For example, make a proposal. Task: Prepare proposal draft Operation: Type proposal

2. Determine the Triggering Event Identify what happens to trigger the activity. Why does the activity get performed at a specific time? What is the status of material or information before the activity occurs? What documentation signifies that the activity is to start? Example: Receive material Material receipt document

3. Define the Activity Document how to perform the activity. Indicate what has to be done and the order in which it is done. This will define all business procedures, policies, and controls. Questions to ask include: • • • • •

What What What What What

are the key process variables? controls these variables? levels lead to optimum performance? are the causes of variation? are the limitations?

Activities should be classified in terms of repetitive versus non-repetitive, primary versus secondary, and required versus discretionary. It is important in this analysis to determine limitations, sources of error, rejects, and delays. Example: Raw material Inspection process manual

SL3151Ch03Frame Page 151 Thursday, September 12, 2002 6:12 PM

Benchmarking

151

4. Determine the Resource Requirements Identify the resources to perform the activity. Include factors such as direct material, direct labor (hours and grade), equipment requirements, information requirements, and space requirements. The resources might come from more than one department. It is crucial to trace all of the resources required to perform the activity. The resources can be determined by a careful analysis of the chart of accounts. When making the cost analysis, carefully choose among using actual, budgeted, standard, or planned cost information. Example: Inspector, material, handling equipment, inspection equipment, inspection area, inspection manual

5. Determine the Activity Drivers What are the factors external to the activity that cause more or less of the resources to be used? What drives the need for the activity and the level of resources required? Consider both efficiency and effectiveness, as follows: 1. Efficiency: Doing things right. 2. Effectiveness: Doing the right things Example: bad weather, poor product quality, automated equipment, workplace layout

6. Determine the Output of the Activity What units can be used to measure the output of the activity? This will be a measure of production level such as pieces produced, lots produced, invoices processed, checks written, or standard hours earned. Example: Lots of raw material inspected, pieces inspected, or material acceptance forms completed

7. Determine the Activity Performance Measure Identify that output measure that most closely controls the level of resources required. For example, when looking at clerical activities, the number of invoices is more significant than the dollar volume of the invoices. When moving material, the tons moved is more significant than the number of invoices represented. In general, the activity measure will be a resource input per unit of an output measure. Examples: • • • • • •

Cost/lot Pieces/hour Cost/unit Square foot per person Patents per engineer Drawings per engineer

SL3151Ch03Frame Page 152 Thursday, September 12, 2002 6:12 PM

152

Six Sigma and Beyond

• • • •

Lines of code per programmer Contact labor/company labor Sales dollars/sales manager Machine changeover –% total

Model the Activity Modeling an activity involves the following: • • • •

Define the process Define the cost or resource requirements Define the output variables Determine the metric or resource per unit of output (This may require the use of regression analysis or the design of experiments.)

Critical considerations are: • What is the relationship between fixed and variable costs? • What determines the capacity limitations of the process? • How much does overhead change with a change in the volume of business? It is important to distinguish between the metric (resource per unit of output) and the cost drivers. The metric or activity measure for inserting pins might be cost per pin inserted. However, the cost drivers might be the product design and the technology used. A different design might require fewer insertions, and a different technology might avoid the need for any insertions. Examples of Modeling The modeling of raw material cost per unit produced might consider the following variables: • • • • • •

The number of parts to be produced The standard raw material per part The percent scrap produced Raw material unit price Raw material quantity discounts Exchange rates

Note that a simple comparison of raw material cost as a percentage of sales dollars provides little real basis for comparing costs and cost improvement. The number of units sold of an item could be modeled as the number of potential buyers times the percentage who become aware of the product if they can get it times the percentage of potential buyers who can get the product times the percentage of triers who will be repeat buyers times the number of units purchased by a repeat buyer.

SL3151Ch03Frame Page 153 Thursday, September 12, 2002 6:12 PM

Benchmarking

153

When working with salaries and wages, it is necessary to take into consideration factors such as headcount, rate by grade, straight time/overtime ratios, benefits, skill level, age, education, union vs. non-union, and incentives. Salary and wage ratios that can be benchmarked are: • • • •

Skilled/unskilled labor Direct/indirect labor Training cost per employee Overtime hours/straight time hours

Flow Chart the Process To determine the sales dollars from a new account, start by flow charting the steps required to sell a new account. Start with cold calls and work through to a close. Use of symbols in flow charting: • • • • • •

Start or stop Flow lines Activity Document Decision Connector

Then ask some key questions: • What are the major activities? • What are the ratios required to forecast sales? • What factors affect the selling cost per rep or the revenue per rep? Does looking at these ratios tell you very much? What would you benchmark? Here is an example of activity performance measures for warehouse operations: Picking operations Orders filled per person per day Line items per person per day Pieces per person per day Number of picks per order Standard hours earned per day Line items per order Receiving operations Number of trucks unloaded per shift Number of pallets received per day Number of cases received per day Number of errors per day Direct labor hours unloading trucks

SL3151Ch03Frame Page 154 Thursday, September 12, 2002 6:12 PM

154

Six Sigma and Beyond

Incoming QC operations Number of inspections per period Number of rejects per period Direct inspection labor hours Putaway/storage operations Number of full pallet putaways per period Number of loose case putaways per period Direct labor hours putaway or storage Cube utilization Truck loading Number of units loaded per truck per period Number of trucks per period Time per trailer Customer service operations Fill rate Elapsed time between order and shipment Error rate Customer calls taken per day Number of problems solved per call Number requiring multiple calls Number of credits issued Number of backorders At this stage we are ready to identify information required when meeting with the benchmark partner. The following information is typical and may be used to focus the meeting with the benchmark partner and to highlight information requirements: 1. Description of company activity and results: 2. Alternative ways of performing the activity: Alternative 1: Alternative 2: Alternative 3: 3. The pros and cons of the alternatives are: Pros Cons Alternative 1: Alternative 2: Alternative 3: 4. Information required to reach a conclusion as to the best approach: Review the assumptions for the study to make sure that the outcomes are correlated to what you were studying. (At this stage, it is not unusual to find surprises. That is, you may find items that you overlooked or you thought were unimportant and so on.)

SL3151Ch03Frame Page 155 Thursday, September 12, 2002 6:12 PM

Benchmarking

ACTIVITIES

DURING THE

155

VISIT

By far the most important characteristic of the visit is to: Observe, question, analyze and learn

Make sure to notice: • • • •

What are they doing? How is it different from what we are doing? Why are they doing it that way? How can the results be measured?

Ask open-ended questions, just as you did when observing your own operations. For example, for “who”: • • • •

Who Who Who Who

does it per the job description? is actually doing it? else could do it? should be doing it?

Ask similar questions for what, where, when, why, and how. Understand the Benchmark Partner’s Activities Follow the procedures described above for analyzing the company activities. You may encounter some analytical difficulties because of the following factors. Accounting differences Account definitions vary in terms of what is included in the account. For example, does the cost of raw material include the cost of freight in and insurance? Where is scrap accounted for? Cost allocations. Identification of all multi-department costs. Different economies of scale/learning curve Specialization Automation Time/unit Identification of All of the Factors Required for Success Factors to consider when trying to determine if you have identified all the factors required for success include the following: Analysis and intuition • MRP and inventory cycle count • Salary and wage comparisons — are the jobs really comparable?

SL3151Ch03Frame Page 156 Thursday, September 12, 2002 6:12 PM

156

Six Sigma and Beyond

• The use of manufacturing work cells (This may require a change in socialization.) • Level of advertising per dollar of sales (Just knowing this may not be very helpful. The relevant question is, “How effectively are the advertising dollars spent?”) Regression analysis Warehouse study Design of experiments

ACTIVITIES

AFTER THE

VISIT

Key activities after the visit include the following: • • • • • •

Be sure to send a thank you note promptly. Document findings for each visit. Summarize all findings — analysis and synthesis. Compare current operations with findings. Gather more specific data if required. Identify opportunities for improvement — combine, eliminate, change order, etc. • Develop team recommendation. • Distribute benchmark report.

Benchmarking Examples 1. Functional Analysis Hours/1000 pcs Company Best Company

Functions Primary machining Heat training Grinding Assembly Packing

.75

.50

Gap .25

2. Cost Analysis

Cost Item Raw material Direct labor

% Total Cost

Cum % Total

Company

Cost per Unit Best

40 20

40 60

17.50 7.50

15.50 5.50

3. Technology Forecasting The benchmarking of competitive technologies can be very critical. This is particularly true when the product or technical life cycle is very short. Keys are:

SL3151Ch03Frame Page 157 Thursday, September 12, 2002 6:12 PM

Benchmarking

157

• Knowing the current technology and its limitations • Identifying the emerging technologies that become the new benchmarks. • Knowing what customers really buy and relating this to the emerging technology. • Having the courage and foresight to change. 4. Financial Benchmarking Financial benchmarking compares a company (or the major segments of a company) relative to the financial performance of other companies. The modified Du Pont chart provides a convenient way to do this. The idea of the modified Du Pont plan is to start with the return on equity and progressively calculate the return on assets, profit margin, gross margin, sales, cost of goods sold (COGS), sales per day, cost of goods sold per day, days inventory (COGS), days receivable (COGS), and days payable (COGS). Company results can be compared with data provided by: • Dun and Bradstreet Industry Norms and Key Business Ratios • Robert Morris Associates Annual Statement Summary • Prentice Hall Almanac of Business and Industrial Financial Ratios 5. Sales Promotion and Advertising The comparison of company strategy versus industry strategy can lead to the need for more specific benchmarking studies. 6. Warehouse Operations The performance of units engaged in essentially the same type of activity can be compared using statistical regression analysis. This technique can be used to determine the significant independent variables and their impact on costs. Exceptionally good and bad performance can be identified and this provides the basis for further benchmarking studies. 7. PIMS Analysis The PIMS analysis is a further application of multiple regression analysis. It can be used to identify the major determinants of company profitability. 8. Purchasing Performance Benchmarks The Center for Advanced Purchasing Studies (Tempe, Arizona) has benchmarked the purchasing activity for the petroleum, banking, pharmaceutical, food service, telecommunication services, computer, semiconductor, chemical, and transportation industries. For a wide variety of activity measures, the reports provide the average value, the maximum, the minimum, and the median value.

SL3151Ch03Frame Page 158 Thursday, September 12, 2002 6:12 PM

158

Six Sigma and Beyond

Motorola Example Perhaps one of the most famous examples of benchmarking in recent history is the Motorola example. Motorola, through “Operation Bandit,” was able to cut the product development time for a new pager in half to 18 months based on traveling the world and looking for “islands of excellence.” These companies were in various industries: cars, watches, cameras, and other technically intensive products. The solution required a variety of actions: • • • •

Automated factories Removing barriers in the workplace Training of 100,000 employees Technical sharing alliance with Toshiba

Motorola was particularly impressed by the P200 program of a Hitachi plant. This stands for a 200% increase in productivity by year end. The plant set immutable deadlines for the solutions to problems. All departments had six sigma goals.

GAP ANALYSIS DEFINITION

OF

GAP ANALYSIS

There are at least two ways to view “gap.” 1. Result Gap — A result gap is the difference between the company performance and the performance of the best in class as determined by the benchmarking process. This gap is defined in terms of the activity performance measure. The gap can be positive or negative. 2. Practice or Process Gap — A practice or process gap is the difference between what the company does in carrying out an activity and what the best in class does as determined by the benchmarking process. This gap is measured in terms of factors such as organizational structure, methods used, technology used, or material used. The gap can be positive or negative. The determination of a gap can be a strong motivator toward the improvement of performance. It can create the tension necessary for change to occur.

CURRENT

VERSUS

FUTURE GAP

It is critical to distinguish between the current gap and the likely future gap. Remember that the benchmark partner’s performance will not remain at the current levels nor will the expectations of the marketplace. More likely than not, performance will improve as time goes on. A company that concentrates only on closing the current gap will find itself in a constant game of catch-up.

SL3151Ch03Frame Page 159 Thursday, September 12, 2002 6:12 PM

Benchmarking

159

The company that ignores likely improvements of the benchmark gets caught in the Z trap. The Z trap, of course, is the step-wise improvement that is OK for catching up but never good enough to be the best in class. To summarize the benchmark findings, it is often helpful to make a tabulation showing the current practice and metric and the expected future practice and metric for the company, the competition, and the best in class. In order for the benchmarking process to be effective, it is critical that management accept the validity of the gap and provide the resources necessary to close the gap.

GOAL SETTING GOAL DEFINITION Two terms that are often used interchangeably are “objective” and “goal.” There is, of course, no one correct definition. As long as the terms are used consistently within an organization, it does not really matter. For our purposes, however, objectives are broad areas where something is to be accomplished, such as sales and marketing or customer service. Goals, on the other hand, are specific and measurable and have a time frame. For example, “Answer all inquiries within 2 hours by the 3rd quarter of 2002.”

GOAL CHARACTERISTICS For best results, goals should be (a) tough (you need to stretch to attain them) and (b) attainable (realistic). When evaluating these two characteristics, always take into consideration the current capabilities of the company versus the benchmark candidate now and projected. A good way to monitor progress towards attainment is through trend charting.

RESULT

VERSUS

EFFORT GOALS

Result goals define the specific performance measure to be achieved. For example, “Sell $4 million of product x to company y in 2003.” Effort goals define specific accomplishments that are completely under the control of the goal setter. They are necessary to achieve the result goals. They can be thought of as action plans. For example, an effort goal would be, “Make x cold calls a week to new departments of x company”.

GOAL SETTING PHILOSOPHY Best of the Best versus Optimization There can be a clear difference between implementing an inventory control system that ensures that a company never runs out of stock and an inventory control system that optimizes the level of inventory. The optimum inventory balances off the cost of holding the inventory and the cost of carrying the inventory.

SL3151Ch03Frame Page 160 Thursday, September 12, 2002 6:12 PM

160

Six Sigma and Beyond

A similar consideration is that of determining the optimum feature set for a product, taking into consideration what specific market segments value and will pay for. Differentiation that is not valued by the market could result in an unnecessary expenditure of funds. The determination of value has to be based on the underlying need of the customer. If this had not been done, there would be no need to have produced a ballpoint pen, only a better fountain pen. Who asked for electricity, the camera, or the copy machine before they existed? No one by name, but many in terms of desire and underlying need. There is a fundamental difference between working within the constraints facing a business and removing the constraints. For example, a company can either (a) optimize production given the setup time for a job or (b) reduce the setup time. Optimization within the constraints leads to larger lots, higher inventory, perhaps poorer quality, and delays. It is much more effective to remove the constraint. The key to manufacturing excellence is to remove the constraints that cause the tradeoffs between cost and customer satisfaction. Kaizen versus Breakthrough Strategies The Kaizen philosophy of management stresses making small, constant improvements as opposed to looking for the one magic silver bullet that will lead to success. Which company is likely to be more innovative: (a) a company that is looking for the one big idea or (b) a company that is constantly making small improvements? Both are appropriate strategies depending on the specific situation. However, if a company is in dire need of improvement there is no better way than to look at benchmarking. The benchmarking in this case will be a true breakthrough. On the other hand, the Kaizen approach tells us that we should not relax in our effort to be the best. There is always something that we can do better.

GUIDING PRINCIPLE IMPLICATIONS The decisions made regarding goals can have a profound interaction with the mission statement of the company and/or the values as defined in the statement of guiding principles. The statement of guiding principles generally consists of: 1. Mission statement — a description of the product and markets served or who, what, and how 2. Values — those human and ethical principles that guide the conduct of the business

GOAL STRUCTURE Cascading Goal Structure A consistent goal structure can provide focus and direction to the entire organization. In order to create this, start with the most important goal, as viewed by the president or chief executive officer, and decompose each of these by functional area working from one management level to the next. For example, starting with a return on equity

SL3151Ch03Frame Page 161 Thursday, September 12, 2002 6:12 PM

Benchmarking

161

goal, what does this mean each department has to do? What does this suggest in the way of specific benchmarking goals? Interdepartmental Goals One of the most elusive tasks of management is to get all departments to work together toward a common set of goals. One way to manage this is to have each department indicate its goals and what it requires in the way of performance from other departments to reach those goals. A cross tabulation can then be used to develop the total goals for a department or function.

ACTION PLAN IDENTIFICATION AND IMPLEMENTATION The benchmarking process has been used to identify the present and projected result and performance gap. The actual solution to closing the gap may be the synthesis of several of the benchmark partner’s ideas. In order to creatively identify new solutions, the following questions can be helpful: Put to other uses? New ways to use as is? Other uses if modified? Adapt? What else is this like? What other ideas does this suggest? Does the past offer a parallel? What could I copy? Whom could I emulate? Modify? New twist? Change meaning, color, motion, sound, odor, form or shape? Other changes? Magnify? What to add? More time? Greater frequency? Stronger? Higher? Longer? Thicker? Extra value? Plus ingredients? Duplicate? Multiply? Exaggerate? Minimize? What to subtract? Smaller? Condensed? Miniature? Lower? Shorter? Lighter? Omit? Streamline? Split up? Understate? Substitute? Who else instead? What else instead? Other ingredients? Other material? Other process? Other power? Other place? Other approach? Other tone of voice? Rearrange? Interchange components? Other pattern? Other layout? Other sequence? Transpose cause and effect? Change place? Change schedule? Reverse? Transpose positive and negative? How about opposites? Turn it backwards? Turn it upside down? Reverse roles? Change shoes? Turn tables? Turn other cheek?

SL3151Ch03Frame Page 162 Thursday, September 12, 2002 6:12 PM

162

Six Sigma and Beyond

Combine? How about a blend, an alloy, an assortment, an ensemble? Combine units? Combine purpose? Combine appeals? Combine ideas?

A CREATIVE PLANNING PROCESS It is highly desirable that more than one alternative way to achieve a goal be identified. It is also critical that each viable alternative be fully evaluated on its own merits and that a conscious choice be made to select the best alternative. For each alternative, consider the following process: 1. 2. 3. 4. 5.

Develop a vision or a dream of what you would like to have happen. Identify the critical success factors for achieving the vision. Determine the required action programs. Match resource requirements and availability. Determine if the vision is feasible and either implement the required action programs or consciously decide to drop or modify the vision. 6. Implement the plan by assigning action plan responsibility. 7. Monitor performance versus expectations and revise the plan as required.

ACTION PLAN PRIORITIZATION If more action plans are identified than can be implemented, it will be necessary to prioritize the action plans relative to the corporate goals and customers’ needs, wants, and expectations. The process identified earlier in the discussion of prioritization of benchmark alternatives may be used for this purpose. One aspect of action plan prioritization is the determination of the most desirable plan from a financial point of view. Evaluations of this type often involve the comparison of cash flows that occur in different years. Consequently, the time value of money has to be taken into consideration when deciding which plan is most desirable.

ACTION PLAN DOCUMENTATION The action plan must be documented and the person(s) responsible for individual tasks must be identified: • Use of Critical Path Scheduling Tools • Action plan format • Technique for sequencing activities using Post It Notes • Importance of identifying milestones, deliverables, and decision-making roles

MONITORING

AND

CONTROL

A good way to maintain monitoring and control is through formalized periodic reporting of performance versus plan. Issues to keep in mind are:

SL3151Ch03Frame Page 163 Thursday, September 12, 2002 6:12 PM

Benchmarking

163

Need to assign responsibility for ongoing review and evaluation Use of a control chart for each variable with the responsible person identified Just because the official benchmarking study has been completed does not mean that you are done. To the contrary, you must be vigilant in monitoring your competitor’s activities by tracking the competitive performance versus plan. This is because things change and modifications must be made to recalibrate the results. Some items of interest are: • • • • • •

Benchmarks may need to be recalibrated. Changes may occur in industry, customers, or competitors. How fast are things moving and in what direction? Critical success factors may change. New competitors may enter the field. Competition may be better or worse than expected.

FINANCIAL ANALYSIS OF BENCHMARKING ALTERNATIVES When comparing benchmarking alternatives, it is often necessary to take into consideration the fact that cash is received and/or disbursed in different time periods for each of the alternatives. Cash received in the future is not as valuable as cash received today because cash received today can be reinvested and earn a return. In order to compare the current value of cash received or disbursed in different periods, it is necessary to convert a future dollar value to its present value. For example, the present value of $1100 received a year from now is $1000 if money can be invested at 10%. The alternative way to view this is to note that the future value of $1000 invested for one year at 10% is equal to 1000 times 1.10 or $1100. The following table can be used to determine the present value of a future cash flow depending upon the discount rate and the number of years from the present that the investment is made. To relate to the discussion above, note that the Present Value Factor for one year at 10% is .9091. Therefore, the present value of $1100 received a year from now is $1000, i.e., $1100 times .9091. A typical capital project of benchmark alternative evaluation is discussed in the following pages. The projected net income after tax as well as a summary of the investments made in the project, the after-tax salvage value, and the cash flow associated with the project are indicated. The assumptions used to generate the net income are indicated below the projection. Note the separation of fixed and variable cost and the relationship between specific assumptions and the level of capacity utilization. In this case, the investment is assumed to occur at the end of the first year. Also, there is no increase in working capital associated with the construction of the plant. The cash flow can be determined in one of two ways: (a) it is equal to the net income after tax plus depreciation or (b) it is equal to revenue minus operating

SL3151Ch03Frame Page 164 Thursday, September 12, 2002 6:12 PM

164

Six Sigma and Beyond

Present Value Factors Discount Rate Year

10%

20%

30%

40%

1 2 3 4 5 6 7 8 9 10

0.9091 0.8264 0.7513 0.6830 0.6209 0.5645 0.5132 0.4665 0.4241 0.3822

0.8333 0.6944 0.5787 0.4823 0.4019 0.3349 0.2791 0.2326 0.1938 0.1615

0.7692 0.5917 0.4552 0.3501 0.2693 0.2072 0.1594 0.1226 0.0943 0.0725

0.7143 0.5102 0.3644 0.2603 0.1859 0.1328 0.0949 0.0678 0.0484 0.0346

expenses minus taxes. The net present value is indicated for several discount rates (10 to 40%). The net present value at 10% is determined, for example, as in Table 3.2. If the company cost of capital is 15%, then this project would be acceptable because the net present value is positive at that discount rate. A similar analysis can be used to determine a breakeven product price — see Table 3.3.

MANAGING BENCHMARKING FOR PERFORMANCE To summarize this chapter, here are some do’s and don’ts for successful benchmarking: Requirements for success • Use goal-oriented management — measure and monitor everything; link to compensation plan. • Start small and showcase. • Recognize that conflict is inevitable because of the need to share resources to reach conflicting goals. Management has to make tough decisions to resolve the healthy conflict. • Link goals to action plans. • Understand that adequate resources are necessary to ensure the success of the plan. • Ensure continuing top management support with the recognition that benchmarking does not necessarily supply a quick fix. • Place emphasis both on the result (what to do) and the process (how to do it). • Accept the concept of constant, incremental change. • A blend of analytical and intuitive skills requiring the ability to synthesize sometimes ambiguous data is needed. • Be willing to admit that change or improvement is possible and perhaps desirable. • Focus on the needs of specific target market segments and business strategy when setting the priorities to benchmark.

SL3151Ch03Frame Page 165 Thursday, September 12, 2002 6:12 PM

Benchmarking

165

TABLE 3.2 An Example of Cash Flow and Present Value End of Year

Cash Flow

Present Value Factor

Present Value

1 2 3 4

–1,000,000 246,680 597,764 1,008,814

0.9091 0.8264 0.7513 0.6830 Total Net Present Value

–909,091 203,868 449,109 689,034 432,919

TABLE 3.3 Benchmark Project Evaluations 2001 Sales (units) Unit price Revenue Operating expense Depreciation Net income before tax Tax Net income after tax Investment Salvage value Cash flow Interest rate (%) Net present value

1,000,000 10,000 –1,000,000 10 432,919

Assumptions Plant capacity (units) Unit price — start Tax rate (%) Depreciation (%) Capacity utilization (%) Price increase (%) Operating Expense Units Fixed Variable 10,000 20,000 30,000 40,000 50,000 60,000 70,000

200 200 200 400 400 500 500

20.00 20.00 20.00 21.00 21.00 21.00 21.00

2002

2003

2004

21,000 38.00 798,000 420,200 50,000 327,800 131,120 196,680

49,000 40.66 1,992,340 1,029,400 50,000 912,940 365,176 547,764

66,500 45.54 3,028,357 1,397,000 50,000 1,581,357 632,543 948,814

246,680 20 170,404

597,764 30 2,030

1,008,814 40 –107,982

70,000 38.00 40 5 30 7

70 7

95 12

SL3151Ch03Frame Page 166 Thursday, September 12, 2002 6:12 PM

166

Six Sigma and Beyond

• Create a corporate culture that thrives on learning and self-improvement with constant, though gradual, change. Constantly apply the Plan, Do, Check, Act cycle. • Use Statistical Process Control to determine when events, results, or processes are out of control. • Change the role of middle management. The middle manager is no longer “the boss.” Middle managers must encourage and enable workers to think. Common mistakes • Giving lip service to the process and not providing the resources to get the job done properly • Failure to effectively communicate the benchmark findings and drive them to implementation: all analysis and no action • Failure to precisely define the expected results of benchmark improvement and to monitor actual performance (In the absence of this, no organizational learning occurs.) • Lack of a comprehensive prioritization of the benchmarking projects to ensure the best cost/benefit results • The expectation of quick results and a short-term focus on quarterly earnings • Lack of constant purpose, focus, and direction • Failure to implement results in small size, meaningful modules with specific deliverables; looking for “the” big win • Unwillingness to face the reality of a situation and recognize that change is necessary and that hard choices have to be made • Not drawing the correct balance between required accuracy and the practical ability to achieve better results; 100 percent accuracy, certainty, or performance is not required • Failure to recognize that the early follower is almost as profitable as the pioneer and sometimes even more so • Reliance on executive office analysis versus observation of the handson experience of others both within and outside the company • Focus on problem reduction and not problems avoidance • Failure to realize that, in most cases, benchmarking follows strategy • Failure to recognize the constantly rising level of expectations in the marketplace • Lack of contingency planning • Failure to get participation at all levels and to break down interdepartmental barriers so that the total resources of the organization can be focused on the solution to common problems

REFERENCES Berle, G., Business Information Sourcebook, Wiley, New York, 1991. Buzzell, R.D. and Gale B.T., The PIMS Principles, Free Press, New York, 1987.

SL3151Ch03Frame Page 167 Thursday, September 12, 2002 6:12 PM

Benchmarking

167

Clifford, D.K. and Cavanagh, R.E., The Winning Performance: How America’s High Growth Midsize Companies Succeed, Bantam Books, New York, 1988. Garvin, D.A, Managing Quality, Free Press, New York, 1988. Hall, W.K., Survival Strategies in a Hostile Environment, Harvard Business Review, Sept./Oct. 1980, pp. 34–38. Higgins, H. and Vincze, A., Strategic Management, Dryden Press, New York, 1989. Smith, G.N. and Brown, P.B., Sweat Equity, Simon and Schuster, New York, 1986. Stamatis, D.H., Total Quality Service, St. Lucie Press, Boca Raton, FL, 1996. Stamatis, D.H., TQM Engineering Handbook, Marcel Dekker, New York, 1997.

SELECTED BIBLIOGRAPHY Balm, G.J., Benchmarking: A Practitioner’s Guide for Becoming and Staying Best of the Best, Quality & Productivity Management Association, Schaumburg, IL, 1992. Barnes, B., Squeeze Play: Satisfaction Programs Are Key for Manufacturers Caught Between Declining and Increasing Raw Material Costs, Quirk’s Marketing Research Review, Oct. 2001, pp. 44–47. Bosomworth, C., The Executive Benchmarking Guidebook, Management Roundtable, Boston, MA, 1993. Boxsvell, R.J., Jr., Benchmarking for Competitive Advantage, McGraw-Hill, New York, 1994. Camp, R., Business Process Benchmarking: Finding and Implementing Best Practices, ASQC Quality Press, Milwaukee, WI, 1995. Chang, R.Y. and Kelly, P.K., Improving through Benchmarking, Richard Chang Associates, Publications Division, Irvine, CA, 1994. Karlof, B. and Ostblom, S., Benchmarking: A Signpost to Excellence in Quality and Productivity, John Wiley & Sons, New York, 1993. Lewis, S., Cleaning Up: Ongoing Satisfaction Measurement Adds to Japanese Janitorial Firm’s Bottom Line, Quirk’s Marketing Research Review, Oct 2001, pp. 20–21, 68–70.

SL3151Ch03Frame Page 168 Thursday, September 12, 2002 6:12 PM

SL3151Ch04Frame Page 169 Thursday, September 12, 2002 6:11 PM

4

Simulation

As companies continue to look for more efficient ways to run their businesses, improve work flow, and increase profits, they increasingly turn to simulation, which is used by best-in-class operations to improve their processes, achieve their goals, and gain a competitive edge. Simulation is used by some of the world’s most successful companies, including Ford, Toyota, Honda, DaimlerChrysler, Volkswagen, Boeing, Delphi Automotive Systems, Dell Corp. Gorton Fish Co., and many others. Both design and process simulations have become increasingly important and integral tools as businesses look for ways to strip non-value-adding steps from their processes and maximize human and equipment effectiveness, all parts of the six sigma philosophy. The beauty of simulation is that, while it complements and aids in the six sigma initiative, it can also stand alone to improve business processes. In this chapter, we do not dwell on the mathematical justification of simulation; rather, we attempt to explain the process and identify some of the key characteristics in any simulation. Part of the reason we do not elaborate on the mathematical formulas is the fact that in the real world, simulations are conducted via computers. Also, readers who are interested in the theoretical concepts of simulation can refer to the selected bibliography found both at the end of the chapter and at the end of this volume.

WHAT IS SIMULATION? Simulation is a technology that allows the analysis of complex systems through statistically valid means. Through a software interface, the user creates a computerized version of a design or a process, otherwise known as a “model.” The model construction is a basic flow chart with great additional capabilities. It is the interface a company uses to build a model of its business process. Simulation technology has been around for a generation or more, with early developments mostly in the area of programming languages. In the last 10 to 15 years, a number of off-the-shelf software packages have become available. More recently, these tools have been simplified to the point that your average business manager with no industrial engineering skills can effectively employ this technology without requiring expert assistance. (Some companies have actually modified the commercial versions to adopt them into their own environments.) Simplicity is the key to today’s simulation software. The basic simulation structure is as follows: after flow charting the process, the user inputs information about how the process operates by simply filling in blanks. While completing a model, the user answers three questions at each step of the process: how long does the step 169

SL3151Ch04Frame Page 170 Thursday, September 12, 2002 6:11 PM

170

Six Sigma and Beyond

take, how often does it happen, and who is involved? After the model is built and verified, it can be manipulated to do two critical things: analyze current operations to identify problem areas and test various ideas for improvement. The latest improvements in simulation software have made it an excellent tool for enhancing the design for six sigma (DFSS) process, which strives to eliminate eight wastes: overproduction, motion, inventory, waiting, transportation, defects, underutilized people, and extra processing. DFSS targets non-value-added activities — the same activities that contribute to poor product quality. In this chapter we are not going to discuss commercial packages; rather we are going to introduce three methodologies that facilitate simulation — Monte Carlo, Finite Element analysis, and Excel’s Solver approach.

SIMULATED SAMPLING The sampling method, known generally as Monte Carlo, is a simulation procedure of considerable value. Let us assume that a product is being assembled by a two-station assembly line. There is one operator at each of the two stations. Operation A is the first of the two operations. The operator completes approximately the first half of the assembly and then sets the half-completed assembly on a section of conveyor where it rolls down to operation B. It takes a constant time of 0.10 minute for the part to roll down the conveyor section and be available to operator B. Operator B then completes the assembly. The average time for operation A is 0.52 minute per assembly and the average time for operation B is 0.48 minute per assembly. We wish to determine the average inventory of assemblies that we may expect (average length of the waiting line of assemblies) and the average output of the assembly line. This may be done by simulated sampling as follows: 1. The distributions of assembly time for operations A and B must be known or procured. Usually this is done through historical data, sometimes with surrogate. A study was taken for both operations, and two frequency distributions were constructed (not shown here). In the case of operation A, the value 0.25 minute occurred three times, 0.30 occurred twice, and so on. For operation A, the mean was 0.52 min with N = 167 and for operation B the mean was 0.48 with N = 115. The two distributions do not necessarily fit mathematical distributions but this is not important. 2. Convert the frequency distributions to cumulative probability distributions. This is done by summing the frequencies that are less than or equal to each performance time and plotting them. The cumulative frequencies are then converted to percents by assigning the number 100 to the maximum value. The cumulative frequency distribution (not shown here) for operation A began at the lowest time, 0.25 minute; there were three observations. Three is plotted on the cumulative chart for the time 0.25 minute. For the performance time 0.30 minute, there were two observations, but there were five observations that measured 0.30 minute or

SL3151Ch04Frame Page 171 Thursday, September 12, 2002 6:11 PM

Simulation

171

less, so the value five is plotted for 0.30 minute. For the performance time 0.35 minute, there were 10 observations recorded, but there were 15 observations that measured 0.35 minute or less. When the cumulative frequency distribution was completed, a cumulative percent scale was constructed on the right, by assigning the number 100 to the maximum value, 167 in this case, and dividing the resulting scale into equal parts. This results in a cumulative probability distribution. We can use this distribution to say for example that 100 percent of the time values were 0.85 minute or less, 55.1 per cent were 0.50 minute or less and so on. 3. Sample at random from the cumulative distributions to determine specific performance times to use in simulating the operation of the assembly line. We do this by selecting numbers between 0 and 100 at random (representing probabilities or percents). The random numbers could be selected by any random process, such as drawing numbered chips from a box, using a random number table, or using computer-generated random numbers. For small studies, the easiest way is to use a table of random numbers. The random numbers are used to enter the cumulative distributions in order to obtain time values. In our example, we start with the random number 10. A horizontal line is projected until it intersects the distribution curve; a vertical line projected to the horizontal axis gives the midpoint time value associated with the intersected point on the distribution curve, which happens to be 0.40 minute for the random number 10. Now we can see the purpose behind the conversion of the original distribution to a cumulative distribution. Only one time value can now be associated with a given random number. In the original distribution, two values would result because of the bell shape of the curve. Sampling from the cumulative distribution in this way gives time values in random order, which will occur in proportion to the original distribution, just as if assemblies were actually being produced. Table 4.1 gives a sample of 20 time values determined in this way from the two distributions. 4. Simulate the actual operation of the assembly line. This is done in Table 4.2, which is very similar to waiting (queuing) line problems. The time values for operation A (Table 4.1) are first used to determine when the half-completed assemblies would be available to operation B. The first assembly is completed by operator A in 0.40 minute. It takes 0.10 minute to roll down to operator B, so this point in time is selected as zero. The next assembly is available 0.40 minute later, and so on. For the first assembly, operation B begins at time zero. From the simulated sample, the first assembly requires 0.60 minute for B. At this point, there is no idle time for B and no inventory. At time 0.40 the second assembly becomes available, but B is still working on the first so the assembly must wait 0.20 minute. Operator B begins

SL3151Ch04Frame Page 172 Thursday, September 12, 2002 6:11 PM

172

Six Sigma and Beyond

TABLE 4.1 Simulated Samples of 20 Performance Time Values for Operations A and B Operation A

Random Number 10 22 24 42 37 77 99 96 89 85 28 63 9 10 7 51 2 1 52 7 Totals

Operation B

Performance Time from Cumulative Distribution for Operation A 0.40 0.40 0.45 0.50 0.45 0.60 0.85 0.75 0.65 0.65 0.45 0.55 0.40 0.40 0.35 0.50 0.30 0.25 0.50 0.35 9.75

Random Number 79 69 33 52 13 16 19 4 14 6 30 25 38 0 92 82 20 40 44 25

Performance Time from Cumulative Distribution for Operation B 0.60 0.50 0.40 0.45 0.35 0.35 0.35 0.30 0.35 0.30 0.40 0.35 0.40 0.25 0.70 0.60 0.35 0.40 0.45 0.35 8.20

work on it at 0.60. From Table 4.1, the second assembly requires 0.50 minute for B. We continue the simulated operation of the line in this way. The sixth assembly becomes available to B at time 2.40, but B was ready for it at time 2.30. He therefore was forced to remain idle for 0.10 minute because of lack of work. The completed sample of 20 assemblies is progressively worked out — see Table 4.2. The summary at the bottom of Table 4.4 shows the result in terms of the idle time in operation B, the waiting time of the parts, the average inventory between the two operations, and the resulting production rates. From the average times given by the original distributions, we would have guessed that A would limit the output of the line since it was the slower of the two operations. Actually, however, the line production rate is less than that dictated by A (116.5 pieces per hour compared to 123 pieces per hour for A as an individual operation). The reason is that the interplay

SL3151Ch04Frame Page 173 Thursday, September 12, 2002 6:11 PM

Simulation

173

TABLE 4.2 Simulated Operation of the Two-Station Assembly Line when Operation A Precedes Operation B Assemblies Available for Operation B Operation B at Begins at

Operation B Time in Ends at Operation B

Idle Waiting Time of Assemblies

Number of Parts in Line, Excluding Assembly Being Processed in Operation B

0.00 0.00 0.60 0 0 0.40 0.60 1.10 0 0.20 0.85 1.10 1.50 0 0.25 1.35 1.50 1.95 0 0.15 1.80 1.95 2.30 0 0.15 2.40 2.40 2.75 0.10 0 3.25 3.25 3.60 0.50 0 4.00 4.00 4.30 0.40 0 4.65 4.65 5.00 0.35 0 5.30 5.30 5.60 0.30 0 5.75 5.75 6.15 0.15 0 6.30 6.30 6.65 0.15 0 6.70 6.70 7.10 0.05 0 7.10 7.10 7.35 0 0 7.45 7.45 8.15 0.10 0 7.95 8.15 8.75 0 0.20 8.25 8.75 9.10 0 0.50 8.50 9.10 9.50 0 0.60 9.00 9.50 9.95 0 0.50 9.35 9.95 10.30 0 0.60 Idle time in operation B = 2.10 minutes Waiting time of parts = 3.15 minutes Avenge inventory of assemblies between A and B = 3.15/9.35 = 0.34 assemblies Average production rate of A = [20 × 60]/9.75 = 123 pieces/hour Average production rate of B (while working) = [20 × 60]/8.20 = 146 pieces/hour Average production rate of A and B together = [20 × 60]/10.30 = 116.5 pieces/hour

0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 1 2 2 2

Note: In the above computations, 20 is the total number of completed assemblies; 9.75 is the total work time of operation A for 20 assemblies from Table 4.1; 8.20 is the total work time, exclusive of idle time, for operation B for 20 assemblies from Table 4.1.

of performance times for A and B does not always match up very well, and sometimes B has to wait for work. B’s enforced idle time plus B’s total work time actually determine the maximum production rate of the line. A little thought should convince us that, if possible, it would have been better to redistribute the assembly work so that A is the faster of the two operations. Then the probability that B will run out of work is reduced. This is demonstrated by Table 4.3, which assumes a simple reversal of the sequence of A and B. The same

SL3151Ch04Frame Page 174 Thursday, September 12, 2002 6:11 PM

174

Six Sigma and Beyond

TABLE 4.3 Simulated Operation of the Two-Station Assembly Line when Operation B Precedes Operation A Assemblies Available for Operation A Operation A at Begins at

Operation A Time in Ends at Operation A

Idle Waiting Time of Assemblies

Number of Parts in Line, Excluding Assembly Being Processed in Operation A

0.00 0.00 0.40 0 0 0.50 0.50 0.90 0.10 0 0.90 0.90 1.35 0 0 1.35 1.35 1.85 0 0 1.70 1.85 2.30 0 0.15 2.05 2.30 2.90 0 0.25 2.40 2.90 3.75 0 0.40 2.70 3.75 4.50 0 1.05 3.05 4.50 5.15 0 1.45 3.35 5.15 5.80 0 1.80 3.75 5.80 6.25 0 2.05 4.10 6.25 6.80 0 2.15 4.50 6.80 7.20 0 2.30 4.75 7.20 7.60 0 2.45 5.45 7.60 7.95 0 2.15 6.05 7.95 8.45 0 1.90 6.40 8.45 8.75 0 2.05 6.80 8.75 9.00 0 1.95 7.25 9.00 9.50 0 1.75 7.60 9.50 9.85 0 1.90 Idle time in operation A = 0.10 minute Waiting time of parts = 25.75 minutes Average inventory of assemblies between A and B = 25.75/7.60 = 3.4 assemblies Average production rate of A (while working) = [20 × 60]/9.75 = 123 pieces/hour Average production rate of B = [20 × 60]/8.20 = 146 pieces/hour Average production rate of A and B together = [20 × 60]/9.85 = 122 pieces/hour

0 0 0 0 1 1 1 2 2 3 3 4 4 5 5 5 5 6 5 6

sample times have been used and the simulated operation of the line has been developed as before. With the faster of the two operations being first in the sequence, the output rate of the line increases and approaches the rate of the limiting operation, and the average inventory between the two operations increases. With the higher average inventory there, the second operation in the sequence is almost never idle owing to lack of work. Actually, this conclusion is a fairly general one with regard to the balance of assembly lines; that is, the best labor balance will be achieved when each succeeding operation in the sequence is slightly slower than the one before it. This minimizes the idle time created when the operators run out of work because of the variable performance times of the various operations. In practical

SL3151Ch04Frame Page 175 Thursday, September 12, 2002 6:11 PM

Simulation

175

situations, it is common to find safety banks of assemblies between operations in order to absorb these fluctuations in performance. We may have wanted to build a more sophisticated model of the assembly line. Our simple model assumed that the performance times were independent of other events in the process. Perhaps in the actual situation, the second operation in the sequence would tend to speed up when the inventory began to build up. This effect could have been included if we had knowledge of how inventory affected performance time. If we have followed this simulation example through carefully, we may be convinced that it would work but that it would be very tedious for problems of practical size. Even for our limited example, we would probably wish to have a larger run on which to base conclusions, and there would probably be other alternatives to test. For example, there may be several alternative ways to distribute the total assembly task between the two stations, or more than two stations could be considered. Which of the several alternatives would yield the smallest incremental cost of labor, inventory costs, etc.? To cope with the problem of tedium and excessive person-hours to develop a solution, the computer may be used. If a computer were programmed to simulate the operation of the assembly line, we would place the two cumulative distributions in the memory unit of the computer. Through the program, the computer would select a performance time value at random from the cumulative distribution for A in much the same fashion as we did by hand. Then it would select at random a time value from the cumulative distribution for B, make the necessary computations, and hold the data in memory. The cycle would repeat, selecting new time values at random, adding and subtracting to obtain the record that we produced by hand. A large run could be made easily and with no more effort than a small run. Various alternatives could be evaluated quickly and easily in the same manner.

FINITE ELEMENT ANALYSIS (FEA) This technique is not thought of as being a reliability improvement method, yet it can contribute significantly to its enhancement. Finite Element Analysis (FEA) is a technique of modeling a complex structure into a collection of structural elements that are interconnected at a given number of nodes. The model is subjected to known loads, whereby the displacement of the structure can be determined through a set of mathematical equations that account for the element interactions. The reader is encouraged to read Buchanan (1994) and Cook (1995) for a more complete and easy understanding of the theoretical aspects of FEA. In commercial use, FEA is a computer-based procedure for analyzing a complex structure by dividing it into a number of smaller, interconnected pieces (the “finite elements”), each with easily definable load and deflection characteristics.

TYPES

OF

FINITE ELEMENTS

The library of finite elements available in general purpose codes can be subdivided into the following categories:

SL3151Ch04Frame Page 176 Thursday, September 12, 2002 6:11 PM

176

Six Sigma and Beyond

1. Point elements: An example of a point element is a lumped mass element or an element specifically created to represent a particular constraint or loading present at that point. 2. Line elements: Truss links, rods, beams, pipes, cables, rigid links, springs, and gaps are examples of line elements. This type of element is usually characterized by two grid points or nodes at each end. 3. Surface elements: Membranes, plates, shells and certain types of fluid and thermal elements fall into this category. The surface elements can be triangular or quadrilateral, and thin or thick; accordingly they are characterized by a connectivity of three or more grid points or nodes. 4. Solid elements: Examples of solid elements include wedges, prisms, cubes, parallelepipeds and three-dimensional fluid and thermal elements. Elements in this category are usually defined using six or more grid points or nodes. 5. Special purpose elements: Combinations of springs, gaps, dampers, electrical conductors, acoustic, fluid, magnetic, mass, superelement, crack tips, radiation links, etc., are included in this category. For example, commonly used elements in the automotive industry (body engineering) are: • • • • • •

TYPES

Beams Rigid links Thin plates — triangular and quadrilateral Solid elements Springs Gaps (contact or interface elements) OF

ANALYSES

There are many combinations of analyses one may perform with FEA as the driving tool. However, the two predominant types are nonlinear and dynamic. Using these types one may focus on specific analysis of — for example nonlinearities types such as: Geometric • Stress less than yield strength • Euler (elastic) buckling • Examples: quarter panel under jacking and towing; hood following front crash Material • Stress greater than yield strength or material is nonlinear elastic • Plastic flow • Examples: seat belt pull; door intrusion beam bending Combination of geometric and material • Stress is greater than yield strength and buckling takes place • Crippling • Examples: rails during crash; roof crush

SL3151Ch04Frame Page 177 Thursday, September 12, 2002 6:11 PM

Simulation

177

The reader should also recognize that combinations of these types exist as well, for example linear/static — the easiest and most economical. Most of the FEA applications involve this kind of analysis. Examples include joint stiffness and door sag. Nonlinear/static is less frequently used. Examples include door intrusion beam, roof crush, and seat belt pull. Linear/dynamic is rarely used. Examples include windshield wipers or latch mechanism. Nonlinear/dynamic is the most complex and most expensive. Examples include knee bolster crash, front crash, and rear crash. Let us look at these combinations a little more closely: Linear static analysis: This is the simplest form of analytical application and is used most frequently for a wide range of structures. The desired results are usually the stress contours, deformed geometry, strain energy distribution, unknown reaction forces, and design optimization. Typical examples are door sag simulation, margin/fit problems, joint stiffness evaluation, high stress location search for all components, spot weld forces, and thermal stresses. Euler buckling analysis: This analysis is also relatively simple to perform and is used to calculate critical buckling loads. Caution should be exercised when performing this analysis because it produces analytical results that are not conservative. In other words, the critical buckling load thus calculated is usually higher than the actual load that would be determined through testing. A typical application is hood buckling. Normal modes analysis: This is an extremely useful technique for determining the natural frequencies (eigenvalues) of components and also the corresponding eigenvectors which represent the modes of deformation. Strictly speaking, this category does not fall under dynamic analysis since the problem is not time dependent. Typical examples include instrument panels, total vehicle or component NVH evaluation, door inner panel flutter, and steering column shake. Nonlinear static analysis: In general, all nonlinear analysis requires advanced methodology and is not recommended for use by inexperienced analysts. Usually, a graduate degree or several graduate level courses in the theory of elasticity, plasticity, vibrations, and solid and fluid mechanics are required to understand nonlinear behavior. Nonlinear FEA tends to be as much an art as it is a science, and familiarity with the subject structure is essential. Typical examples are seat back distortion, door beam bending rigidity studies, underbody components such as front and rear rails and wheel housings, bumper design, and crush analysis of several components. Nonlinear dynamic analysis: This FEA category is the most advanced. It involves very complex ideas and techniques and has become practicable only due to the availability of super-high-speed computers. This class of analysis involves all the complexities of nonlinear static analysis as well as additional problems involved with iterative time step selection and contact simulation at impact. Typical applications are related to crash evaluation and energy management.

SL3151Ch04Frame Page 178 Thursday, September 12, 2002 6:11 PM

178

Six Sigma and Beyond

PROCEDURES INVOLVED

IN

FEA

The procedures involved in FEA include: 1. Problem definition: Specification of concerns and expected results 2. Planning of analysis: Making decisions regarding the applicability of FEA, which code to use, and the size and the type of model to be constructed 3. Digitizing: The translation of a drawing into line data that is available to the modeler 4. Modeling: Creating the desired finite element model as planned (Many sophisticated tools are available such as the PDGS-FAST system, PATRAN, and so on.) 5. Input of data: Creating, editing and storing a formatted data file that includes a description of the model geometry, material properties, constraints, applied loading, and desired output 6. Execution: Processing the input data in either the batch or the interactive mode through the finite element code residing on the computer system and receiving the output in the form of a printout and/or post-processor data 7. Interpretation of output: A study of the output to check the validity of the input parameters as well as the solution of the structural problem 8. Feasibility considerations: Utilizing the output to make intelligent technical decisions about the acceptability of the structural design and the scope for design enhancement 9. Parametric studies: Redesign using parametric variation (The easiest changes to study are those involving different gages, materials, constraints, and loading. Geometric changes require repetition of steps 3 through 8; the same is true about remodeling of the existing geometry.) 10. Design optimization: An iterative process involving the repetition of steps 3 through 9 to optimize the design from considerations of weight, cost, manufacturing feasibility, and durability

STEPS

IN

ANALYSIS PROCEDURE

The steps in the analysis procedure are: 1. Establish objective. 2. What type of analysis? What program? Statics Mechanical Loads • Forces • Displacements • Pressure • Temperatures

SL3151Ch04Frame Page 179 Thursday, September 12, 2002 6:11 PM

Simulation

3.

4.

5.

6.

179

Heat Transfer • Conduction • Convection • 1-D radiation Dynamics Mode frequency Mechanical load • Transient (direct or reduced) linear • Sinusoidal Shock spectra Heat transfer direct transient Special features Nonlinear • Buckling • Large displacement • Elasticity • Creep • Friction, gaps Substructuring What is minimum portion of system or structure required? Known forces or displacements at a point Allows for separation Structural symmetry Isolation through test data Cyclic symmetry What are loading and boundary conditions? Loading known Loading can be calculated from simplistic analysis Loading to be determined from test data Support of excluded part of system established on modeled portion Test data taken to establish stiffness of partial constraints Determine model grid. Choose element types. Establish grid size to satisfy cost versus accuracy criterion. Develop bulk data. Establish coordinate systems. Number node or order elements to minimize cost. Develop node coordinates and element connectivity description. Code load and B.C. description. Check geometry description by plotting.

OVERVIEW

OF

FINITE ELEMENT ANALYSIS — SOLUTION PROCEDURE

The process of FEA may be summarized with a flow chart of linear static structural analysis in seven steps. The steps are:

SL3151Ch04Frame Page 180 Thursday, September 12, 2002 6:11 PM

180

Six Sigma and Beyond

1. Represent continuous structure as a collection of discrete elements connected by node points. 2. Formulate element stiffness matrices from element properties, geometry, and material. 3. Assemble all element stiffness matrices into global stiffness matrix. 4. Apply boundary conditions to constrain model (i.e., remove certain degrees of freedom). 5. Apply loads to model (forces, moments, pressure, etc.). 6. Solve matrix equation {F} = [K]{u} for displacements. 7. Calculate element forces and stresses from displacement results.

INPUT

TO THE

FINITE ELEMENT MODEL

Once the user is satisfied with the model subdivision, the following classes of input data must be prepared to provide a detailed description of the finite element model to typical FEA software such as MSC/NASTRAN (1998): Geometry: This refers to the locations of grid points and the orientations of coordinate systems that will be used to record components of displacements and forces at grid points. Element connectivities: This refers to identification numbers of the grid points to which each element is connected. Element properties: Examples of element properties are the thickness of a surface element and the cross-sectional area of a line element. Each element type has a specific list of properties. Material properties: Examples of material properties are Young’s modulus, density, and thermal expansion coefficient. There are several material types available in MSC/NASTRAN. Each has a specific list of properties. Constraints: Constraints are used to specify boundary conditions, symmetry conditions, and a variety of other useful relationships. Constraints are essential because an unconstrained structure is capable of free-body motion, which will cause the analysis to fail. Loads and enforced displacements: Loads may be applied at grid points or within elements.

OUTPUTS

FROM THE

FINITE ELEMENT ANALYSIS

Once the data describing the finite element model have been assembled and submitted to the computer, they will be processed by a software package such as MSC/NASTRAN to produce information requested by the user. The classes of output data are: 1. Components of displacements at grid points 2. Element data recovery: stresses, strains, strain energy, and internal forces and moments 3. Grid point data recovery: applied loads, forces of constraint, and forces due to elements

SL3151Ch04Frame Page 181 Thursday, September 12, 2002 6:11 PM

Simulation

181

It is the responsibility of the user to verify the accuracy of the finite element analysis results. Some suggested checks to perform are: Generate plots to visually verify the geometry. Verify overall model response for loadings applied. Check input loads with reaction forces. Perform hand checks of results whenever possible. Review and check results. Plot deformation and stress contour. Check equilibrium and reaction forces. Check concentration region for fineness of grid (compare calculated stress distribution with assumed element distribution). Check peak deflection and/or stress for ballpark accuracy. Special note: How a structure actually behaves under loading is determined by four characteristics: (a) the shape of the structure, (b) the location and type of constraints that hold the structure in place, (c) the loads applied to the structure — their magnitude, location and direction, and (d) the characteristics of the materials that comprise the structure. For example, glass, steel, and rubber have significantly different characteristics and different stiffnesses.

ANALYSIS

OF

REDESIGNS

OF

REFINED MODEL

At this stage, generally a correlation is attempted even though it is very difficult and presents many potential problems. These problems are about 60% associated with analysis and 40% associated with the actual testing. Remember that correlations at this stage commonly (over 50 projects) may run from 5 to 30%. Obviously, the focus should be on testing and test-related correlation with real world usage. Items of concern should be: Loads: • Isolation of single component of assembly • Hard to put assumed load in controlled lab test (linear loads causing moments) Strain gages: • Gage locations and orientation • Single leg gages versus rosettes • Improper gage lead hookup Non-linearities: • Plasticity • Pin joint clearance • Bolted joints In a typical analysis, the related correlation issues/problems/concerns examples are:

SL3151Ch04Frame Page 182 Thursday, September 12, 2002 6:11 PM

182

Six Sigma and Beyond

• Mesh size (for localized stress concentration, isolate concentration region and refine mesh) • Element type • Load distribution and B.C. isolation • Input error/bad data • Weld details Common problems that may be encountered in the FEA are: • Part not to size • Misunderstanding or interpretation of results Therefore, to make sure that the FEA is worth the effort, the following steps are recommended: 1. Initially, take simple, well-isolated components, with simple well-defined loads. 2. Do not expect miracles. 3. Use a joint test/analysis program. It can improve the capabilities of each step and serves as a check on techniques. 4. Work together. This is the key. The test results supplement weakness of analysis and vice versa.

SUMMARY — FINITE ELEMENT TECHNIQUE: A DESIGN TOOL • • • • •

Proven tool — approximate but very accurate if applied properly. Fine enough grid to match true strain field. Need to know loads accurately. Are supports rigid? What spring stiffness? Do not let FEA become just a research tool searching for an absolute answer. Use in all stages of design cycle as relative comparison tool in conjunction with test. • FEA if nothing else forces someone to examine in detail a component design. • A check on geometry itself. • The experimenter must think in detail about loads and interaction with rest of system

EXCEL’S SOLVER Yet another simple simulation tool is found in the Tools (add in) category of the Excel software program. Its simplicity is astonishing, and the results may be indeed phenomenal. What is required is the transformation function. Once that is identified, then the experimenter defines the constraints and the rest is computed by Solver.

DESIGN OPTIMIZATION In dealing with DFSS, a frequent euphemism is “design optimization.” What is design optimization? Design optimization is a technique that seeks to determine an

SL3151Ch04Frame Page 183 Thursday, September 12, 2002 6:11 PM

Simulation

183

optimum design. By “optimum design,” we mean one that meets all specified requirements but with a minimum expense of certain factors such as weight, surface area, volume, stress, cost, and so on. In other words, the optimum design is one that is as efficient and as effective as possible. To calculate an optimum design, many methods can be followed. Here, however, we focus on the ANSYS program, as defined by Moaveni (1999), which performs a series of analysis-evaluation-modification cycles. That is, an analysis of the initial design is performed, the results are evaluated against specified design criteria, and the design is modified as necessary. This process is repeated until all specified criteria are met. Design optimization can be used to optimize virtually any aspect of the design: dimensions (such as thickness), shape (such as fillet radii), placement of supports, cost of fabrication, natural frequency, material property, and so on. Actually, any ANSYS item that can be expressed in terms of a parameter can be subjected to design optimization. One example of optimization is the design of an aluminum pipe with cooling fins where the objective is to find the optimum diameter, shape, and spacing of the fins for maximum heat flow. Before describing the procedure for design optimization, we will define some of the terminology: design variables, state variables, objective function, feasible and unfeasible designs, loops, and design sets. We will start with a typical optimization problem statement: Find the minimum-weight design of a beam of rectangular cross section subject to the following constraints: Total stress σ should not exceed σmax

[σ ≤ σmax]

Beam deflection δ should not exceed δmax

[δ ≤ δmax]

Beam height h is limited to hmax

[h ≤ hmax]

Design Variables (DVs) are independent quantities that can be varied in order to achieve the optimum design. Upper and lower limits are specified on the design variables to serve as “constraints.” In the above beam example, width and height are obvious candidates for DVs, since they both cannot be zero or negative, so their lower limit would be some value greater than zero. State Variables (SVs) are quantities that constrain the design. They are also known as “behavioral constraints” and are typically response quantities that are functions of the design variables. Our beam example has two SVs: σ(the total stress) and δ(the beam deflection). You may define up to 100 SVs in an ANSYS design optimization problem. The Objective Function is the quantity that you are attempting to minimize or maximize. It should be a function of the DVs, i.e., changing the values of the DVs should change the value of the objective function. In our beam example, the total weight of the beam could be the objective function (to be minimized). Only one objective function may be defined in a design optimization problem. A design is simply a set of design variable values. A feasible design is one that satisfies all specified constraints, including constraints on the SVs as well as constraints

SL3151Ch04Frame Page 184 Thursday, September 12, 2002 6:11 PM

184

Six Sigma and Beyond

on the DVs. If even one of the constraints is not satisfied, the design is considered infeasible. An optimization loop (or simply loop) is one pass through the analysis-evaluation-modification cycle. Each loop consists of the following steps: 1. Build the model with current values of DVs and analyze. 2. Evaluate the analysis results in terms of the SVs and objective function. 3. Modify the design by calculating new values of DVs. These new values are calculated by ANSYS and are used to define the new version of the model. At the end of each loop, new values of DVs, SVs, and the objective function are available and are collectively referred to as a design set (or simply set).

HOW TO DO DESIGN OPTIMIZATION Design optimization requires a thorough understanding of the concept of ANSYS parameters, which are simply user-named variables to which you can assign numeric values. The model must be defined in terms of parameters (which are usually the DVs), and results data must be retrieved in terms of parameters (for SVs and the objective function). The usual procedure for design optimization consists of six main steps: 1. 2. 3. 4.

Initialize the design variable parameters. Build the model parametrically. Obtain the solution. Retrieve the results data parametrically and initialize the state variable and objective function parameters. 5. Declare optimization variables and begin optimization. 6. Review and verify optimum results. Details of these steps are beyond the scope of this volume. However, the reader may find the information in Moaveni (1999).

UNDERSTANDING

THE

OPTIMIZATION ALGORITHM

Understanding the algorithm used by a computer program is always helpful, and this is particularly true in the case of design optimization. Perhaps one of the most important issues is the notion of approximation. For simple mathematical functions that are continuously differentiable, minima can be found by analytical techniques such as solving for points of zero slope. The mathematical relationship between an arbitrary objective function and the DVs, however, is generally not known, so the program has to establish the relationship by curve fitting. This is done by calculating the objective function for several sets of DV values (i.e., for several designs) and performing a least squares fit among the data points. The resulting curve (or surface) is called an approximation. Each optimization loop generates a new data point, and the objective function is updated. It is this approximation that is minimized, not the actual objective function.

SL3151Ch04Frame Page 185 Thursday, September 12, 2002 6:11 PM

Simulation

185

State variables are handled in the same manner. An approximation is generated for each state variable and updated at the end of each loop. (Because approximations are used for the objective function and SVs, the optimum design will be only as good as the approximations.)

CONVERSION

TO AN

UNCONSTRAINED PROBLEM

State variables and limits on design variables are used to constrain the design and make the optimization problem a constrained one. The ANSYS program converts this problem to an unconstrained optimization problem because minimization techniques for the latter are more efficient. The conversion is done by adding penalties to the objective function approximation to account for the imposed constraints. You can think of penalties as causing an upturn of the objective function approximation at the constraints. The ANSYS program uses extended interior penalty functions. (For more information on penalty functions see sources in the selected bibliography for this chapter.) The search for a minimum is then performed on the unconstrained objective function approximation using the Sequential Unconstrained Minimization Technique (SUMT), which is explained in most texts on engineering design and optimization.

SIMULATION AND DFSS In summary, simulation is of value in connection with DFSS because: Design problems are discovered sooner. • Shortens development time • Provides better overall quality • Permits early optimization of the design Build-and-test is supplemented by computer simulations. • Permits lower testing budgets • Shortens development time • Permits evaluation of alternative designs • Minimizes overdesign by evaluating early in cycle Therefore, with the aid of simulation we are capable of: • Less time spent designing • Less time spent testing • Less time spent changing Result: Better products … in less time… at a lower cost. And that is what DFSS is all about.

SL3151Ch04Frame Page 186 Thursday, September 12, 2002 6:11 PM

186

Six Sigma and Beyond

REFERENCES Buchanan G.R., Schaum’s Outline of Finite Element Analysis, McGraw-Hill Professional Publishing, New York, 1994. Cook, R., Finite Element Modeling for Stress Analysis, Wiley, New York, 1995. Moaveni, S., Finite Element Analysis: Theory and Applications with ANSYS, Prentice Hall, Upper Saddle River, NJ, 1999. Schaeffer, H.G., MSC/NASTRAN Primer: Static and Normal Modes Analysis, MSC, New York, 1998.

SELECTED BIBLIOGRAPHY Adams, V. and Askenazi, A., Building Better Products with Finite Element Analysis, OnWord Press, New York, 1998. Belytschko, T., Liu, W.K., and Moran, B., Nonlinear Finite Elements for Continua and Structures, Wiley, New York, 2000. Hughes, T.J.R., The Finite Element Method: Linear Static and Dynamic Finite Element Analysis, Dover Publications, New York, 2000. Malkus, D.S. et al., Concepts and Applications of Finite Element Analysis, 4th ed., Wiley, New York, 2001. Rieger, M. and Steele, J., Basic Course in FEA Modeling, Machine Design, June 6, 1981, pp. 7–8. Rieger, M. and Steele, J., Basic Course in FEA Modeling, Machine Design, July 9, 1981, pp. 8–10. Rieger, M. and Steele, J., Advanced Techniques in FEA Modeling, Machine Design, July 23, 1981, pp. 7–12. Shih, R., Introduction to Finite Element Analysis Using I-DEAS Master Series 7, Schroff Development Corp. Publications, New York, 1999. Zienkiewics, O.C. and Taylor, R.L., Finite Element Method: Volume 1, The Basis, ButterworthHeinsmann, London, 2000. Zienkiewics, O.C. and Taylor, R.L., Finite Element Method: Volume 2, Solid Mechanics, Butterworth-Heinsmann, London, 2000. Zienkiewics, O.C. and Taylor, R.L., Finite Element Method: Volume 3, Fluid Dynamics, Butterworth-Heinsmann, London, 2000.

SL3151Ch05Frame Page 187 Thursday, September 12, 2002 6:10 PM

5

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

When we talk about design for manufacturability/assembly (DFM/DFA or DFMA), we describe a methodology that is concerned with reducing the cost of a product through simplification of its design. In other words, we try to reduce the number of individual parts that must be assembled and ultimately, increase the ease with which these parts can be put together. By focusing on these two items we are able to: 1. Design for a competitive advantage 2. Design for manufacturability, assembliability 3. Design for testability, serviceability, maintainability, quality, reliability, work-in-process (wip), cost, profitability, and so on. This, of course, brings us to the objectives of DFM/DFA, which are: To maximize a. Simplicity of design b. Economy of materials, parts, and components c. Economy of tooling/fixtures, process, and methods d. Standardization e. Assembliability f. Testability g. Serviceability h. Integrity of product features To minimize a. Unique processes b. Critical, precise processes c. Material waste, or scrap due to process d. Energy consumption e. Generation of pollution, liquid or solid f. Waste g. Limited available materials, components, and parts h. Limited available, proprietary, or long lead time equipment i. Degree of ongoing product and production support 187

SL3151Ch05Frame Page 188 Thursday, September 12, 2002 6:10 PM

188

Six Sigma and Beyond

Producibility Trade-off

Trade-offs Reliability

Performance

a) Old Design Goals: Reliability Better performance

Life Cycle Costs

Trade-offs

Producibility

Trade-offs Reliability

Trade-offs

Trade-offs

Trade-offs

Performance

(b) New Design Goals: Balanced Design Low support cost Low acquisition cost

FIGURE 5.1 Trade-off relationships between program objectives (balance design).

Therefore, one may describe the DFM/DFA process as a common-sense approach consistent with the old maxim, “Get it done right the first time.” In DFM/DFA, we strive to get it done right the first time with the most practical and affordable methods in order to meet the customer’s expectations in terms of time, process, costs, value, needs, and wants. This approach is quite different from the old way of doing business. Figure 5.1 shows the old and new ways of design. So, in a formal way we can say that design for manufacturing and assembly is a way of focusing on designing the product with manufacturability and assembliability in mind, to ensure the product can be produced with an affordable manufacturing effort and cost and also, after the manufacturing process, to ensure that the original designed product reliability can be maintained, if not enhanced. This approach may seem time-consuming and not value added, but if we consider the possible alternatives available we can appreciate the benefit of any DFM/DFA initiative. For example, consider the following: • What good is the design, if nobody can produce it? • What good is the design, if nobody can produce it with an affordable effort (in terms of manufacturing cost, scrap, rework, production cycle/turn-around, wip, and so on)?

SL3151Ch05Frame Page 189 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

• • • • •

What What What What What

good good good good good

is is is is is

189

the product, if nobody can afford it? the product, if we cannot market it in time? the product, if it does not sell? the product, if it is not profitable? it, if it does not work?

By doing a DFM/DFA, we are able to take into consideration many inputs with the intent of optimizing the design in terms of the following characteristics: • Design/development lead time vs. marketing time • Customer needs/wants vs. field application/performance vs. engineering specifications • Production launch efforts • Manufacturing cost • Flexibility and obsolescence of process and equipment • Maintainability/serviceability of product • Profitability Specifically, we are looking for the: 1. DFA to minimize total product cost by targeting: a. Part count — the major product cost driver b. Assembly time c. Part cost d. Assembly process 2. DFM to minimize part cost by: a. Optimizing manufacturing process b. Optimizing material selection c. Evaluating tooling and fabrication strategies d. Estimating tooling costs

BUSINESS EXPECTATIONS AND THE IMPACT FROM A SUCCESSFUL DFM/DFA Perhaps one of the major reasons why we do a DFMA is that in the final analysis we expect tremendous results with a measurable impact in the organization. Typical expectations are: • • • • • • •

Product development time improvement by 50–75% Product design cost reduction by 25–50% Product liability improvement by 10–25% Product field performance chosen to customer’s needs/wants Product production launch time reduction by 25–50% Total manufacturing cost reduction by 25–75% Reduction or even elimination of additional tooling/fixture cost

SL3151Ch05Frame Page 190 Thursday, September 12, 2002 6:10 PM

190

Six Sigma and Beyond

• Reduction, if not total elimination, of the engineering change notice by 75–99% • Increase in engineering and technical personnel’s work morale, and also letting them feel and assume ownership • Ability to be competitive, be profitable, be successful The impact, of course, becomes obvious. The entire organization is impacted for the better — it becomes business focused. For example: marketing becomes focused on the customer; engineering becomes focused on design; and manufacturing becomes focused on process. Specifically, the impact may be in the following areas: • Product closer to what customer expects • Reduction of time to market • Enhanced product liability, not just from original product design point of view but also from a manufacturing process point of view • Improved profit margins by reducing product cost • Improved operating efficiency by reducing work-in-process • Enhanced return on assets • Reduced technical personnel turnover rate by improving group and individual satisfaction with the job/work • Making the organization be profitable Traditional Approach — In the past, product design/development, manufacturing process design/development, and equipment selection/capability assessment were typically discrete activities — a sequential and discrete approach. That approach may be shown as in Figure 5.2. New Way — In order to let the manufacturing process and equipment have a head start, all three activities of design, process, and equipment occur simultaneously — a simultaneous equipment approach. This is where DFMA can help. This process may be shown as in Figure 5.3. The business strategy here becomes a pursuit to articulate the: Customer needs, wants and expectations → product/process engineering specification by asking a series of specific questions such as: • • • • • • •

What is the voice of the customer (VOC)? What regulations have to be met? What is the relative importance of requirement? Which product characteristics impact the VOC? Which process characteristics impact the VOC? What price and profit margin impact to meet VOC? Are there delivery schedule impacts?

SL3151Ch05Frame Page 191 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

Product selection and development assessment

Design/development manufacturing process

191

Equipment selection and capability assessment Time

Marketing specification and function confirmation

Engineering product design

Mfg process design

Mfg production Quality inspection

Product to customer FIGURE 5.2 Sequential approach.

Product design/development

Manufacturing process design/development assessment

Equipment design capability

FIGURE 5.3 Simultaneous approach.

• Any competition? Targeted competitor? • Continuing improvement opportunity? • Future cost reduction opportunity to meet future customer price reduction demands? Figure 5.4 shows the modern way of addressing these concerns. The arrows between product and process indicate possible alternatives. For example, if we

SL3151Ch05Frame Page 192 Thursday, September 12, 2002 6:10 PM

192

Six Sigma and Beyond

Product alternative(s) Voice of the customer

Process alternative(s)

Business decision (cost and investment)

Manufacturing production and quality

Product

FIGURE 5.4 Tomorrow’s approach … if not today’s.

examine the producibility for a textile component, we could look at the following material considerations: • • • • •

Natural Synthetic Properties Processes Applications

On the other hand, if we were to evaluate the manufacturing process we might want to examine: • • • • • •

Pattern layout Cutting Sewing assembly Types Processes Characteristics

THE ESSENTIAL ELEMENTS FOR SUCCESSFUL DFM/DFA The very minimum requirements for a successful DFMA are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Form a charter that includes all key functions. Establish the product plan. Define product performance requirement. Develop a realistic, agreed upon engineering specification. Establish product’s character/features. Define product architectural structure. Develop a realistic, detailed project schedule. Manage the project — schedule, performance, and results. Make efforts to reduce costs. Plan for continuing improvement.

SL3151Ch05Frame Page 193 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

193

The details of some of these elements are outlined below: Form a DFMA charter With any charter there are two primary responsibilities: (a) to identify the roles and (b) to identify the functions. i. Roles A. Charter members — designer, manufacturing engineer, material/component engineer, product engineer, reliability/quality engineer, and purchasing. B. Team leader — program manager is a good candidate, but not necessary. Any one of the charter members can be an adequate team leader. Some companies/organizations assign an integrator to be the DFMA leader. ii. Charter’s functions A. Determining the character of the product, to see what it is and thus, what design and production methods are appropriate B. Subjecting the product to a product function analysis, so that all design decisions can be made with full knowledge of how the item is supposed to work and so that all team members understand it well enough to contribute optimally C. Carrying out a design for producibility, usability, and maintainability study to determine if these factors can be improved without impairing functionality D. Designing an assembly process appropriate to the product’s particular character (This involves creating a suitable assembly sequence, identifying subassemblies, control plan, and designing each part so that its quality is compatible with the assembly/manufacturing method.) E. Designing a factory system that fully involves workers in the production strategy, operates on adequate inventory, and is integrated with suppliers’/vendors’ capabilities and manufacturing processes Establish product’s character/feature • QFD approach • Value analysis • Effectiveness study on function and appearance/cosmetic • Product character risk assessment Define product architectural structure a. Functional block approach b. Hardware approach c. Software approach d. Component approach Develop a project schedule a. Agreed to by all functions on: • Tasks • Objectives

SL3151Ch05Frame Page 194 Thursday, September 12, 2002 6:10 PM

194

Six Sigma and Beyond

• Duration • Responsibility b. Specific performance test: • Function • Appearance • Durability c. Use project management techniques. d. Concentrate on the concept of getting it done right the first time, not only doing it right the first time. e. Focus on the high leverage items — get some encouraging news first. f. Locate and prioritize the resource. g. Management commitment. h. Individual commitment. Manage the DFMA project • Ensure regular and formal review of the status by charter members. • Regularly prepare and formalize executive reports; get feedback. • Ensure total team inputs and contributions, not only involvement. • Utilize proven tools/methodologies. • Make adjustment with team consensus. • Ensure adequate resources with proper priorities. • Control the progress of the project.

THE PRODUCT PLAN It is imperative that the following considerations, all of which have a major impact on the manufacturing process, must be discussed and resolved as early as possible in the design cycle: 1. Nature of program — crash program, perfect design, or some other alternative 2. Product design itself 3. Production volume 4. Product life cycle 5. Funding 6. Cost of goods sold Product Design The focuses of marketing, engineering, manufacturing, and business/finance are quite different, yet they all push for the same interest for the organization. Our task then is to make sure that we balance out the different interests and priorities among the four functions of an organization. How do we do that? To make a long story short: How to decide between a crash program and a perfect product? When we talk about perfect product we mean it from a definitional perspective. There is no such a thing as a perfect product, but because of the operating definition we choose, we can indeed call something a perfect product.

SL3151Ch05Frame Page 195 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

195

Criteria for Decision between Crash Program and Perfect Product There are three issues here: 1. Opportunity cost 2. Development risk 3. Manufacturing risk For a short life cycle product or a highly innovative product in a competitive environment that changes rapidly, a company must react quickly to each new product that enters the market. Getting the product to market fast is the name of the game. However, being fast to the market is no advantage if the company chooses inadequate technology, creates a product that cannot meet the potential customer’s wants/needs/expectations, designs a product that cannot be manufactured, or must set the price so high that nobody can afford the product. The opportunity cost of missing a fast-moving market window, the risk of entering a market with the wrong product, and the risk of introducing a product nobody can produce pulls managers in opposite directions. So, the choice of a crash program (CP) or a perfect product (PP) approach is a necessary step prior to any product design taking place. Two examples will make the point of a CP and a PP: Case #1 — Crash Program Company: IBM Product: Personal computer Environment: Forecasted annual growth rate of 60%. Competitors, i.e., Apple, Tandy are controlling market developments and are beginning to cut into IBM’s traditional office market. Analysis: Opportunity cost is high. Development cost is low ($10 million compared to IBM’s equity value of $18 billion). The technology of design and process are stable and internally available. Decision: Crash program approach — develop, design, manufacture, and market the product within 2 years. Approach details: Deviate the standard eight phases design procedure. Give the development team complete freedom in product planning; keep interference to a minimum; and allow the use of streamlined, relatively informational management system. Use a so-called zero procedure approach, focusing on development speed rather than risk reduction of product, manufacturing, and so on. Results: Introduce the product within 2 years. Customer acceptance is good. Cost overrun by 15%. Cost of goods sold is about 5% unfavorable to the original estimate. Market share is questionable. Long term effects — ??? (Does this sound familiar? Quite a few organizations take this approach and of course, they fail.)

SL3151Ch05Frame Page 196 Thursday, September 12, 2002 6:10 PM

196

Six Sigma and Beyond

Case #2 — Perfect Product Design Company: Boeing Product: Boeing 727 replacement aircraft (767) Environment: Replacement within ten years is inevitable (may be speeded up to 5 years). Competitor, i.e., Airbus, has started its design. A new mid-range aircraft may take 727 replacement market away due to the operating/fuel inefficiency, comfortability, and Environmental Protection Agency (EPA) restraints. Analysis: Opportunity cost is high. (There is a need for 200–300 seat market; 727 is becoming obsolete.) Development cost is high (estimated $1.5 billion compared to entire company equity of $1.4 billion). Development and manufacturing risk is high. Technology and customer preferences are predictable but not yet crystallized. (Should it have two engineers or three? Should its cockpit allow for two people or three? Cruise range? Fuel consumption? Pricing?) Decision: Perfect product design approach. Complete the development of all new technologies of design and manufacturing processes in the early stages of research and development (R and D). Test everything in sight, and move product to launch only when success is nearly guaranteed. Eight-year design lead time. Approach details: Form an R and D team of 400 engineers/managers that includes designer, manufacturing engineer, quality, purchasing, and marketing. (The team member number goes up to 1000 right before go-ahead.) Apply concurrent engineering and DFMA process fully in the product R and D stage. Results: Introduce the 767 on schedule (which compares to Airbus’ 310 eight months behind schedule). Although Boeing had missed the 300–350 seat market and lost some of the 727 replacement market to Airbus 300, Boeing got to keep 200–300 seat market with a successful 767. Development costs were within budget and cost of goods sold was 4% favorable to the original estimates. No recall record so far. Long term effects — likely good. Most likely you are the in-betweens. The other approaches (see Figure 5.5) include: • • • •

Quantum leap — parallel program Acquisition Joint venture Leapfrog (Purchase a facility to maintain and manufacture current technology/design. Focus R and D on next generation technology/design.)

The Product Plan — Product Design Itself Product design has dedicated (whether one wants to admit it or not) the future of the product. About 95% of the material costs and 85% of the design/labor and

SL3151Ch05Frame Page 197 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

Crash program

Acquisition

197

Leapfrog exit

Step-by-step design approach

Joint venture

Opportunity cost

Development risk and manufacturing risk

FIGURE 5.5 The product development map/guide.

overhead costs are controlled by the design itself. Once the design is complete, about 85% of the manufacturing process has been locked in. Design-related factors affecting the manufacturing process include: • • • • • • •

Product size/weight Reliability/quality requirement Architectural structure Fastener/joint methods Parts/components/materials Size, shape, and weight of parts/components Appearance/cosmetic requirement

Other factors affecting the manufacturing process include: • • • • • • • •

Floor space Material flow and process flow Power, compressed air, a/c and heating, and facility Quality plan Manual operation mandatory Mechanized operation or automation operation mandatory System interfacing requirement Manufacturing process concepts/philosophy — cpf vs. in-line vs. batch vs. cellar approaches

SL3151Ch05Frame Page 198 Thursday, September 12, 2002 6:10 PM

198

Six Sigma and Beyond

• Management commitment • Production volume Volume requirements have the major influence on the choice of the manufacturing process. • Product life cycle As with volume requirements, product life has a significant influence on the manufacturing process. • Funding Since most of mechanization and automation are heavily capitalized, funding plays a major role in determining the product plan, which has a significant influence on the manufacturing process. • Cost of goods sold What is affordable capital/tooling/fixture amortization? What is the targeted cost of goods sold? Define Product Performance Requirement Minimum requirements are the collection and understanding of the following information: • • • •

Customer wants vs. customer needs vs. customer expectations Field condition and environment Performance standards Durability

The result of this understanding will facilitate the development of realistic and agreed upon specification(s). Some of the specific items that will guide realistic specifications are: • • • • • • • •

Engineering interpretation of customer needs Correlation between engineering specification and product specification Reliability study in terms of MTBF Manufacturing process reliability assessment in terms of maintaining original designed product standard Manufacturing cost assessment Option structure Control plan Qualification plan

AVAILABLE TOOLS AND METHODS FOR DFMA Infinite tools and methodologies may be used to accomplish the goal of a DFMA program. However, all of them fall into two categories: (a) approach alternatives and (b) mechanics. Some of the most important ones are listed below: Approach alternatives:

SL3151Ch05Frame Page 199 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

1. 2. 3. 4. 5. 6.

199

Ongoing program/project manager approach Manufacturing engineering sign-off approach Design engineering use simulation software package approach Simultaneous engineering approach Concurrent engineering approach Integrator approach

Mechanics: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Quality function development (QFD) Design of experiments (DOE) Potential failure mode and effects analysis (FMEA) Value engineering and value analysis (VE/VA) Group technology (GT) Geometric dimensioning and tolerancing (GD&T) Dimensional assembly analysis (DAA) Process capability study (Cpk, Ppk, Cp, Cr, ppm indices) Just-in-time (JIT) Qualitative assembly analysis (QAA)

COOKBOOKS

FOR

DFM/DFA

There are no cookbooks for DFMA. However, three organized instruction manuals may be close to most engineers’ terms of guidelines. They are: 1. Mitsubishi method 2. U-MASS method 3. MIL-HDB-727 design guidance for producibility All of the above methods utilize the principles of Taylor’s motion economy, which have been proven to be quite helpful, especially in the DFA area. We identify some of these principles here that may be profitably applied to shop and office work alike. Although not all are applicable to every operation, they do form a basis or a code for improving efficiency and reducing fatigue in manual work: 1. Smooth continuous curved motions of the hands are preferable to straightline motions involving sudden and sharp changes in direction. 2. Ballistic movements are faster, easier, and more accurate than restricted (fixation) or “controlled” movements. 3. Work should be arranged to permit easy and natural rhythm wherever possible. Use of the Human Body 4. The two hands should begin as well as complete their motions at the same time.

SL3151Ch05Frame Page 200 Thursday, September 12, 2002 6:10 PM

200

Six Sigma and Beyond

5. The two hands should not be idle at the same time except during rest periods. 6. Motions of the arms should be made in opposite and symmetrical directions and should be made simultaneously. 7. Hand and body motions should be confined to the lowest classification with which it is possible to perform the work satisfactorily. 8. Momentum should be employed to assist the worker wherever possible, and it should be reduced to a minimum if it must be overcome by muscular effort. 9. Eye fixations should be as few and as close together as possible. Arrangement of the Work Place 10. There should be a definite and fixed place for all tools and materials. 11. Tools, materials, and controls should be located close to the point of use. 12. Gravity feed bins and containers should be used to deliver material close to the point of use. 13. Drop deliveries should be used wherever possible. 14. Materials and tools should be located to permit the best sequence of motions. 15. Provisions should be made for adequate conditions for seeing. Good illumination is the first requirement for satisfactory visual perception. 16. The height of the work place and the chair should preferably be arranged so that alternate sitting and standing at work are easily possible. 17. A chair of the type and height to permit good posture should be provided for every worker. Design of Tools and Equipment 18. The hands should be relieved of all work that can be done more advantageously by a jig, a fixture, or a foot-operated device. 19. Two or more tools should be combined wherever possible. 20. Tools and materials should be pre-positioned whenever possible. 21. Where each finger performs some specific movement, such as in typewriting, the load should be distributed in accordance with the inherent capacities of the fingers. 22. Levers, crossbars, and hand wheels should be located in such positions that the operator can manipulate them with the least change in body position and with the greatest mechanical advantage.

MITSUBISHI METHOD The Mitsubishi method was developed and fine-tuned by Japanese engineers in Mitsubishi’s Kobe shipyard. The primary principle is the combination of QFD and Taylor’s motion economy. The Mitsubishi method is very popular in Japan’s heavy industries, i.e., shipbuilding industry, steel industry, and heavy equipment industry. There is also evidence of some application of this method in Japan’s automotive,

SL3151Ch05Frame Page 201 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

201

motorcycle, and office equipment industries. More efforts are needed to promote and share these techniques, and some effort is needed to fine-tune the Mitsubishi method and make it practical to fit U.S. manufacturing companies’ cultures and traditions. The process is based on the following principles: • The Mitsubishi method focuses on the product design’s reflection of the customer’s desires and tastes. Thus, marketing people, design engineers, and manufacturing staff must work together from the time a product is first conceived. • The Mitsubishi method is a kind of conceptual map that provides the means for inter-functional planning and communications. People with different problems and responsibilities can thrash out design priority while referring to patterns of evidence on the house’s grid. • The method involves 12 steps for each part in design/manufacturing, as follows: 1. Customer attributes (CA) analysis — also called voice of customer (VOC) evaluation — is performed. 2. Relative-importance weights of CA are determined. 3. Data is collected on customer evaluations of competitive products. 4. Engineering characteristics tell how to change the product. 5. Relationship matrix shows how engineering decisions affect customer perceptions. 6. Objective measures evaluate competitive products. 7. Roof matrix facilitates engineering creativity. 8. QFD is finalized. 9. Parts development is based on manufacturing process planning and handling planning (i.e., start the basic manufacturing process with materials in liquid state, feeding raw materials with elevator feeder, handling the wip with center board hopper, and continuing the forthcoming sequential operation with carousel assembly machine). 10. Manufacturing process and handling operation are based on the principles of motion economy. 11. Process planning is guided by parts/component characteristics, which are based on engineering characteristics, and the latter are based on customer attributes (compare to step #9). 12. Integrator coordinates/controls the project. • Analysis procedure. • Continuing improvement: Voice of customer, design alternative, and process alternative continue to interface with each other. It is a dynamical situation — no ending improvement. • Software package. Table 5.1 shows an example of customer attributes (Cas) and bundles of CAs for a car door. An example of relative importance weights of customer attributes is

SL3151Ch05Frame Page 202 Thursday, September 12, 2002 6:10 PM

202

Six Sigma and Beyond

TABLE 5.1 Customer Attributes for a Car Door Primary

Good operation and use

Secondary

Tertiary

Easy to open and close

Easy to close from outside Stays open on a hill Easy to open from outside Does not kick back Easy to close from inside Easy to open from inside Does not leak in rain No road noise Does not leak in car wash No wind noise Does not drip water or snow when open Does not rattle Soft, comfortable In right position Material will not fade Attractive (non-plastic look) Easy to clean No grease from door Uniform gaps between matching panels

Isolation

Arm rest Interior trim Good appearance

Clean Fit

TABLE 5.2 Relative Importance of Weights Bundles

Customer Attributes

Easy to open and close door

Easy to close from outside Stays open on a hill Does not leak in rain No road noise A complete list totals

Isolation

Relative Importance 7 5 3 2 100%

shown in Table 5.2. An example of customer evaluations of competitive products is shown in Table 5.3.

U-MASS METHOD The U-MASS method is named for the University of Massachusetts, where it was developed by two professors, Geoffrey Boothroyd and Peter Dewhurst, and their graduate students. It is the most common DFM/DFA approach used in the U.S. The

SL3151Ch05Frame Page 203 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

203

TABLE 5.3 Customer’s Evaluations of Competitive Products Customer Attributes

Relative Importance

Customer Perceptions

Easy to close from outside Stays open on a hill Does not leak in rain No road noise

7 5 3 2

Worst Best 1 2 3 4 5 Worst Best 1 2 3 4 5 Comparison is based on individual attributes as compared to: Our car door Competitor A’s Competitor B’s And so on…

Bundles Easy to open and close door Isolation

A complete list totals

100%

primary principle is the conventional motion and time study, while keeping in mind the component counts and motion economy. This method is heavily promoted in academic communities or institute-related manufacturing companies located in the New England area, such as Digital Equipment Corp. and Westinghouse Electric Company. Other companies are using it as well, such as Ford Motor Co., DaimlerChrysler, and many others. Its appeal seems to be the availability of the software that may be purchased from Boothroyd and Dewhurst. (Some practitioners find the software very time-consuming in design efficiency calculation and believe that more work is needed to fine tune its efficiency, as well as make it more user friendly.) The process is based on the following principles: 1. Determine the theoretical minimum part count by applying minimum part criteria. 2. Estimate actual assembly time using DFA database. 3. Determine DFA Index by comparing actual assembly time with theoretical minimum assembly time. 4. Identify assembly difficulties and candidates for elimination that may lead to manufacturing and quality problems.

MIL-HDBK-727 This method was developed by the U.S. Army material command and published by the naval publications and forms center. The first edition was published in 1971, and the latest revision was published in April 1984. The primary principle is Taylor’s motion economy and some other design tools, i.e., DOE. This method is not too popular. Not many people know about it, and it is not used very much outside of the military. Some updates and revisions are needed to make it more practical to general manufacturing companies.

SL3151Ch05Frame Page 204 Thursday, September 12, 2002 6:10 PM

204

Six Sigma and Beyond

FUNDAMENTAL DESIGN GUIDANCE The core of the DFM/DFA process is to make sure that the design and assembly are planned in terms of: 1. 2. 3. 4. 5. 6.

Simplicity (as opposed to complexity) Standardization (commonality) Flexibility Capability Suitability Carryover

So, a designer designing a product should be cognizant of the effects on product design. Some of these are: • Materials selection is based on the targeted manufacturing process. • The forms/shapes of parts are based on the targeted transportation, handling, and parts feeding system. • Field environment can affect the production durability, which contributes variation to the components/parts as well as the manufacturing process. • Shelf life. • Operating life. • Product MTBF and MTBR. In the development of the primary design, consideration must be given to whether to start with a basic process or to start with secondary process with purchased raw or semi-raw materials. If the decision is to start with a basic process, then the next question will be — what kind of materials to start with? There are three options: 1. Start with materials in liquid state, i.e., casting. 2. Start with materials in plastic state, i.e., forging. 3. Start with materials in solid state, i.e., roll forming (sheet), extrusion (rod, sheet), electroforming (powder), automatic screw machine work (rod). If a secondary process is needed, either as a sequential operation of a basic process or a fresh starting point, consideration must be given to the selection of the most favorable forming and sizing operations. A number of factors relating to a given design that need to be considered include: 1. 2. 3. 4. 5. 6. 7.

The The The The The The The

shape desired characteristics of the materials tolerance required surface finish quantity to be produced average run size cost

SL3151Ch05Frame Page 205 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

205

The focus then of a product design is to: 1. Minimize parts/components: The fewer parts/components and the fewer manufacturing/assembly operations, the better, i.e., • Combine mating parts, unless isolation is needed. • Eliminate screws and loose pieces. Replace screws with snap-on parts or fasten rivet, if practical. If screws are a necessary evil, try to make them all the same type and size. • Do not use a screw to locate. Remember that a screw is a fastener. 2. Use common/popular components/parts: Off-the-shelf type components/parts usually are user friendly and less expensive. Tooling/setup charges also can be avoidable beyond the pilot try headache, i.e., • Use fasteners with common/popular/standard length and diameter. • Use common values of resistors, capacitors, diodes, etc. • Use standard color chip of paints and coatings, if possible. 3. Design the parts to be symmetrical: If you must use customized unique parts, try to design the parts to be symmetrical, and use a jigless assembly method, if at all possible, i.e., • Avoid internal orientations. • Design an external accentuated locating feature, if it cannot be internally symmetrical. 4. Design the parts to be self-aligned, self-locating, and self-locking, i.e., • Design locating pins and small snap protrusions on mating parts. • Chamfers and tapers. • Use mechanical entrapments and snap-on approach. • Connect necessary wires/harnesses directly and use locking connectors. • Make sure that parts are easy to grip. • Avoid flexible parts — the more rigid the part, the more easily handled and assembled. • Avoid cables, if practical. • Avoid complicated fastening process, if practical. (Special note: If screws must be used, remember these rules: • Shank to head ratio: l greater than or equal to 1.5; l greater than or equal to 1.8 if tube feed • Head design • Thread consideration: Tapped holes? Thread cutting screws? Thread forming screws? • Quality screws) 5. Design for simple or no adjustment at all: • Remember, adjustment is a non-value added operation. Minimum adjustment — if necessary — with one-hand operation should be at most. 6. Modularize sub-assembly design: • Modularize sub-assemblies. Assemble and test them prior to final assembly.

SL3151Ch05Frame Page 206 Thursday, September 12, 2002 6:10 PM

206

Six Sigma and Beyond

Manufacturing System Schematic

Input

Activities

Output

Design drawings Specifications, standards Requirements Materials

Manufacturing Controlling Planning Scheduling

Products

Constraints Personnel Policies Quality Control/Assurance Purchasing

FIGURE 5.6 Manufacturing system schematic.

THE MANUFACTURING PROCESS Figure 5.6 shows a schematic of a manufacturing system. There are four categories of manufacturing processes. They are: 1. Fabrication process — which can be further categorized as basic process, secondary process, or finishing process. Typical types are: • Single station • Continuous production flow • Pace production line • Manufacturing cell approach 2. Assembly process — which can be further categorized as manual assembly, mechanical assembly, automatic assembly, or computer-aided assembly. Typical types are: • Continuous transfer • Intermittent transfer • Indexing mechanisms • Operator-paced free-transfer machine 3. Inspection or quality control process • Inspection check point(s) 4. Material handling process • Conveyors • Tractors

SL3151Ch05Frame Page 207 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

207

• Fork lifts • Parts/component feeding system: • Vibratory bowl feeder • Reciprocating tube hopper feeder • Centerboard hopper feeder • Reciprocating fork hopper feeder • External gate hopper feeder • Rotary disk feeder • Centrifugal hopper feeder • Revolving hook hopper feeder • Stationary hook hopper feeder • Bladed wheel hopper feeder • Tumbling barrel hopper feeder • Rotary centerboard hopper feeder • Magnetic disk feeder • Elevating hopper feeder • Magnetic elevating hopper feeder Approaches to manufacturing processes include the job shop approach, the assembly line approach, and the one in, one out approach. Details of these processes are as follows: Singled station manufacturing process — job shop approach Definition: Single fixture with one or more operations performed Advantages: • Capital investment — low • Line balance — not needed • Interference with other operations (downtime) — minimum, if any • Flexibility — easy to expand or rearrange • Employment fulfillment — high Disadvantages: • Multiple tooling/fixture investment — high • Material handling — high • Material flow — easy to congest at in/out • Operation cycle time — long • Operator skills — moderate Continuous production flow manufacturing process — assembly line approach Definition: Continuous, sequential motion assembly/manufacturing approach Advantages: • Work-in-process — low • Manufacturing/assembly cycle time — low • Material handling — very low, if not eliminated • Material flow — good • Operator skill/training — only in specialized areas

SL3151Ch05Frame Page 208 Thursday, September 12, 2002 6:10 PM

208

Six Sigma and Beyond

Disadvantages: • Capital investment — high • Preventative maintenance and corrective maintenance — absolute necessity (If one part breaks down, the entire line is down.) • Engineering, technician, and flow disciplines — absolute necessity • Flexibility — low • Production changeover — complicated Pace production line — one in, one out Definition: Same cycle time at all work stations, and likely all work pieces transfer at the same time Advantages: • Work-in-process — very low and can be calculated • Material handling — automatic • Material flow — good • Productivity — best Disadvantages: • Capital investment — high • Preventative maintenance and corrective maintenance — absolute necessity (If one part breaks down, the entire line is down.) • Engineering, technician, and flow disciplines — absolute necessity • Flexibility — very low • Production changeover — difficult

MISTAKE PROOFING DEFINITION Mistake proofing by definition is a process improvement system that prevents personal injury, promotes job safety, prevents faulty products, and prevents machine damage. It is also known as the Shingo method, Poka Yoke, error proofing, fail safe design, and by many other names.

THE STRATEGY Establish a team approach to mistake proof systems that will focus on both internal and external customer concerns with the intention of maximizing value. This will include quality indicators such as on-line inspection and probe studies. The strategy involves: • Concentrating on the things that can be changed rather than on the things that are perceived as having to be changed to improve process performance • Developing the training required to prepare team members • Involving all the appropriate people in the mistake proof systems process • Tracking quality improvements using in-plant and external data collection systems (before/after data)

SL3151Ch05Frame Page 209 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

209

• Developing a “core team” to administer the mistake proof systems process This core team will be responsible for tracking the status of the mistake proof systems throughout the implementation stages. • Creating a communication system for keeping plant management, local union committee, and the joint quality committee informed of all progress — as applicable • Developing a process for sharing the information with all other departments and/or plants — as applicable • Establishing the mission statement for each team and objectives that will identify the philosophy of mistake proof systems as a means to improve quality A typical mission statement may read: to protect our customers by developing mistake proofing systems that will detect or eliminate defects while continuing to pursue variation reduction within the process. • Developing timing for completion of each phase of the process • Establishing cross-functional team involvement with your customer(s) Typical objectives may be to: • Become more aware of quality issues that affect our customer • Focus our efforts on eliminating these quality issues from the production process • Expose the conditions that cause mistakes • Understand source investigation and recognize its role in preventing defects • Understand the concepts and principles that drive mistake prevention • Recognize the three functional levels of mistake proofing systems • Be knowledgeable of the relationships between mistake proof system devices and defects • Recognize the key mistake proof system devices • Share the mistake proof system knowledge with all other facilities within the organization

DEFECTS Many things can and often do go wrong in our ever-changing and increasingly complex work environment. Opportunities for mistakes are plentiful and often lead to defective products. Defects are not only wasteful but result in customer dissatisfaction if not detected before shipment. The philosophy behind mistake proof systems suggests that if we are going to be competitive and remain competitive in a world market we cannot accept any number of defects as satisfactory. In essence, not even one defect can be tolerated. Mistake proof systems are a simple method for making this philosophy become a daily practice. Simple concepts and methods are used to accomplish this objective.

SL3151Ch05Frame Page 210 Thursday, September 12, 2002 6:10 PM

210

Six Sigma and Beyond

Humans tend to be forgetful, and as a result, we make mistakes. In a system where blame is practiced and people are held accountable for their mistakes and mistakes within the process, we discourage the worker and lower morale of the individual, but the problem continues and remains unsolved.

MISTAKE PROOF SYSTEM IS IN THE WORKPLACE

A

TECHNIQUE

FOR

AVOIDING ERRORS

The concept of error proof systems has been in existence for a long time, only we have not attempted to turn it into a formalized process. It has often been referred to as idiot proofing, goof proofing, fool proofing, and so on. These terms often have a negative connotation that appears to attack the intelligence of the individual involved and therefore are not used in today’s work environment. For this reason we have selected the term “mistake proof system.” The idea behind a mistake proof system is to reduce the opportunity for human error by taking over tasks that are repetitive or actions that depend solely upon memory or attention. With this approach, we allow the worker to maintain dignity and self-esteem without the negative connotation that the individual is an idiot, goof, or fool.

TYPES

OF

HUMAN MISTAKES

Forgetfulness There are times when we forget things, especially when we are not fully concentrating or focusing. An example that can result in serious consequences is the failure to lock out a piece of equipment or machine we are working on. To preclude this, precautionary measures can be taken: post lock out instructions at every piece of equipment and/or machine; have an ongoing program to continuously alert operators of the danger. Mistakes of Misunderstanding Jumping to conclusions before we are familiar with the situation often leads to mistakes. For example, visual aids are often prepared by engineers who are thoroughly familiar with the operation or process. Since the aid is completely clear from their perspective, they may make the assumption (and often do) that the operator fully understands as well. This may not be true. To preclude this, we may test this hypothesis before we create an aid; provide training/education; standardize work methods and procedures. Identification Mistakes Situations are often misjudged because we view them too quickly or from too far away to clearly see them. One example of this type of mistake is misreading the identification code on a component of a piece of equipment and replacing that component with the wrong part. To prevent these errors, we might improve legibility

SL3151Ch05Frame Page 211 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

211

of the data/information; provide training; improve the environment (lighting); reduce boredom of the job, thus increasing vigilance and attentiveness. Amateur Errors Lack of experience often leads to mistakes. Newly hired workers will not know the sequence of operations to perform their tasks and often, due to inadequate training, will perform those tasks incorrectly. To prevent amateur errors, provide proper training; utilize skill building techniques prior to job assignment; use work standardization. Willful Mistakes Willful errors result when we choose to ignore the rules. One example of this type of error is placing a rack of material outside the lines painted on the floor that clearly designate the proper location. The results can be damage to the vehicle or the material or perhaps an unsafe work condition. To prevent this situation, provide basic education and/or training; require strict adherence to the rules. Inadvertent Mistakes Sometimes we make mistakes without even being aware of them. For example, a wrong part might be installed because the operator was daydreaming. To minimize this, we may standardize the work, through discipline if necessary. Slowness Mistakes When our actions are slowed by delays in judgment, mistakes are often the result. For example, an operator unfamiliar with the operation of a fork lift might pull the wrong lever and drop the load. Methods to prevent this might be: skill building; work standardization. Lack of Standards Mistakes Mistakes will occur when there is a lack of suitable work standards or when workers do not understand instructions. For example, two inspectors performing the same inspection may have different views on what constitutes a reject. To prevent this, develop operation definitions of what the product is expected to be that are clearly understood by all; provide proper training and education. Surprise Mistakes When the function or operation of a piece of equipment suddenly changes without warning, mistakes may occur. For example, power tools that are used to supply specific torque to a fastener will malfunction if an adequate oil supply is not maintained in the reservoir. Errors such as these can often be prevented by work standardization; having a total productive maintenance system in place.

SL3151Ch05Frame Page 212 Thursday, September 12, 2002 6:10 PM

212

Six Sigma and Beyond

Intentional Mistakes Mistakes are sometimes made deliberately by some people. These fall in the category of sabotage. Disciplinary measures and basic education are the only deterrents to these types of mistakes. There are many reasons for mistakes to happen. However, almost all of these can be prevented if we diligently expend the time and effort to identify the basic conditions that allow them to occur, such as: • When they happen • Why they happen and then determine what steps are needed to prevent these mistakes from recurring — permanently. The mistake proof system approach and the methods used give you an opportunity to prevent mistakes and errors from occurring.

DEFECTS

AND

ERRORS

Mistakes are generally the cause of defects. Can mistakes be avoided? To answer this question requires us to realize that we have to look at errors from two perspectives: 1. Errors are inevitable: People will always make mistakes. Accepting this premise makes one question the rationale of blaming people when mistakes are committed. Maintaining this “blame” attitude generally results in defects. Also, quite often errors are overlooked when they occur in the production process. To avoid blame, the discovery of defects is postponed until the final inspection, or worse yet, until the product reaches the customer. 2. Errors can be eliminated: If we utilize a system that supports (a) proper training and education and (b) fostering the belief that mistakes can be prevented, then people will make fewer mistakes. This being true, it is then possible that mistakes by people can be eliminated. Sources of mistakes may be any one of the six basic elements of a process: 1. 2. 3. 4. 5. 6.

Measurement Material Method Manpower Machinery Environment

Each of these elements may have an effect on quality as well as productivity. To make quality improvements, each element must be investigated for potential mistakes of operation. To reduce defects, we must recognize that defects are a

SL3151Ch05Frame Page 213 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

213

TABLE 5.4 Examples of Mistakes and Defects Mistake

Resulting Defects

Failure to put gasoline in the snow blower Failure to close window of unit being tested Failure to reset clock for daylight savings time Failure to show operator how to properly assemble components Proper weld schedule not maintained on welding equipment Low charged battery placed in griptow

Snow blower will not start Seats and carpet are wet Late for work Defective or warped product Bad welds, rejectable and/or scrap material Griptow will not pull racks resulting in lost production, downtime, etc.

consequence of the interaction of all six elements and the actual work performed in the process. Furthermore, we must recognize that the role of inspection is to audit the process and to identify the defects. It is an appraisal system and it does nothing for prevention. Product quality is changed only by improving the quality of the process. Therefore, the first step toward elimination of defects is to understand the difference between defects and mistakes (errors): Defects are the results. Mistakes are the causes of the results. Therefore, the underlying philosophy behind the total elimination of defects begins with distinguishing between mistakes and defects. Examples of mistakes and defects are shown in Table 5.4.

MISTAKE TYPES

AND

ACCOMPANYING CAUSES

The following categories with the associated potential causes are given as examples, rather than exhaustive lists: Assembly mistakes Inadequate training Symmetry (parts mounted backwards) Too many operations to perform Multiple parts to select from with poor or no identification Misread or unfamiliar with parts/products Tooling broken and/or misaligned New operator Processing mistakes Part of process omitted (inadvertent/deliberate) Fixture inadequate (resulting in parts being set into incorrectly) Symmetrical parts (wrong part can be installed)

SL3151Ch05Frame Page 214 Thursday, September 12, 2002 6:10 PM

214

Six Sigma and Beyond

Irregular shaped/sized part (vendor/supplier defect) Tooling damaging part as it is installed Carelessness (wrong part or side installed) Process/product requirements not understood (holes punched in wrong location) Following instructions for wrong process (multiple parts) Using incorrect tooling to complete operations (impact versus torque wrench) Inclusion of wrong part or item Part codes wrong/missing Parts for different products/applications mixing together Similar parts confused Misreading prints/schedules/bar codes etc. Operations mistakes Process elements assigned to too many operators Operator error Consequential results Setup mistakes Improper alignment of equipment Process or instructions for setup not understood or out of date Jigs and fixtures mislocated or loose Fixtures or holding devices will accept mislocated components Assembly omissions — missing parts Special orders (high or low volume parts missing) No inspection capability (hidden parts omitted) Substitutions (unexpected deviations from normal production) Misidentified build parameters (heavy duty versus standard) Measurement or dimensional mistakes Flawed measuring device Operator skill in measuring Inadequate system for measuring Using “best guess” system Processing omissions Operator fatigue (part assembled incorrectly/omitted) Cycle time (incomplete/poor weld) Equipment breakdown (weld omitted) New operator Tooling omitted Automation malfunction Instructions for operation incomplete/missing Job not set up for changeover Operator not trained/improper training Sequence violation Mounting mistakes Symmetry (parts can be installed backwards) Tooling wrong/inadequate

SL3151Ch05Frame Page 215 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

215

Operator dependency (parts installed upside down) Fixtures or holding devices accept mispositioned parts Miscellaneous mistakes Inadequate standards Material misidentified No controls on operation Counting system flawed/operating incorrectly Print/specifications incorrect

SIGNALS

THAT

ALERT

Signals that “alert” are conditions present in a process that commonly result in mistakes. Some signals that alert are: • • • • • • • • • • • •

Many parts/mixed parts Multiple steps needed to perform operation Adjustments Tooling changes Critical conditions Lack of or ineffective standards Infrequent production Extremely high volume Part symmetry Asymmetry Rapid repetition Environmental • Housekeeping • Material handing • Poor lighting • Foreign matter and debris • Other

Ten of the most common types of mistakes are: Assembly mistakes Inclusion of wrong part or item Setup mistakes Measurement mistakes Mounting mistakes

APPROACHES

TO

Processing mistakes Operations mistakes Assembly omissions (missing parts) Process omissions Miscellaneous

MISTAKE PROOFING

As we already mentioned, any mistake proofing system is a process that focuses on producing zero defects by eliminating the human element from assembly. There are two approaches to this — see Figure 5.7.

SL3151Ch05Frame Page 216 Thursday, September 12, 2002 6:10 PM

216

Six Sigma and Beyond

Operation #1

Operation #2

Ship to Customer

Reactive Systems

Proactive Systems

Focus on defect identification Alerts (signals) operator that failure has occurred Provides immediate feedback to operator Points to area of cause of defect Points to apparent cause (symptom of defect stops production until defective item removed or repaired) Protects customer from receiving defective product Does not prevent mistakes or defects from recurring

Focus on defect prevention Utilizes source inspection to detect when a mistake is about to occur before a defect is produced Halts production before mistake occurs Utilizes ideal Mistake Proofing Methods that eliminate the possibility of mistakes so that defective product cannot be produced Performs 100% inspection without inspection costs Prevents defects and mistakes from occurring

FIGURE 5.7 Approaches to mistake proofing.

1. Reactive systems (defect detection) This approach relies on halting production in order to sort out the good from the bad for repair or scrap. 2. Proactive systems (defect prevention) This approach seeks to eliminate mistakes so that defective products are not produced, production downtime is reduced, costs are lowered, and customer satisfaction is increased. Major Inspection Techniques Figure 5.8 shows major inspection techniques. Source inspection utilizing mistake proofing system devices is the most logical method of defect prevention. Mistake Proof System Devices Mistake proof system “devices” are simple and inexpensive. There are essentially two types of devices used: 1. Detectors (sensors) — to detect mistakes that have occurred or are about to occur 2. Preventers — to prevent mistakes from occurring

SL3151Ch05Frame Page 217 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

Operation #1

Operation #2

Source Inspection A defect is a result of a mistake. Source inspection looks at the cause(s) for the mistake, rather than the actual defect. By conducting inspection at the source, mistakes can be corrected before they become defects. Inspection utilizing Mistake Proofing System Devices to automatically inspect for mistakes or defective operating conditions is an effective low-cost solution for eliminating defects and resulting defective product. Informative Inspection Looks at the cause(s) of defects and feeds this information back to the appropriate personnel/process so that defects can be reduced/eliminated

217

Ship to Customer Inspect Finished Product Sort “good” from “bad” BAD

Scrap

GOOD

Repair

Judgment Inspection Distinguishes good product from bad. This method prevents defective product from being delivered to the customer but: Does nothing to prevent production of defective products

FIGURE 5.8 Major inspection techniques.

Devices Used as “Detectors of Mistakes” When used as detectors (sensors), these devices: 1. Provide prompt feedback (signals) to the operator that a mistake has occurred or is about to occur 2. Initiate an action or actions to prevent further mistakes from occurring Devices Used as “Preventers of Mistakes” When used to prevent mistakes, these devices prevent mistakes from occurring or initiate an action or actions to prevent mistakes from occurring.

SL3151Ch05Frame Page 218 Thursday, September 12, 2002 6:10 PM

218

Six Sigma and Beyond

Operation #1

Operation #2

First Function Eliminates the mistake at the source before it occurs

Ship to customer Third Function Detects a defect that has occurred before it is sent to the next operation or shipped to the customer

Second Function Detects mistakes as they are occurring, but before they result in defects

FIGURE 5.9 Functions of mistake-proofing devices.

EQUATION

FOR

SUCCESS

To be successful with a mistake proofing initiative one must keep in mind the following equation: Source investigation + Mistake proofing = Defect free system However, to reach the state of defect free system, in addition to signals and inspection we must also incorporate appropriate sensors to identify, stop, and/or correct a problem before it goes to the next operation. Sensors are very important in mistake proofing, so let us look at them little closer. A sensor is an electrical device that detects and responds to changes in a given characteristic of a part, assembly, or fixture — see Figure 5.9. A sensor can, for example, verify with a high degree of accuracy the presence and position of a part on an assembly or fixture and can identify damage or wear. Some examples of types of sensors and typical uses are: Welding position indicators: Determine changes in metallic composition, even on joints that are invisible to the surface Fiber sensors: Observe linear interruptions utilizing fiber optic beams Metal passage detectors: Determine if parts have a metal content or mixed metal content, for example in resin materials Beam sensors: Observe linear interruptions using electronic beams Trimetrons: Exclude or detect preset measurement values using a dial gauge (Value limits can be set on plus or minus sides, as well as on nominal values.) Tap sensors: Identify incomplete or missing tap screw machining Color marking sensors: Identify differences in color or colored marking

SL3151Ch05Frame Page 219 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

219

Area sensors: Determine random interruptions over a fixed area Double feed sensors: Identify when two products are fed at the same time Positioning sensors: Determine correct/incorrect positioning Vibration sensors: Identify product passage, weld position, broken wires, loose parts, etc. Displacement sensors: Identify thickness, height, warpage, surface irregularities, etc. Typical Error Proofing Devices Some of the most common mistake proofing devices used are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Sensors Sequence restrictors Odd part out method Limit or microswitches, proximity detectors Templates Guide rods or pins Stoppers or gates Counters Standardized methods of operation and/or material usage Detect delivery chute Critical condition indicators Probes Mistake proof your mistake proof system

and so on

REFERENCES Boothroyd, G. and Dewhurst, P., Product Design for Assembly, Boothroyd Dewhurst, Inc., Wakefield, RI, 1991. MIL-HDBK-727, Design Guidance for Producibility, U.S. Army Material Command, Washington, DC, 1986. Mitsubishi, Mitsubishi Design Engineering Handbook, Mitsubishi, Kobe, Japan, 1976. Munro, A., S. Munro and Associates, Inc., Design for Manufacture, training manual, 1992.

SELECTED BIBLIOGRAPHY Anon., How To Achieve Error Proof Manufacturing: Poka-Yoke and Beyond: A Technical Video Tutorial, SAE International, undated (may be ordered online for $895 [$25 preview copy]). Anon., 21st Century Manufacturing Enterprise Strategy, An Industry Led View, Volumes 1 and 2, Iacocca Institute, Lehigh University, PA, 1991. Anon., Mistake-Proofing for Operators: The ZQC System, The Productivity Press Development Team, Productivity Press, Portland, OR, 1997.

SL3151Ch05Frame Page 220 Thursday, September 12, 2002 6:10 PM

220

Six Sigma and Beyond

Anon., Manufacturing Management Handbook for Program Manager, ABN Fort Belvoir, VA, 1982. Anon., Product Engineering Design Mannual, Litton Industries, Beverly Hills, CA, 1978. Azuma, L. and Tada, M., A case history development of a foolproofing interface documentation system, IEEE Transactions on Software Engineering, 19, 765–773, 1993. Bandyopadhyay, J.K., Poka Yoke systems to ensure zero defect quality manufacturing, International Journal of Management, 10(1), 29–33, 1993. Barkers, R., Motion and Time Study: Design and Measurement of Work, Cot Loge Book Company, Los Angeles, 1970. Barkman, W.E., In-Process Quality Control for Manufacturing, Marcel Dekker, New York, 1989. (Preface and Chapter 3 are of particular interest.) Bayer, P.C., Using Poka Yoke (mistake proofing devices) to ensure quality, IEEE 9th Applied Power Electronics Conference Proceedings, 1, 201–204, 1994. Bodine, W.E., The Trend: 100 Percent Quality Verification, Production, June 1993, pp. 54–55. Bosa, R., Despite fuzzy logic and neural networks, operator control is still a must, CMA, 69, 7, 995. Boothroyd, G. and Murch, P., Automatic Assembly, Marcel Dekker, New York, 1982. Brehmer, B., Variable errors set a limit to adaptation, Ergonomics 33, 1231–1239, 1990. Brall, J., Product Design for Manufacturing, McGraw-Hill, New York, 1986. Casey, S., Set Phasers on Stun and Other True Tales of Design, Technology, and Human Error, Aegean, Santa Barbara, CA, 1993. Chase, R.B., and Stewart, D.M., Make Your Service Fail-safe, Sloan Management Review, Spring 1994, pp. 35–44. Chase, R.B. and Stewart, D.M., Designing Errors Out, Productivity Press, Portland, OR, 1995. Note of interest: Productivity Press has discontinued sales of this book (a very sad outcome). Some copies may be available in local bookstores. It is both more readable and broader in application than Shingo but does not have a catalog of examples as Shingo does. Damian, J., “Agile Manufacturing” Can Revive U.S. Competitiveness, Industry Study Says — A Modest Proposal, Electronics, Feb. 1992, pp. 34, 42–44. Dove, R., Agile and Otherwise — Measuring Agility: The Toll of Turmoil, Production, Jan. 1995, pp. 12–15. Dove, R., Agile and Otherwise — The Challenges of Change, Production, Feb. 1995, pp. 14–16. Gross, N., This Is What the U.S. Must Do To Stay Competitive, Business Week, Dec. 1991, pp. 21–24. Grout, J.R., Mistake-Proofing Production, working paper 75275–0333, Cox School of Business, Southern Methodist University, Dallas, 1995. Grout, J.R. and Downs, B.T., An Economic Analysis of Inspection Costs for Failsafing Attributes, working paper 95–0901, Cox School of Business, Southern Methodist University, Dallas, 1995. Grout, J.R. and Downs. B.T., Fail-safing and Measurement Control Charts, 1995 Proceedings, Decision Sciences Institute Annual Meetings, Boston, MA, 1995. Henricks, M., Make No Mistake, Entrepreneur, Oct. 1996, pp. 86–89. (Last quote should read “average net savings of around $2500 a piece...” not average cost.) Hinckley, C.M. and Barkan, P., The role of variation, mistakes, and complexity in producing nonconformities, Journal of Quality Technology 27(3), 242–249, 1995. Jaikumar, R., Manufacturing a’la Carte Agile Assembly Lines, Faster Development Cycles, 200 Years to CIM, IEEE Spectrum, 76–82, Sept. 1993.

SL3151Ch05Frame Page 221 Thursday, September 12, 2002 6:10 PM

Design for Manufacturability/ Assembly (DFM/DFA or DFMA)

221

Kaplan, G., Manufacturing a’la Carte Agile Assembly Lines, Faster Development Cycles, Introduction, IEEE Spectrum, 46–51, Sept. 1993. Kelly, K., Your Job Managing Error is Out of Control, Addison-Wesley, New York, 1994. Kletz, T., Plant Design for Safety: A User-Friendly Approach, Hemisphere Publishing Corp., New York, 1991. Lafferty, J.P., Cpk of 2 Not Good Enough for You? Manufacturing Engineering, Oct. 1992, p. 10. Ligus, R.G., Enterprise Agility: Jazz in the Factory, Industrial Engineering, Nov. 1994, pp. 19–23. Lucas Engineering and Systems Ltd., Design for Manufacture Reference Tables, University of Hull, Hull, England, Lucas Industries, Jan. 1994. Manji, J.F., Sharpen, C., Your Competitive Edge Today and Into the 21st Century, CALS El Journal, Date unknown, pp. 56–61. Marchwinski, C., Ed., Company Cuts the Risk of Defects During Assembly and Maintenance, MfgNet: The Manufacturer’s Internet Newsletter, Productivity, Inc. Norwalk, CT, 1996. Marchwinski, C., Ed., Mistake-proofing, Productivity, 17(3), 1–6, 1995. Marchwinski, C., Ed., (1997). SPC vs. ZQC. Productivity, 18(1), 1–4 1997. (Note: ZQC is another name for mistake proofing. It stands for Zero Quality Control.) McClelland, S., Poka-Yoke and the Art of Motorcycle Maintenance, Sensor Review, 9(2), 63, 1989. Monden, Y., Toyota Production System, Industrial Engineering and Management Press, Norcross, GA, 1983, pp. 10, 137–154. Munro, A., S. Munro and Associates, Inc., Design for Manufacture, training manual, 1994. Munro, A., S. Munro and Associates, Inc., Trainers for Design for Manufacture, analysis, undated. Myers, M., Poka/Yoke-ing Your Way to Success, Network World, Sept. 11, 1995, p. 39. Nakajo, T. and Qume, H., The principles of foolproofing and their application in manufacturing, Reports of Statistical Application Research, Union of Japanese Scientists and Engineers, 32(2), 10–29, 1985. Niebel, C. and Baldwin, J., Designing for Production, Irwin, Homewood, IL, 1963. Nieber, C. and Draper, G., Product Design and Process Engineering, McGraw-Hill, New York, 1974. Noaker, P.M., The Search for Agile Manufacturing, Manufacturing Engineering, Nov. 1994, pp. 57–63. Norman, D.A., The Design of Everyday Things, Doubleday, New York, 1989. O’Connor, L., Agile Manufacturing in a Responsive Factory, Mechanical Engineering, July 1994, pp. 43–46. Otto, K. and Wood, K., Product Design: Techniques in Reverse Engineering and New Product Development, Prentice Hall, Upper Saddle River, NJ, 2001. Port, O., Moving Past the Assembly Line — “Agile” Manufacturing Systems May Bring a U.S. Revival, Business Week/Re-Inventing America, 1992, pp. 17–20. Reason, J., Human Error, Cambridge University Press, New York, 1990. Robinson, A.G. and Schroeder, D.M., The limited role of statistical quality control in a zero defects environment, Production and Inventory Management Journal, 31(3), 60–65, 1990. Robinson, A.G., Ed., Modern Approaches to Manufacturing Improvement: The Shingo System, Productivity Press, Portland, OR, 1991. Shandle, J., Sandia Labs Launches Agile Factory Program, Electronics, Mar. 8, 1993, pp. 48–49.

SL3151Ch05Frame Page 222 Thursday, September 12, 2002 6:10 PM

222

Six Sigma and Beyond

Sheridan, J.H., A Vision of Agility, Industry Week, Mar. 21, 1994, pp. 22–24. Shingo, S., Zero Quality Control: Source Inspection and the Poka-Yoke System, Trans. A.P. Dillion, Productivity Press, Portland, OR, 1986. Shingo S., A Study of the Toyota Production System from an Industrial Engineering Viewpoint, Productivity Press, Portland, OR, 1989, online excerpts. Steven, S. and Bowen., H.K., Decoding the DNA of the Toyota Production System, Harvard Business Review, Sept./Oct, 1999, pp. 97–106. Texas Instruments, Design to Cost: An Introduction, Corporate Engineering Council, Texas Instruments, Inc., Dallas, 1977. Trucks, H.E., Designing for Economical Production, SME, Dearborn, MI, 1974. Tsuda, Y., Implications of fool proofing in the manufacturing process, in Quality Through Engineering Design, Kuo, W., Ed., Elsevier, New York, 1993. Vasilash, G.S., Re-engineering, Re-energizing, Objects and Other Issues of Interest, Production, Jan. 1995, pp. 38–41. Vasilash, G.S., On training for mistake-proofing, Production, Mar. 1995, pp. 42–44. Ward, C., What Is Agility? Industrial Engineering, Nov. 1994, pp. 38–44. Warm, J.S., An introduction to vigilance, in Sustained Attention in Human Performance, Warm, J.S., Ed., Wiley, New York, 1984. Weimer, G., Is an American Renaissance at Hand? Industry Week, May 1992, pp. 14–17. Weimer, G., U.S.A. 2006: Industry Leader or Loser, Industry Week, Jan. 20, 1992, pp. 31–34.

SL3151Ch06Frame Page 223 Thursday, September 12, 2002 6:09 PM

6

Failure Mode and Effect Analysis (FMEA)

This chapter has been developed to assist and instruct design, manufacturing, and assembly engineers in the development and execution of a potential Failure Mode and Effect Analysis (FMEA) for design considerations, manufacturing, assembly processes, and machinery. An FMEA is a methodology that helps identify potential failures and recommends corrective action(s) for fixing these failures before they reach the customer. A concept (system) FMEA is conducted as early as possible to identify serious problems with the potential concept or design. A design FMEA is conducted prior to production and involves the listing of potential failure modes and causes. An FMEA identifies actions required to prevent defects and thus keeps products that may fail or not be fit from reaching the customer. Its purpose is to analyze the product’s design characteristics relative to the planned manufacturing or assembly process to ensure that the resultant product meets customer needs and expectations. When potential failure modes are identified, corrective action can be initiated to eliminate them or continuously reduce their potential occurrence. The FMEA also documents the rationale for the manufacturing or assembly process involved. Changes in customer expectations, regulatory requirements, attitudes of the courts, and the industry’s needs require disciplined use of a technique to identify and prevent potential problems. That disciplined technique is the FMEA. A process FMEA is an analytical technique that identifies potential productrelated process failure modes, assesses the potential customer effects of the failures, identifies the potential manufacturing or assembly process causes, and identifies significant process variables to focus controls for prevention or detection of the failure conditions. (Also, process FMEAs can assist in developing new machine or equipment processes. The methodology is the same; however, the machine or equipment being designed would be considered the product.) A machinery FMEA is a methodology that helps in the identification of possible failure modes and determines the cause for and effect of these failures. The focus of the machinery FMEA is to eliminate any safety issues and to resolve them according to specified procedures between customer and supplier. In addition, the purpose of this particular FMEA is to review both design and process with the intent to reduce risk. All FMEAs utilize occurrence and detection probability in conjunction with severity criteria to develop a Risk Priority Number (RPN) for prioritization of corrective action considerations. This is a major departure in methodology from the Failure Mode and Critical Analysis (FMCA), which focuses primarily on the severity of the failure as a priority characteristic.

223

SL3151Ch06Frame Page 224 Thursday, September 12, 2002 6:09 PM

224

Six Sigma and Beyond

In its most rigorous form, an FMEA summarizes the engineer’s thoughts while developing a process. This systematic approach parallels and formalizes the mental discipline that an engineer normally uses to develop processing requirements.

DEFINITION OF FMEA FMEA is an engineering “reliability tool” that: 1. Helps to define, identify, prioritize, and eliminate known and/or potential failures of the system, design, or manufacturing process before they reach the customer, with the goal of eliminating the failure modes or reducing their risks 2. Provides structure for a cross-functional critique of a design or a process 3. Facilitates inter-departmental dialog (It is much more than a design review.) 4. Is a mental discipline “great” engineering teams go through, when critiquing what might go wrong with the product or process 5. Is a living document that reflects the latest product and process actions 6. Ultimately helps prevent and not react to problems 7. Identifies potential product- or process-related failure modes before they happen 8. Determines the effect and severity of these failure modes 9. Identifies the causes and probability of occurrence of the failure modes 10. Identifies the controls and their effectiveness 11. Quantifies and prioritizes the risks associated with the failure modes 12. Develops and documents action plans that will occur to reduce risk

TYPES OF FMEAS There are many types of FMEAs (see Figure 6.1). However, the main ones are: • System/Concept — S/CFMEA. These are driven by system functions. A system is an organized set of parts or subsystems to accomplish one or more functions. System FMEAs are typically done very early, before specific hardware has been determined. • Design — DFMEA. A design FMEA is driven by part or component functions. A design/part is a unit of physical hardware that is considered a single replaceable part with respect to repair. Design FMEAs are typically done later in the development process when specific hardware has been determined. • Manufacturing or Process — PFMEA. A process FMEA is driven by process functions and part characteristics. A manufacturing process is a sequence of tasks that is organized to produce a product. A process FMEA can involve fabrication as well as assembly.

SL3151Ch06Frame Page 225 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

225

Types of FMEA Design FMEA

Process FMEA

Component Subsystem System

System FMEA

Machinery FMEA Focus: Design changes to lower life cycle costs Objective: Improve the reliability and maintain ability of the machinery and equipment

Focus: Minimize failure effects on the system Objective: Maximize system quality, reliability cost, and maintain ability

Machines Methods Material Manpower Measurement Environment Focus: Minimize production process failure effects on the system Objective: Maximize the system quality, reliability, cost, maintain ability, and productivity

FIGURE 6.1 Types of FMEA.

• Machinery — MFMEA is driven by low volume machinery and equipment where large-scale testing is impractical prior to production and manufacture of the machinery and equipment. The MFMEA focuses on design changes to lower life cycle costs by improving the reliability and maintainability of the machinery and equipment. Note: Service, software, and environmental FMEAs are additional variations. However, in this chapter we will focus only on design, process, and machinery FMEAs. The other FMEAs follow the same rationale as the design and process FMEAs.

IS FMEA NEEDED? If any answer to the following questions is positive, then you need an FMEA: • • • • •

Are Are Are Are Are

customers becoming more quality conscious? reliability problems becoming a big concern? regulatory requirements harder to meet? you doing too much problem solving? you addicted to problem solving?

SL3151Ch06Frame Page 226 Thursday, September 12, 2002 6:09 PM

226

Six Sigma and Beyond

“Addiction” to problem solving is a very important consideration in the application of an active FMEA program. When the thrill and excitement of solving problems become dominant, your organization is addicted to problem solving rather than preventing the problem to begin with. A proper FMEA will help break your addiction by: • Reducing the percentage of time devoted to problem solving • Increasing the percentage of time in problem prevention • Increasing the efficiency of resource allocation Note: The emphasis is always on reducing complexity and engineering changes.

BENEFITS OF FMEA When properly conducted, product and process FMEAs should lead to: 1. Confidence that all risks have been identified early and appropriate actions have been taken 2. Priorities and rationale for product and process improvement actions 3. Reduction of scrap, rework, and manufacturing costs 4. Preservation of product and process knowledge 5. Reduction of field failures and warranty cost 6. Documentation of risks and actions for future designs or processes By way of comparison of FMEA benefits and the quality lever, Figure 6.2 may help. In essence, one may argue that the most important benefit of an FMEA is that it helps identify hidden costs, which are quite often greater than visible costs. Some of these costs may be identified through: 1. 2. 3. 4. 5.

Customer dissatisfaction Development inefficiencies Lost repeat business (no brand loyalty) High employee turnover And so on

FMEA HISTORY This type of thinking has been around for hundreds of years. It was first formalized in the aerospace industry during the Apollo program in the 1960s. The initial automotive adoption was in the 1970s in the area of safety issues. FMEA was required by QS-9000 and the advanced product quality planning process in 1994 for all automotive suppliers. It has now been adopted by many other industries.

SL3151Ch06Frame Page 227 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

227

Payback: Effort Product design fix 100:1

Process design fix 10:1 Production fix 1:1 Customer fix 1:10

Planning and definition Product design and development Mfg process design and development Product and process validation

Production

FIGURE 6.2 Payback effort.

INITIATION OF THE FMEA Regardless of the type, all FMEAs should be conducted as early as possible. FMEA studies can be carried out at any stage during the development of a product or process. However, the ideal time to start the FMEA is: • When new systems, designs, processes, or machines are being designed, but before they are finalized • When systems design, process, or machine modifications are being contemplated • When new applications are used for the systems, designs, processes, or machines • When quality concerns become visible • When safety issues are of concern Note: Once the FMEA is initiated, it becomes a living document, is updated as necessary, and is never really complete. Therefore: • “FMEA-type thinking” is central to reliability and continual improvement in products and manufacturing processes to remain competitive in our global marketplace. It must be understood that an FMEA conducted after production serves as a reactive tool, and the user has not taken full advantage of the FMEA process.

SL3151Ch06Frame Page 228 Thursday, September 12, 2002 6:09 PM

228

Six Sigma and Beyond

• A typical system FMEA should begin even before the program approval stage. The design FMEA should start right after program approval and continue to be updated through prototypes. A process FMEA should begin just before prototypes and continue through pilot build and sometimes into product launching. As for the MFMEA, it should also start at the same time as the design FMEA. It is imperative for a user of an FMEA to understand that sometimes information is not always available. During these situations, users must do the best they can with what they have, recognizing that the document itself is indeed a living document and will change as more information becomes available. • History has shown that a majority of product warranty campaigns and automotive recalls could have been prevented by thorough FMEA studies.

GETTING STARTED Just as with anything else, before the FMEA begins there are some assumptions and preparations that must be taken care of. These are: 1. 2. 3. 4.

Know your customers and their needs. Know the function. Understand the concept of priority. Develop and evaluate conceptual designs/processes based on your customer’s needs and business strategy. 5. Be committed to continual improvement. 6. Create an effective team. 7. Define the FMEA project and scope.

1. UNDERSTAND YOUR CUSTOMERS

AND

THEIR NEEDS

A product or a process may perform functions flawlessly, but if the functions are not aligned with the customer’s needs, you may be wasting your time. Therefore, you must: • Determine all (internal or external) relevant customers. • Understand the customer’s needs better than the customers understand their own needs. • Document the customer’s needs and develop concepts. For example, customers need: • Chewable toothpaste • Smokeless cigarettes • Celery-flavored gum • ???? In FMEA, a customer is anyone/anything that has functions/needs from your product or manufacturing process. An easy way to determine customer needs is to understand the Kano model — see Figure 6.3.

SL3151Ch06Frame Page 229 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

229

Satisfied Excitement needs Performance needs

Did not do it at all

Did it very well

Basic needs

Time

Dissatisfied FIGURE 6.3 Kano model.

The model facilitates understanding of all the customer needs, including: Excitement needs: Generally, these are the unspoken “wants” of the customer. Performance needs: Generally, these are the spoken “needs” of the customer. They serve as the neutral requirements of the customer. Basic needs: Generally, these are the unspoken “needs” of the customer. They serve as the very minimum of requirements. It is important to understand that these needs are always in a state of change. They move from basic needs to performance to excitement depending on the product or expectation, as well as value to the customer. For example: SYSTEM customers may be viewed as: other systems, whole product, government regulations, design engineers, and end user. DESIGN customers may be viewed as: higher assembly, whole product, design engineers, manufacturing engineers, government engineers, and end user. PROCESS customers may be viewed as: the next operation, operators, design and manufacturing engineering, government regulations, and end user. MACHINE customers may be viewed as: higher assembly, whole product, design engineers, manufacturing engineers, government regulations, and end user. Another way to understand the FMEA customers is through the FMEA team, which must in no uncertain terms determine: 1. Who the customers are 2. What their needs are 3. Which needs will be addressed in the design/process

SL3151Ch06Frame Page 230 Thursday, September 12, 2002 6:09 PM

230

Six Sigma and Beyond

The appropriate and applicable response will help in developing both the function and effects.

2. KNOW

THE

FUNCTION

The dictionary definition of a function is: The natural, proper, or characteristic action of any thing. This is very useful because it implies performance. After all, it is performance that we are focusing in the FMEA. Specifically, a function from an FMEA perspective is the task that a system, part, or manufacturing process performs to satisfy a customer. To understand the function and its significance, the team conducting the FMEA must have a thorough list of functions to evaluate. Once this is done, the rest of the FMEA process is a mechanical task. For machinery, the function may be analyzed through a variety of methodologies including but not limited to: • • • • • • • •

Describing the design intent either through a block diagram or a P-diagram Identifying an iterative process in terms of what can be measured Describing the ideal function — what the machine is supposed to do Identifying relationships in verb–noun statements — function tree analysis Considering environmental and safety conditions Accounting for all R & M parameters Accounting for the machine’s performance conditions Analyzing all other measurable engineering attributes

3. UNDERSTAND

THE

CONCEPT

OF

PRIORITY

One of the outcomes of an FMEA is the prioritization of problems. It is very important for the team to recognize the temptation to address all problems, just because they have been identified. That action, if taken, will diminish the effectiveness of the FMEA. Rather, the team should concentrate on the most important problems, based on performance, cost, quality, or any characteristic identified on an a priori basis through the risk priority number.

4. DEVELOP AND EVALUATE CONCEPTUAL DESIGNS/PROCESSES BASED ON CUSTOMER NEEDS AND BUSINESS STRATEGY There are many methods to assist in developing concepts. Some of the most common are: 1. 2. 3. 4.

Brainstorming Benchmarking TRIZ (the theory of inventive problem solving) Pugh concept selection (an objective way to analyze and select/synthesize alternative concepts)

SL3151Ch06Frame Page 231 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

231

Eval. Criteria Stubble length Pain level Mfg. Costs Price/Use Etc Etc

Razor D A T U M

A B

C D

E F

+ S

+ S S S

+ -

-

Totals

+ S

2 1 1

1 3

3 1

+ + S S

S S S +

+ S + -

1 3

1 2 1

4 2 2

G

H

Legend: Evaluation Criteria: These are the criteria that we are comparing the razor with the other approaches. Datum: These are the basic razor characteristics that we are comparing the other concepts to. A: Chemical D: Duct tape G: Straight edge

B: Electric E: Epilady H: ?

C: Electrolysis F: Laser beam

+ : Better than the basic razor requirement - : Worse than the basic razor requirement S : Same as the basic razor requirement

FIGURE 6.4 A Pugh matrix — shaving with a razor.

Figure 6.4 shows what a Pugh matrix may look like for the concept of “shaving” with a base that of a “razor.”

5. BE COMMITTED

TO

CONTINUAL IMPROVEMENT

Everyone in the organization and especially management must be committed to continual improvement. In FMEA, that means that once recommendations have been made to increase effectiveness or to reduce cost, defects, or any other characteristic, a proper corrective action must be developed and implemented, provided it is sound and it complements the business strategy.

6. CREATE

AN

EFFECTIVE FMEA TEAM

Perhaps one of the most important issues in dealing with the FMEA is that an FMEA must be done with a team. An FMEA completed by an individual is only that individual’s opinion and does not meet the requirements or the intent of an FMEA. The elements of an effective FMEA team are: • Expertise in subject (five to seven individuals) • Multi-level/consensus based

SL3151Ch06Frame Page 232 Thursday, September 12, 2002 6:09 PM

232

Six Sigma and Beyond

• Representing all relevant stakeholders (those who have ownership) • Possible change in membership as work progresses • Cross-functional and multidisciplinary (One person’s best effort cannot approach the knowledge of an effective cross-functional and multidisciplinary team.) • Appropriate and applicable empowerment The structure of the FMEA team is based on: Core team The experts of the project and the closest to the project. They facilitate honest communication and encourage active participation. Support membership may vary depending on the stage of the project. Champion/sponsor • Provides resources and support • Attends some meetings • Supports team • Promotes team efforts and implements recommendations • Shares authority/power with team • Kicks off team • Higher up in management the better Team leader A team leader is the “watchdog” of the project. Typically, this function falls upon the lead engineer. Some of the ingredients of a good team leader are: • Possesses good leadership skills • Is respected by team members • Leads but does not dominate • Maintains full team participation Recorder Keeps documentation of team’s efforts. The recorder is responsible for coordinating meeting rooms and times as well as distributing meeting minutes and agendas. Facilitator The “watchdog” of the process. The facilitator keeps the team on track and makes sure that everyone participates. In addition, it the facilitator’s responsibility to make sure that team dynamics develop in a positive environment. For the facilitator to be effective, it is imperative for the facilitator to have no stake on the project, possess FMEA process expertise, and communicate assertively. Important considerations for a team include: • Continuity of members • Receptive and open-minded • Committed to success

SL3151Ch06Frame Page 233 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

• • • • •

233

Empowered by sponsor Cross-functionality Multidiscipline Consensus Positive synergy

Ingredients of a motivated FMEA team include: • • • • • • • • •

Realistic agendas Good facilitator Short meetings Right people present Reach decisions based on consensus Open minded, self initiators, volunteers Incentives offered Ground rules established One individual responsible for coordination and accountability of the FMEA project (Typically for the design, the design engineer is that person and for the process, the manufacturing engineer has that responsibility.)

To make sure the effectiveness of the team is sustained throughout the project, it is imperative that everyone concerned with the project bring useful information to the process. Useful information may be derived due to education, experience, training, or a combination of these. At least two areas that are usually underutilized for useful information are background information and surrogate data. Background information and supporting documents that may be helpful to complete system, design, or process FMEAs are: • • • • • • • • • •

Customer specifications (OEMs) Previous or similar FMEAs Historical information (warranty/recalls etc.) Design reviews and verification reports Product drawings/bill of material Process flow charts/manufacturing routing Test methods Preliminary control and gage plans Maintenance history Process capabilities

Surrogate data are data that are generated from similar projects. They may help in the initial stages of the FMEA. When surrogate data are used, extra caution should be taken. Potential FMEA team members include: • Design engineers • Manufacturing engineers

SL3151Ch06Frame Page 234 Thursday, September 12, 2002 6:09 PM

234

Six Sigma and Beyond

• • • • • • • • •

Quality engineers Test engineers Reliability engineers Maintenance personnel Operators (from all shifts) Equipment suppliers Customers Suppliers Anyone who has a direct or indirect interest • In any FMEA team effort the individuals must have interaction with manufacturing and/or process engineering while conducting a design FMEA. This is important to ensure that the process will manufacture per design specification. • On the other hand, interaction with design engineering while conducting a process or assembly FMEA is important to ensure that the design is right. • In either case, group consensus will identify the high-risk areas that must be addressed to ensure that the design and/or process changes are implemented for improved quality and reliability of the product

Obviously, these lists are typical menus to choose an appropriate team for your project. The actual team composition for your organization will depend upon your individual project and resources. Once the team is chosen for the given project, spend 15–20 minutes creating a list of the biggest (however you define “biggest”) concerns for this product or process. This list will be used later to make sure you have a complete list of functions.

7. DEFINE

THE

FMEA PROJECT

AND

SCOPE

Teams must know their assignment. That means that they must know: • • • •

What they are working on (scope) What they are not working on (scope) When they must complete the work Where and how often they will meet

Two excellent tools for such an evaluation are (1) block diagram for system, design, and machinery and (2) process flow diagram for process. In essence, part of the responsibility to define the project and scope has to do with the question “How broad is our focus?” Another way to say this is to answer the question “How detailed do we have to be?” This is much more difficult than it sounds and it needs some heavy discussion from all the members. Obviously, consensus is imperative. As a general rule, the focus is dependent upon the project and the experience or education of the team members. Let us look at an example. It must be recognized that sometimes due to the complexity of the system, it is necessary to narrow the scope of the FMEA. In other words, we must break down the system into smaller pieces — see Figure 6.5.

SL3151Ch06Frame Page 235 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

Master cylinder Pedals and linkages Hydraulics Brake System

Back plate and hardware

235

Cylinder, fluid bladder, etc Pedal, rubber cover, cotter pins, etc. Rubber hose, metal tubing, proportioning valve, fitting, etc. Back plate, springs, washer, clips, etc.

Caliper system

Pistons, cylinder, casting, plate, etc.

Rotor and studs

Rotor hat, rotor, studs, etc.

Pads and hardware

Friction material, substrate, rivets, clip etc.

OUR FMEA SCOPE

FIGURE 6.5 Scope for DFMEA — braking system.

THE FMEA FORM There are many forms to develop a typical FMEA. However, all of them are basically the same in that they are made up of two parts, whether they are for system, design, process, or machinery. A typical FMEA form consists of the header information and the main body. There is no standard information that belongs in the header of the form, but there are specific requirements for the body of the form. In the header, one may find the following information — see Figure 6.7. However, one must remember that this information may be customized to reflect one’s industry or even the organization: • • • • • • • • • •

Type of FMEA study Subject description Responsible engineer FMEA team leader FMEA core team members Suppliers Appropriate dates (original issue, revision, production start, etc.) FMEA number Assembly/part/detail number Current dates (drawings, specifications, control plan, etc.)

The form may be expanded to include or to be used for such matters as: Safety: Injury is the most serious of all failure effects. As a consequence, safety is handled either with an FMEA or a fault tree analysis (FTA) or critical

SL3151Ch06Frame Page 236 Thursday, September 12, 2002 6:09 PM

236

Six Sigma and Beyond

Yes Good? No

No

M

Inpect print

H

Apply paste

L

M

Wash board

Load board

L

L Run, package and ship

Our scope

Dispense paste

H

Set up machine

H

Load screen

L Load sqeegee

L Load tool plate

L Develop program

Legend: L: Low risk M: Medium risk H: High risk Note: Just as in design FMEA, sometimes it is necessary to “narrow the scope” of the process FMEA. FIGURE 6.6 Scope for PFMEA — printed circuit board screen printing process.

FMEA WORKSHEET System FMEA ____Design FMEA ____Process FMEA ____FMEA Number ____ Subject: ______________Team Leader.________________Page ____ of _____ Part/Proc. ID No. __________Date Orig. _____________Date Rev. __________ Key Date. ____________Team Members: ___________________

FIGURE 6.7 Typical FMEA header.

Failure Mode Analysis

Action Plan

Action Results

FIGURE 6.8 Typical FMEA body.

S

Part name or Potential Potential S C Potential O Current D RPN Recommended Target Actual Actions S O D R Remarks L cause of P effect of controls action and finish finish taken process step failure A failure N failure responsibility date date and function mode S mode mode

Description

SL3151Ch06Frame Page 237 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA) 237

SL3151Ch06Frame Page 238 Thursday, September 12, 2002 6:09 PM

238

Six Sigma and Beyond

analysis (FMCA). In the traditional FTA, the starting point is the list of hazard or undersized events for which the designer must provide some solution. Each hazard becomes a failure mode and thus it requires an analysis. Effect of downtime: The FMEA may incorporate maintenance data to study the effects of downtime. It is an excellent tool to be used in conjunction with total preventive maintenance. Repair planning: An FMEA may provide preventive data to support repair planning as well as predictive maintenance cycles. Access: In the world of recycling and environmental conscience, the FMEA can provide data for tear downs as well as information about how to get at the failed component. It can be used with mistake proofing for some very unexpected positive results. A typical body of an FMEA form may look like Figure 6.8. The details of this form will be discussed in the following pages. We begin with the first part of the form; that is the description in the form of: Part name/process step and function (verb/noun) In this area the actual description is written in concise, exact and simple language.

DEVELOPING

THE

FUNCTION

A fundamental principle in writing functions is the notion that they must be written either in action verb format or as a measurable noun. Remember, a function is a task that a component, subsystem, or product must perform, described in language that everyone understands. Stay away from jargon. To identify appropriate functions, leading questions such as the following may help: • • • •

What does the product/process do? How does the product/process do that? If a product feature or process step is deleted, what functions disappear? If you were this task, what are you supposed to accomplish? Why do you exist?

The priority of asking function questions for a system/part FMEA is: 1. A system view 2. A subsystem view 3. A component view Typical functions are:

SL3151Ch06Frame Page 239 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

HOW?

239

Primary supporting function

Supporting functions

Primary supporting function

Secondary supporting function

Primary supporting function Task function

Ensure dependability Ensure convenience

WHY?

Secondary enhancing function

Please senses Delight customer

Enhancing functions

Tertiary supporting function Tertiary supporting function

Tertiary enhancing function Tertiary enhancing function Tertiary enhancing function

FIGURE 6.9 Function tree process.

• • • • •

Position Support Seal in, out Retain Lubricate

ORGANIZING PRODUCT FUNCTIONS After the brainstorming is complete, a function tree — see Figure 6.9 — can be used to organize the functions. This is a simple tree structure to document and help organize the functions, as follows: Purposes of the function tree a. To document all the functions b. To improve team communication c. To document complexity and improve team understanding of all the functions Steps a. Brainstorm all the functions. b. Arrange functions into function tree. c. Test for completeness of function (how/why). Building the function tree Ask: What does the product/process do? Which component/process step does that?

SL3151Ch06Frame Page 240 Thursday, September 12, 2002 6:09 PM

240

Six Sigma and Beyond

How does it do that? • Primary functions provide a direct answer to this question without conditions or ambiguity. • Secondary functions explain how primary functions are performed. • Continue until the answer to “how” requires using a part name, labor operation, or activity. • Ask “why” in the reverse direction. • Add additional functions as needed. The function tree process can be summarized as follows: 1. Identify the task function. Place on the far left side of a chart pad. 2. Identify the supporting functions. Place on the top half of the pad. 3. Identify enhancing functions. Place on the bottom half of the pad. 4. Build the function tree. Include the secondary/tertiary functions. Place these to the right of the primary functions. 5. Verify the diagram: Ask how and why. For an example of a function tree for a ball point pen (tip), see Figure 6.10.

FAILURE MODE ANALYSIS The second portion of the FMEA body form deals with the failure mode analysis. A typical format is shown in Figure 6.11. Understanding Failure Mode Failure mode (a specific loss of a function) is the inability of a component/subsystem/system/process/part to perform to design intent. In other words, it may potentially fail to perform its function(s). Design failure mode is a technical description of how the system, subsystem, or part may not adequately perform its function. Process failure mode is a technical description of how the manufacturing process may not perform its function, or the reason the part may be rejected. Failure Mode Questions The process of brainstorming failure modes may include the following questions: DFMEA • Considering the conditions in which the product will be used, how can it fail to perform its function? • How have similar products failed in the past?

SL3151Ch06Frame Page 241 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

241

Super pen makes marks on varied surfaces User grasps barrel and moves pen axially while simultaneously pressing down on the tip at a vector to the 180 degree plane Axial Force Function

Vector Force Function

The inside diameter of the barrel tip end transmits axial force to the tip system housing sheath O.D.

The end of the barrel and the barrel I.D. (tip end) simultaneously apply force to the tip system housing end and sheath.

The tip system housing tip retainer I.D. transmits axial force to the ball housing

The tip assembly housing transmits the vector force to the O.D. of the ball housing

The ball housing I.D. (ball) transmits axial force on the ball

The ball transmits axial force to the marking surface, however, the marking surface is stationary, which causes the ball rotational motion

The ball housing transmits the vector force to the ball, the ball moves up into the ball housing creating a gap between the ball and ball housing The ink flows through the ink tube contacting the ball surface

The ball rotates through the ink supply, picking up a film of ink on the ball surface The ink is transferred from the ball surface to the marking surface The ink remains on the marking surface (3mm width) area, drying in 3 seconds FIGURE 6.10 Example of ballpoint pen.

PFMEA • Considering the conditions in which the process will be used, what could possibly go wrong with the process? • How have similar processes failed in the past? • What might happen that would cause a part to be rejected?

SL3151Ch06Frame Page 242 Thursday, September 12, 2002 6:09 PM

242

Six Sigma and Beyond

Potential Failure Mode

Potential Effects of Failure Mode

S E V E R I T Y

C Potential L Causes of A S Failure Mode S

O C C U R R E N C E

Current Controls

D E T E C T I O N

Risk Priority Number (RPN)

Identify the Potential Failure FIGURE 6.11 FMEA body.

Determining Potential Failure Modes Failure modes are when the function is not fulfilled in five major categories. Some of these categories may not apply. As a consequence, use these as “thought provokers” to begin the process and then adjust them as needed: 1. 2. 3. 4. 5. 6.

Absence of function Incomplete, partial, or decayed function Related unwanted “surprise” failure mode Function occurs too soon or too late Excess or too much function Interfacing with other components, subsystems or systems. There are four possibilities of interfacing. They are (a) energy transfer, (b) information transfer, (c) proximity, and (d) material compatibility.

Failure mode examples using the above categories and applied to the pen case include: 1. Absence of function: • DFMEA: Make marks • PFMEA: Inject plastic 2. Incomplete, partial or decayed function: • DFMEA: Make marks • PFMEA: Inject plastic 3. Related unwanted “surprise” failure mode • DFMEA: Make marks • PFMEA: Inject plastic 4. Function occurs too soon or too late • DFMEA: Make marks • PFMEA: Inject plastic 5. Excess or too much function • DFMEA: Make marks • PFMEA: Inject plastic

SL3151Ch06Frame Page 243 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

243

General examples of failure modes include: Design FMEA: No power Water leaking Open circuit Releases too early Noise Vibration Does not cut

Failed to open Partial insulation Loss of air No spark Insufficient torque Paper jams And so on

Process FMEA: Four categories of process failures: 1. 2. 3. 4.

Fabrication failures Assembly failures Testing failures Inspection failures

Typical examples of these categories are: • • • • • • • • • • • •

Warped Too hot RPM too slow Rough surface Loose part Misaligned Poor inspection Hole too large Leakage Fracture Fatigue And so on

Note: At this stage, you are ready to transfer the failure modes in the FMEA form — see Figure 6.12.

FAILURE MODE EFFECTS A failure mode effect is a description of the consequence/ramification of a system, part, or manufacturing process failure. A typical failure mode may have several effects depending on which customer(s) are considered. Consider the effects/consequences on all the “customers,” as they are applicable, as in the following FMEAs: SFMEA • System • Other systems

SL3151Ch06Frame Page 244 Thursday, September 12, 2002 6:09 PM

244

Six Sigma and Beyond

Potential Potential S Failure Effects E V Mode of E Failure R Mode I

C L A S S

Potential Causes of Failure Mode

T Y

O C C U R R E N C E

Current Controls

D E T E C T I O N

Risk Priority Number (RPN)

Does Not Transfer Ink Partial Ink and so on FIGURE 6.12 Transferring the failure modes to the FMEA form.

• Whole product • Government regulations • End user DFMEA • Part • Higher assembly • Whole product • Government regulations • End user PFMEA • Part • Next operation • Equipment • Government regulations • Operators • End user Effects and Severity Rating Effects and severity are very related items. As the effect increases, so does the severity. In essence, two fundamental questions have to be raised and answered: 1. What will happen if this failure mode occurs? 2. How will customers react if these failures happen? • Describe as specifically as possible what the customer(s) might notice once the failure occurs. • What are the effects of the failure mode? • How severe is the effect on the customers? The progression of function, cause, failure mode, effect, and severity can be illustrated by the following series of questions: In function: What is the individual task intended by design? In failure mode: What can go wrong with this function?

SL3151Ch06Frame Page 245 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

245

In cause: What is the “root cause” of the failure mode? In effect: What are the consequences of this failure mode? In severity: What is the seriousness of the effect? The following are some examples of DFMEA and PFMEA effects: Customer gets wet System failure Loss of efficiency Reduced life Degraded performance Cannot assemble Violate Gov. Reg. XYZ Damaged equipment

Loss of performance Scrap Rework Becomes loose Hard to load in next operation Operator injury Noise, rattles And so on

Special Note: Please note that the effect remains the same for both DFMEA and PFMEA. Severity Rating (Seriousness of the Effect) Severity is a numerical rating — see Table 6.1 for design and Table 6.2 for process — of the impact on customers. When multiple effects exist for a given failure mode, enter the worst-case severity on the worksheet to calculate risk. (This is the excepted method for the automotive industry and for the SAE J1739 standard. In cases where severity varies depending on timing, use the worst-case scenario. Note: There is nothing special about these guidelines. They may be changed to reflect the industry, the organization, the product/design, or the process. For example, the automotive industry has its own version and one may want to review its guidelines in the AIAG (2001). To modify these guidelines, keep in mind: 1. 2. 3. 4. 5.

List the entire range of possible consequences (effects). Force rank the consequences from high to low. Resolve the extreme values (rating 10 and rating 1). Fill in the other ratings. Use consensus.

At this point the information should be transferred to the FMEA form — see Figure 6.13. The column identifying the “class” is the location for the placement of the special characteristic. The appropriate response is only “Yes” or “No.” A Yes in this column indicates that the characteristic is special, a No indicates that the characteristic is not special. In some industries, special characteristics are of two types: (a) critical and (b) significant. “Critical” refers to characteristics associated with safety and/or government regulations, and “significant” refers to those that affect the integrity of the product. In design, all special characteristics are potential. In process they become critical or significant depending on the numerical values of severity and occurrence combinations.

SL3151Ch06Frame Page 246 Thursday, September 12, 2002 6:09 PM

246

Six Sigma and Beyond

TABLE 6.1 DFMEA — Severity Rating Effect

Description

None

No effect noticed by customer; the failure will not have any perceptible effect on the customer Very minor effect, noticed by discriminating customers; the failure will have little perceptible effect on discriminating customers Minor effect, noticed by average customers; the failure will have a minor perceptible effect on average customers Very low effect, noticed by most customers; the failure will have some small perceptible effect on most customers Primary product function operational, however at a reduced level of performance; customer is somewhat dissatisfied Primary product function operational, however secondary functions inoperable; customer is moderately dissatisfied Failure mode greatly affects product operation; product or portion of product is inoperable; customer is very dissatisfied Primary product function is non-operational but safe; customer is very dissatisfied. Failure mode affects safe product operation and/or involves nonconformance with government regulation with warning Failure mode affects safe product operation and/or involves nonconformance with government regulation without warning

Very minor

Minor Very low Low Moderate High Very high Hazard with warning Hazard with no warning

FAILURE CAUSE

AND

Rating 1 2

3 4 5 6 7 8 9 10

OCCURRENCE

The analysis of the cause and occurrence is based on two questions: 1. What design or process choices did we already make that may be responsible for the occurrence of a failure? 2. How likely is the failure mode to occur because of this? For each failure mode, the possible mechanism(s) and cause(s) of failure are listed. This is an important element of the FMEA since it points the way toward preventive/corrective action. It is, after all, a description of the design or process deficiency that results in the failure mode. That is why it is important to focus on the “global” or “root” cause. Root causes should be specific and in the form of a characteristic that may be controlled or corrected. Caution should be exerted not to overuse “operator error” or “equipment failure” as a root cause even though they are both tempting and make it easy to assign “blame.” You must look for causes, not symptoms of the failure. Most failure modes have more than one potential cause. An easy way to probe into the causes is to ask: What design choices, process variables, or circumstances could result in the failure mode(s)?

SL3151Ch06Frame Page 247 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

247

TABLE 6.2 PFMEA — Severity Rating Effect

Description

None

No effect noticed by customer; the failure will not have any effect on the customer Very minor disruption to production line; a very small portion of the product may have to be reworked; defect noticed by discriminating customers Minor disruption to production line; a small portion (much 1 in 2

Rating 1 2 3 4 5 6 7 8 9 10

• Experiments — When many causes are suspect or specific cause is unknown • Classical • Taguchi methods Occurrence Rating The occurrence rating is an estimated number of frequencies or cumulative number of failures (based on experience) that will occur in our design concepts for a given cause over the intended life of the design. For example: cause of staples falling out = soft wood. The likelihood of occurrence is a 9 if we pick balsa wood but a 2 if we choose oak. Just as with severity, there are standard tables for occurrence — see Table 6.3 for design and Table 6.4 for process — for each type of FMEA. The ratings on these tables are estimates based on experience or similar products or processes. Nonstandard occurrence tables may also be used, based on specific characteristics. However, reliability expertise is needed to construct occurrence tables. (Typical characteristics may be historical failure frequencies, Cpks, theoretical distributions, and reliability statistics.) At this point the data for causes and their ratings should be transferred to the FMEA form — see Figure 6.14. Current Controls and Detection Ratings Design and process controls are the mechanisms, methods, tests, procedures, or controls that we have in place to prevent the cause of the failure mode or detect the failure mode or cause should it occur. (The controls currently exist.) Design controls prevent or detect the failure mode prior to engineering release. Process controls prevent or detect the failure mode prior to the part or assembly leaving the area.

SL3151Ch06Frame Page 250 Thursday, September 12, 2002 6:09 PM

250

Six Sigma and Beyond

TABLE 6.4 PFMEA — Occurrence Rating Occurrence

Description

Remote

Failure is very unlikely; no failures associated with similar processes Few failures; isolated failures associated with like processes Occasional failures associated with similar processes, but not in major proportions

Low Moderate

High Very high

Repeated failures; similar processes have often failed Process failure is almost inevitable

Frequency

Cpk

Rating

1.67

1

1 in 150,000 1 in 15,000 1 in 2,000 1 in 400 1 in 80 1 in 20 1 in 8 1 in 3 >1 in 2

1.50 1.33 1.17 1.00 0.83 0.67

2 3 4 5 6 7 8 9 10

0.51 0.33

A good control prevents or detects causes or failure modes. • As early as possible (ideally before production or prototypes) • As early as possible • Using proven methods So, the next step in the FMEA process is to: • Analyze planned controls for your system, part, or manufacturing process • Understand the effectiveness of these controls to detect causes or failure modes Detection Rating Detection rating — see Table 6.5 for design and Table 6.6 for process — is a numerical rating of the probability that a given set of controls will discover a specific cause or failure mode to prevent bad parts from leaving the operation/facility or getting to the ultimate customer. Assuming that the cause of the failure did occur, assess the capabilities of the controls to find the design flaw or prevent the bad part from leaving the operation/facility. In the first case, the DFMEA is at issue. In the second case, the PFMEA is of concern. When multiple controls exist for a given failure mode, record the best (lowest) to calculate risk. In order to evaluate detection, there are appropriate tables for both design and process. Just as before, however, if there is a need to alter them, remember that the change and approval must be made by the FMEA team with consensus. At this point, the data for current controls and their ratings should be transferred to the FMEA form — see Figure 6.15. There should be a current control for every cause. If there is not, that is a good indication that a problem might exist.

SL3151Ch06Frame Page 251 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

Potential Failure Mode

Potential Effects of Failure Mode

S E V E R I T Y

Does not Pan does 8 transfer ink not work; customer tries and eventually tears paper and scraps the pen Old pen stops 7 Partial ink writing, customer scraps pen Customer 7 has to retrace Writing or drawing looks bad

251

C Potential L Causes of A Failure Mode S S

N

O C C U R R E N C E Ball housing 2 I.D. deform

Current Controls

D E T E C T I O N

Risk Priority Number (RPN)

Ink viscosity 9 too high Debris build- 5 up

N

Inconsistent ball rolling due to deformed housing

2

Ball does not 7 always pick up ink due to ink viscosity Housing I.D. 1 variation due to mfg and so on

and so on

and so on FIGURE 6.14 Transferring causes and occurrences to the FMEA form.

UNDERSTANDING

AND

CALCULATING RISK

Without risk, there is very little progress. Risk is inevitable in any system, design, or manufacturing process. The FMEA process aids in identifying significant risks, then helps to minimize the potential impact of risk. It does that through the risk priority number or as it is commonly known, the RPN index. In the analysis of the RPN, make sure to look at risk patterns rather than just a high RPN. The RPN is the product of severity, occurrence, and detection or:

SL3151Ch06Frame Page 252 Thursday, September 12, 2002 6:09 PM

252

Six Sigma and Beyond

TABLE 6.5 DFMEA Detection Table Detection

Description

Almost certain

Design control will almost certainly detect the potential cause of subsequent failure modes Very high chance the design control will detect the potential cause of subsequent failure mode High chance the design control will detect the potential cause of subsequent failure mode Moderately high chance the design control will detect the potential cause of subsequent failure mode Moderate chance the design control will detect the potential cause of subsequent failure mode Low chance the design control will detect the potential cause of subsequent failure mode Very low chance the design control will detect the potential cause of subsequent failure mode Remote chance the design control will detect the potential cause of subsequent failure mode Very remote chance the design control will detect the potential cause of subsequent failure mode There is no design control or control will not or cannot detect the potential cause of subsequent failure mode

Very high High Moderately high Moderate Low Very low Remote Very remote Very uncertain

Rating 1 2 3 4 5 6 7 8 9 10

Risk = RPN = S × O × D Obviously the higher the RPN the more the concern. A good rule-of-thumb analysis to follow is the 95% rule. That means that you will address all failure modes with a 95% confidence. It turns out the magic number is 50, as indicated in this equation: [(S = 10 × O = 10 × D = 10) – (1000 × .95)]. This number of course is only relative to what the total FMEA is all about, and it may change as the risk increases in all categories and in all causes. Special risk priority patterns require special attention, through specific action plans that will reduce or eliminate the high risk factor. They are identified through: 1. High RPN 2. Any RPN with a severity of 9 or 10 and occurrence > 2 3. Area chart The area chart — Figure 6.16 — uses only severity and occurrence and therefore is a more conservative approach than the priority risk pattern mentioned previously. At this stage, let us look at our FMEA project and calculate and enter the RPN — see Figure 6.17. It must be noted here that this is only one approach to evaluating risk. Another possibility is to evaluate the risk based on the degree of severity first,

SL3151Ch06Frame Page 253 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

253

TABLE 6.6 PFMEA Detection Table Detection

Description

Almost certain

Process control will almost certainly detect or prevent the potential cause of subsequent failure mode Very high chance process control will detect or prevent the cause of subsequent failure mode High chance the process control will detect or prevent the potential cause of subsequent failure mode. Moderately high chance the process control will detect or prevent the potential cause of subsequent failure mode Moderate chance the process control will detect or prevent the potential cause of subsequent failure mode Low chance the process control will detect or prevent the potential cause of subsequent failure mode Very low chance the process control will detect or prevent the potential cause of subsequent failure mode Remote chance the process control will detect or prevent the potential cause of subsequent failure mode Very remote chance the process control will detect or prevent the potential cause of subsequent failure mode There is no process control or control will not or cannot detect the potential cause of subsequent failure mode

Very high High Moderately high Moderate Low Very low Remote Very remote Very uncertain

Rating 1 2 3 4 5 6 7 8 9 10

in which case the engineer tries to eliminate the failure; evaluate the risk based on a combination of severity (values of 5–8) and occurrence (>3) second, in which case the engineer tries to minimize the occurrence of the failure through a redundant system; and to evaluate the risk through the detection of the RPN third, in which case the engineer tries to control the failure before the customer receives it.

ACTION PLANS AND RESULTS The third portion of the FMEA form deals with the action plans and results analysis. A typical format is shown in Figure 6.18. The idea of this third portion of the FMEA form is to generate a strategy that reduces severity and occurrence and makes the detection effective to reduce the total RPN: Reducing the severity rating (or reducing the severity of the failure mode effect) • Design or manufacturing process changes are necessary. • This approach is much more proactive than reducing the detection rating. Reducing the occurrence rating (or reducing the frequency of the cause) • Design or manufacturing process changes are necessary. • This approach is more proactive than reducing the detection rating.

SL3151Ch06Frame Page 254 Thursday, September 12, 2002 6:09 PM

254

Potential Failure Mode

Six Sigma and Beyond

Potential Effects of Failure Mode

S E V E R I T Y

Does not Pan does 8 transfer ink not work; customer tries and eventually tears paper and scraps the pen Old pen stops 7 Partial ink writing, customer scraps pen Customer has to 7 retrace

Writing or drawing looks bad

C Potential L Causes of A Failure Mode S S

N Ball housing I.D. deform

Ink viscosity 9 Test # X too high

D E T E C T I O N

Risk Priority Number (RPN)

2 10

Debris build- 5 Design review 7 Prototype test # up XY

N Inconsistent ball rolling due to deformed housing

2 Test # X

10

Ball does not 7 None always pick up ink due to ink viscosity

10

Housing I.D. 1 None variation due to mfg

10

and so on

and so on

O Current C Controls C U R R E N C E 2 Life test Test # X

and so on

and so on FIGURE 6.15 Transferring current controls and detection to the FMEA form.

Reducing the detection rating (or increasing the probability of detection) • Improving the detection controls is generally costly, reactive, and does not do much for quality improvement, but it does reduce risk. • Increased frequency of inspection, for example, should only be used as a last resort. It is not a proactive corrective action.

CLASSIFICATION

AND

CHARACTERISTICS

Different industries have different criteria for classification. However, in all cases the following characteristics must be classified according to risk impact:

SL3151Ch06Frame Page 255 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

Occurrence

Severity

10 9 8 7 6 5 4 3 2 1 1

255

High Priority Medium Priority Low Priority 2

3

4

5

6

7

8

9

10

FIGURE 6.16 Area chart.

• Severity 9, 10: Highest classification (critical) These product- or process-related characteristics: • May affect compliance with government or federal regulations (EPA, OSHA, FDA, FCC, FAA, etc.) • May affect safety of the customer • Require specific actions or controls during manufacturing to ensure 100% compliance • Severity between 5 and 8 and occurrence greater than 3: Secondary classification (significant) These product- or process-related characteristics: • Are non-critical items that are important for customer satisfaction (e.g., fit, finish, durability, appearance) • Should be identified on drawings, specifications, or process instructions to ensure acceptable levels of capability • High RPN: Secondary classification (see Table 6.7) Product Characteristics/“Root Causes” Examples include size, form, location, orientation, or other physical properties such as color, hardness, strength, etc. Process Parameters/“Root Causes” Examples include pressure, temperature, current, torque, speeds, feeds, voltage, nozzle diameter, time, chemical concentrations, cleanliness of incoming part, ambient temperature, etc.

DRIVING

THE

ACTION PLAN

For each recommended action, the FMEA team must:

SL3151Ch06Frame Page 256 Thursday, September 12, 2002 6:09 PM

256

Six Sigma and Beyond

Potential Failure Mode

Potential Effects of Failure Mode

S E V E R I T Y

Does not Pan does 8 transfer ink Pan does not work; customer tries and eventually tears paper and scraps the pen Old pen stops writing, Partial ink customer scraps pen

7

Customer 7 has to retrace

Writing or drawing looks bad

C Potential L Causes A of S Failure Mode S

O Current C Controls C U R R E N C E N Ball housing 2 Life test Test # X I.D. deform

Risk Priority Number (RPN)

2

32

Ink viscosity 9 Test # X 10 too high Design review Debris build- 5 Prototype test # 7 up XY

N Inconsistent ball rolling due to deformed housing

2 Test # X

Ball does not 7 None always pick up ink due to ink viscosity 1 None Housing I.D. variation due to mfg and so on

and so on

D E T E C T I O N

and so on

720 280

10

140

10

490

10

70

and so on

and so on FIGURE 6.17 Transferring the RPN to the FMEA form.

• Plan for implementation of recommendations • Make sure that recommendations are followed, demonstrate improvement, and are completed Implementation of action plans requires answering the classic questions… • WHO … (will take the lead) • WHAT… (specifically is to be done)

SL3151Ch06Frame Page 257 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

Action Plan Recommended Actions and Responsibility

257

Action Results

Target Finish Date

Actual Finish Date

Actions S O D RPN Taken

Remarks

FIGURE 6.18 Action plans and results analysis.

TABLE 6.7 Special Characteristics for Both Design and Process FMEA Type

Classification

Purpose

Criteria

Design

YC

Severity = 9–10

Does not apply

Design

YS

Severity = 5–8 Occurrence = 4–10

Does not apply

Design

Blank

Severity < 5

Does not apply

Process Process

Inverted delta SC

A potential critical characteristic (Initiate PFMEA) A potential significant characteristic (Initiate PFMEA) Not a potential critical or significant characteristic A critical characteristic A significant characteristic

Required Required

Process

HI

High impact

Process Process

OS Blank

Operator safety Not a special characteristic

Severity = 9–10 Severity = 5–8 Occurrence = 4–10 Severity = 5–8 Occurrence = 4–10 Severity = 9–10 Other

• • • •

Control

Emphasis Safety sign-off Does not apply

WHERE … (will the work get done) WHY… (this should be obvious) WHEN… (should the actions be done) HOW… (will we start)

Additional points concerning the action plan include the following: • Accelerate implementation by getting buy-in (ownership). • It is important to draw out and address objections. • When plans address objections in a constructive way, stakeholders feel ownership in plans and actions. Ownership aids in successful implementation.

SL3151Ch06Frame Page 258 Thursday, September 12, 2002 6:09 PM

258

Six Sigma and Beyond

• Typical questions that begin a fruitful discussion are: • Why are we…? • Why not this…? • What about this…? • What if…? • Timing and actions must be reviewed on a regular basis to: • Maintain a sense of urgency • Allow for ongoing facilitation • Ensure work is progressing • Drive team members to meet commitments • Surface new facts that may affect plans • Fill in the actions taken. • The “Action Taken” column should not be filled out until the actions are totally complete. • Record final outcomes in the Action Plan and Action Results sections of the FMEA form. Remember, because of the actions you have taken you should expect changes in severity, occurrence, detection, RPN, and new characteristic designations. Of course, these changes may be individual or in combination. The form will look like Figure 6.19.

LINKAGES AMONG DESIGN AND PROCESS FMEAS AND CONTROL PLAN FMEAs are not islands unto themselves. They have continuity, and the information must be flowing throughout the design and process FMEAs as well as to the control plan. A typical linkage is shown in Figure 6.20. In addition to the control plan, the FMEA is also linked with robustness. To appreciate these linkages in FMEA, we must recall that design for six sigma (DFSS) must be a robust process. In fact, to see the linkages of this robustness we may begin with a P diagram (see Volume V) and identify its components. It turns out that the robustness in the FMEA usage is to make sure that the part, subsystem, or system is going to perform its intended function, in spite of problems in both manufacturing and environment. Of particular interest are the error states, control factors, and noise factors. Error states may help in identifying the failures, noise factors may help us in identifying causes, and control factors may help us in identifying the recommendations. The signal and response become the functions or the starting point of the FMEA. The linkages then help generate the inputs and outputs of the FMEA. Typical inputs are: System (concept) inputs P diagram Boundary diagram Interface matrix Potential design verification tests

SL3151Ch06Frame Page 259 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

Description Failure Action Mode Plan Analysis RPN Recommended Target Action and Finish Responsibility Date 32 No action required 720 DOE - Taguchi 3/22/99

259

Action Results Actual Finish Date

Action Taken

2/15/99 Optimize ink formula

8

DR R Remarks P N 2 2 32 None

8

1 10

80

5 1

40

1 4

28

1 10

70

SO

280 Develop accel. 2/18/99 2/3/99 Test 8 procedure test (thermal revised vibration) 4/3/99 D. Robins 140 Develop new test # ABC

2/2/99 2/2/99

Test implemented 2/2/99

7

5/3/99 4/30/99 Optimized DOE - Taguchi 7 ink optimize formula viscosity on C. Abrams 4/30/99 70 Evaluate TBD machining process and so on and so on 490

7

FIGURE 6.19 Transferring action plans and action results on the FMEA form.

Surrogate data for reliability and robustness considerations Corporate requirements Benchmarking results Customer functionality in terms of engineering specifications Regulatory requirements review Design inputs P diagram Boundary diagram Interface matrix Customer functionality in terms of engineering specifications Regulatory requirements review Process inputs P diagram Process flow diagram Special characteristics from the DFMEA Process characteristics Regulatory requirements review

SL3151Ch06Frame Page 260 Thursday, September 12, 2002 6:09 PM

260

Six Sigma and Beyond

Design FMEA Quality Function Deployment

Function Failure Effect Severity Class Cause Controls Rec. Action

System Design Specifications

Sign-Off Report Design Verification Plan and Report Process FMEA Part Characteristic 1 2 3 4 etc.

C Function Failure Effect Controls L Normal A S S

Cause

Controls Reaction Special

Remove the classification symbol

Dynamic Control Plan

Part Drawing (Inverted Delta and Special Characteristics)

FIGURE 6.20 FMEA linkages.

Machinery inputs P diagram Boundary diagram Interface matrix Customer functionality in terms of engineering specifications Regulatory requirements review

GETTING THE MOST FROM FMEA Common team problems that may make it difficult to get the most from FMEA include: • Poor team composition (not cross-functional or multidisciplinary) • Low expertise in FMEA • Not multi-level

SL3151Ch06Frame Page 261 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

• • • • • • • •

261

• Low experience/expertise in product • One-person FMEA Lack of management support Not enough time Too detailed, could go on forever Arguments between team members (Opinions should be based on facts and data.) Lack of team enthusiasm/motivation Difficulty in getting team to start and stay with the process Proactive vs. reactive (a “before the event” not “after the fact” exercise) Doing it for the wrong reason

Common procedural problems include: • Confusion about, poorly defined or incomplete functions, failure modes, effects, or causes • Subgroup discussion • Using symptoms or superficial causes instead of root causes • Confusion about ratings as estimates and not absolutes (It will take time to be consistent.) • Confusion about the relationship between causes, failure modes, and effects • Using “customer dissatisfied” as failure effect • Shifting design concerns to manufacturing and vice-versa • Doing FMEAs by hand • Dependent on the engineers’ “printing skills” • RPNs or criticality cannot be ranked easily • Hard to update • Much space taken up by complicated FMEAs • Time consuming • Resistance to being the “recorder” when done manually • Inefficient means of storing and retrieving info Note: With FMEA software these are all eliminated. • Working non-systematically on the form (It is suggested that the failure analysis should progress from left to right, with each column being completed before the next is begun.) • Resistance of individuals to taking responsibility for recommended actions • Doing a reactive FMEA as opposed to a proactive FMEA (FMEAs are best applied as a problem prevention tool, not problem solving tool, although one may use them for both. However, the value of a reactive FMEA is much less.) • Not having robust FMEA terminology (A robust communication process is one that delivers its “function” [imparting knowledge and understanding] without being affected by “noise factors” [varying degrees of training]. Simply stated, the process should be as clear as possible with minimum possibility for misunderstanding.)

SL3151Ch06Frame Page 262 Thursday, September 12, 2002 6:09 PM

262

Six Sigma and Beyond

Stages of Learning Unconscious incompetence

Stages of FMEA Maturity Never heard of FMEA

Conscious incompetence

We talked about it

Conscious competence

Customer made us do it

Unconscious competence

Some small successes Proper and regular use

FIGURE 6.21 The learning stages.

Institutionalizing FMEA in your company is challenging, and its success is largely dependent upon the culture in the organization as well as the reason it is being utilized. Below are some main considerations: • • • •

Selecting pilot projects (Start small and build successes.) Identifying team participants Developing and promoting FMEA successes Developing templates (databases of failure modes, functions, controls, etc.) • Addressing training needs

Figure 6.21 shows the learning stages (the direction of the arrows indicates the increasing level) in a company that is developing maturity in the use of FMEA.

SYSTEM OR CONCEPT FMEA A concept FMEA is used to analyze concepts at very early stages with new ideas. Concept FMEAs can be design, process, or even machinery oriented. However, in practical terms, most of them are done on a system or subsystem level. The process of the system or concept FMEA is practically the same as that of a design FMEA. In fact, the evaluation guidelines are exactly the same as those of DFMEA. The difference is that in the system FMEA a great effort is made to identify gross failures with high severities. If these problems cannot be overcome, then the project most likely will be killed. If the failures can be fixed through reasonable design changes, then the project moves to a second stage and the design FMEA takes over.

DESIGN FAILURE MODE AND EFFECTS ANALYSIS (DFMEA) The Design Failure Mode and Effects Analysis (Design FMEA) is a method for identifying potential or known failure modes and providing follow-up and corrective actions.

SL3151Ch06Frame Page 263 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

263

OBJECTIVE The design FMEA is a disciplined analysis of the part design with the intent to identify and correct any known or potential failure modes before the manufacturing stage begins. Once these failure modes are identified and the cause and effects are determined, each failure mode is then systematically ranked so that the most severe failure modes receive priority attention. The completion of the design FMEA is the responsibility of the individual product design engineer. This individual engineer is the most knowledgeable about the product design and can best anticipate the failure modes and their corrective actions.

TIMING The design FMEA is initiated during the early planning stages of the design and is continually updated as the program develops. The design FMEA must be totally completed prior to the first production run.

REQUIREMENTS The requirements for a design FMEA include: 1. Forming a team 2. Completing the design FMEA form 3. FMEA risk ranking guidelines

DISCUSSION The effectiveness of an FMEA is dependent on certain key steps in the analysis process, as follows: Forming the Appropriate Team A typical team for conducting a design FMEA is the following: • • • • • • •

Design engineer(s) Test/development engineer Reliability engineer Materials engineer Field service engineer Manufacturing/process engineer Customer

A design and a manufacturing engineer are required to be team members. Others may participate as needed or as the project calls for their knowledge or experience. The leader for the design FMEA is typically the design engineer.

SL3151Ch06Frame Page 264 Thursday, September 12, 2002 6:09 PM

264

Six Sigma and Beyond

Describing the Function of the Design/Product There are three types of functions: 1. Task functions: These functions describe the single most important reason for the existence of the system/product. (Vacuum cleaner? Windshield wiper? Ballpoint pen?) 2. Supporting functions: These are the “sub” functions that are needed in order for the task function to be performed. 3. Enhancing functions: These are functions that enhance the product and improve customer satisfaction but are not needed to perform the task function. After computing the function tree or a block diagram, transfer functions to the FMEA worksheet or some other form of a worksheet to retain. Add the extent of each function (range, target, specification, etc.) to test the measurability of the function. Describing the Failure Mode Anticipated The team must pose the question to itself, “How could this part, system or design fail? Could it break, deform, wear, corrode, bind, leak, short, open, etc.?” The team is trying to anticipate how the design being considered could possibly fail; at this point, it should not make the judgment as to whether it will fail but should concentrate on how it could fail. The purpose of a design FMEA (DFMEA) is to analyze and evaluate a design on its ability to perform its functions. Therefore, the initial assumption is that parts are manufactured and assembled according to plan and in compliance with specifications. Once failure modes are determined under this assumption, then determine other failure modes due to purchased materials, components, manufacturing processes, and services. Describing the Effect of the Failure The team must describe the effect of the failure in terms of customer reaction or in other words, e.g., “What does the customer experience as a result of the failure mode of a shorted wire?” Notice the specificity. This is very important, because this will establish the basis for exploratory analysis of the root cause of the function. Would the shorted wire cause the fuel gage to be inoperative or would it cause the dome light to remain on? Describing the Cause of the Failure The team anticipates the cause of the failure. Would poor wire insulation cause the short? Would a sharp sheet metal edge cut through the insulation and cause the short? The team is analyzing what conditions can bring about the failure mode. The more specific the responses are, the better the outcome of the FMEA.

SL3151Ch06Frame Page 265 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

265

The purpose of a design FMEA (DFMEA) is to analyze and/or evaluate a design on its ability to perform its functions (part characteristics). Therefore, the initial assumption in determining causes is that parts are made and assembled according to plan and in compliance with specifications, including purchased materials, components, and services. Then and only then, determine causes due to purchased materials, components, and services. Some cause examples include: Brittle material Weak fastener Corrosion Low hardness Too small of a gap Wrong bend angle Stress concentration Ribs too thin Wrong material selection Poor stitching design High G forces Part interference Tolerance stack-up Vibration Oxidation And so on Estimating the Frequency of Occurrence of Failure The team must estimate the probability that the given failure is going to occur. The team is assessing the likelihood of occurrence, based on its knowledge of the system, using an evaluation scale of 1 to 10. A 1 would indicate a low probability of occurrence whereas a 10 would indicate a near certainty of occurrence. Estimating the Severity of the Failure In estimating the severity of the failure, the team is weighing the consequence of the failure. The team uses the same 1 to 10 evaluation scale. A 1 would indicate a minor nuisance, while a 10 would indicate a severe consequence such as “loss of brakes” or “stuck at wide open throttle” or “loss of life.” Identifying System and Design Controls Generally, these controls consist of tests and analyses that detect failure modes or causes during early planning and system design activities. Good system controls detect faults or weaknesses in system designs. Design controls consist of tests and analyses that detect failure causes or failure modes during design, verification, and validation activities. Good design controls detect faults or weaknesses in component designs.

SL3151Ch06Frame Page 266 Thursday, September 12, 2002 6:09 PM

266

Six Sigma and Beyond

Special notes: • Just because there is a current control in place that does not mean that it is effective. Make sure the team reviews all the current controls, especially those that deal with inspection or alarms. • To be effective (proactive), system controls must be applied throughout the pre-prototype phase of the Advanced Product Quality Planning (APQP) process. • To be effective (proactive), design controls must be applied throughout the pre-launch phase of the APQP process. • To be effective (proactive), process controls should be applied during the post-pilot build phase of APQP and continue during the production phase. If they are applied only after production begins, they serve as reactive plans and become very inefficient. Examples of system and design controls include: Engineering analysis • Computer simulation • Mathematical modeling/CAE/FEA • Design reviews, verification, validation • Historical data • Tolerance stack studies • Engineering reviews, etc. System/component level physical testing • Breadboard, analog tests • Alpha and beta tests • Prototype, fleet, accelerated tests • Component testing (thermal, shock, life, etc.) • Life/durability/lab testing • Full scale system testing (thermal, shock, etc) • Taguchi methods • Design reviews Estimating the Detection of the Failure The team is estimating the probability that a potential failure will be detected before it reaches the customer. Again, the 1 to 10 evaluation scale is used. A 1 would indicate a very high probability that a failure would be detected before reaching the customer. A 10 would indicate a very low probability that the failure would be detected, and therefore, be experienced by the customer. For instance, an electrical connection left open preventing engine start might be assigned a detection number of 1. A loose connection causing intermittent no-start might be assigned a detection number of 6, and a connection that corrodes after time causing no start after a period of time might be assigned a detection number of 10.

SL3151Ch06Frame Page 267 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

267

Detection is a function of the current controls. The better the controls, the more effective the detection. It is very important to recognize that inspection is not a very effective control because it is a reactive task. Calculating the Risk Priority Number The product of the estimates of occurrence, severity, and detection forms a risk priority number (RPN). This RPN then provides a relative priority of the failure mode. The higher the number, the more serious is the mode of failure considered. From the risk priority numbers, a critical items summary can be developed to highlight the top priority areas where actions must be directed. Recommending Corrective Action The basic purpose of an FMEA is to highlight the potential failure modes so that the responsible engineer can address them after this identification phase. It is imperative that the team provide sound corrective actions or provide impetus for others to take sound corrective actions. The follow-up aspect is critical to the success of this analytical tool. Responsible parties and timing for completion should be designated in all corrective actions. Strategies for Lowering Risk: (System/Design) — High Severity or Occurrence To reduce risk, you may change the product design to: • • • •

Eliminate the failure mode cause or decouple the cause and effect Eliminate or reduce the severity of the effect Make the cause less likely or impossible to occur Eliminate function or eliminate part (functional analysis)

Some “tools” to consider: • • • • •

Quality Function Deployment (QFD) Fault Tree Analysis (FTA) Benchmarking Brainstorming TRIZ, etc.

Evaluate ideas using Pugh concept selection. Some specific examples: • • • •

Change material, increase strength, decrease stress Add redundancy Constrain usage (exclude features) Develop fail safe designs, early warning system

Strategies for Lowering Risk: (System/Design) — High Detection Rating Change the evaluation/verification/tests to:

SL3151Ch06Frame Page 268 Thursday, September 12, 2002 6:09 PM

268

Six Sigma and Beyond

• Make failure mode easier to perceive • Detect causes prior to failure Some “tools” to consider: • • • •

Benchmarking Brainstorming Process control (automatic corrective devices) TRIZ, etc.

Evaluate ideas using Pugh concept selection. Some specific examples: • • • •

Change testing and evaluation procedures Increase failure feedback or warning systems Increase sampling in testing or instrumentation Increase redundancy in testing

PROCESS FAILURE MODE AND EFFECTS ANALYSIS (FMEA) The Process Failure Mode and Effects Analysis (process FMEA) is a method for identifying potential or known processing failure modes and providing problem follow-up and corrective actions.

OBJECTIVE The process FMEA is a disciplined analysis of the manufacturing process with the intent to identify and correct any known or potential failure modes before the first production run occurs. Once these failure modes are identified and the cause and effects are determined, each failure mode is then systematically ranked so that the most severe failure modes receive priority attention. The completion of the process FMEA is the responsibility of the individual product process engineer. This individual process engineer is the most knowledgeable about the process structure and can best anticipate the failure modes and their effects and address the corrective actions.

TIMING The process FMEA is initiated during the early planning stages of the process before machines, tooling, facilities, etc., are purchased. The process FMEA is continually updated as the process becomes more clearly defined. The process FMEA must be totally completed prior to the first production run.

REQUIREMENTS The requirements for a process FMEA are as follows:

SL3151Ch06Frame Page 269 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

269

1. Form team 2. Complete the process FMEA form 3. FMEA risk ranking guidelines

DISCUSSION The effectiveness of an FMEA on a process is dependent on certain key steps in the analysis, including the following: Forming the Team A typical team for the process/assembly FMEA is the following: • • • • • • • •

Design engineer Manufacturing or process engineer Quality engineer Reliability engineer Tooling engineer Responsible operators from all shifts Supplier Customer

A design engineer, a manufacturing engineer, and representative operators are required to be team members. Others may participate as needed or as the project calls for their knowledge or experience. The leader for the process FMEA is typically the process or manufacturing engineer. Describing the Process Function The team must identify the process or machine and describe its function. The team members should ask of themselves, “What is the purpose of this operation?” State concisely what should be accomplished as a result of the process being performed. Typically, there are three areas of concern. They are: 1. Creating/constructing functions: These are the functions that add value to the product. Examples include cutting, forming, painting, drying, etc. 2. Improving functions: These are the functions that are needed in order to improve the results of the creating function. Examples include deburring, sanding, cleaning, etc. 3. Measurement functions: These are functions that measure the success of the other functions. Examples include SPC, gauging, inspections, etc. Manufacturing Process Functions Just as products have functions, manufacturing processes also have functions. The goal is to concisely list the function(s) for each process operation. The first step in improving any process is to make the current process visible by developing a process

SL3151Ch06Frame Page 270 Thursday, September 12, 2002 6:09 PM

270

Six Sigma and Beyond

flow diagram (a sequential flow of operations by people and/or equipment). This helps the team understand, agree, and define the scope. Three important questions exist for any existing process: 1. What do you think is happening? 2. What is actually happening? 3. What should be happening? Special reminder for manufacturing process functions: Remember, if the process flow diagram is too extensive for a “timely” FMEA, a risk assessment may be done on each process operation to narrow the scope. The PFMEA Function Questions Each manufacturing step typically has one or more functions. Determine what functions are associated with each manufacturing process step and then ask: 1. What does the process step do to the part? 2. What are you doing to the part/assembly? 3. What is the goal, purpose, or objective of this process step? For example, consider the pen assembly process (see Figure 6.22), which involves the following steps: 1. 2. 3. 4. 5. 6. 7. 8.

Inject ink into ink tube (0.835 cc) Insert ink tube into tip assembly housing (12 mm) Insert tip assembly into tip assembly housing (full depth until stop) Insert tip assembly housing into barrel (full depth until stop) Insert end cap into barrel (full depth until stop) Insert barrel into cap (full depth until stop) Move to dock (to dock within 8 seconds) Package and ship (12 pens per box)

Note: At the end of this function analysis you are ready to transfer the information to the FMEA form. Remember that another way to reduce the complexity or scope of the FMEA is to prioritize the list of functions and then take only the ones that the team collectively agrees are the biggest concerns. Describing the Failure Mode Anticipated The team must pose the question to itself, “How could this process fail to complete its intended function? Could the resulting workpiece be oversize, undersize, rough, eccentric, misassembled, deformed, cracked, open, shorted, leaking, porous, damaged, omitted, misaligned, out of balance, etc.?” The team members are trying to anticipate how the workpiece might fail to meet engineering requirements; at this point in their analysis they should stress how it could fail and not whether it will fail.

SL3151Ch06Frame Page 271 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

271

Ink

Ink tube

Inject ink into ink tube

Insert ink tube into tip assembly

Tip assembly housing

Tip assembly

Insert tip assembly into tip assembly housing Insert tip assembly housing

Barrel

Insert end cap into barrel

End cap

Insert barrel into cap

Cap

Move to dock Package and ship FIGURE 6.22 Pen assembly process.

The purpose of a process FMEA (PFMEA) is to analyze and evaluate a process on its ability to perform its functions. Therefore, the initial assumptions are: 1. The design intent meets all customer requirements. 2. Purchased materials and components comply with specifications. Once failure modes are determined under these assumptions, then determine other failure modes due to: 1. Design flaws that cause or lead to process problems 2. Problems with purchased materials, components, or services Describing the Effect(s) of the Failure The team must describe the effect of the failure on the component or assembly. What will happen as a result of the failure mode described? Will the component or

SL3151Ch06Frame Page 272 Thursday, September 12, 2002 6:09 PM

272

Six Sigma and Beyond

assembly be inoperative, intermittently operative, always on, noisy, inefficient, surging, not durable, inaccurate, etc.? After considering the failure mode, the engineer determines how this will manifest itself in terms of the component or assembly function. The open circuit causes an inoperative gage. The rough surface will cause excessive bushing wear. The scratched surface will cause noise. The porous casting will cause external leaks. The cold weld will cause reduced strength, etc. In some cases the process engineer (the leader) must interface with the product design engineer to correctly describe the effect(s) of a potential process failure on the component or total assembly. Describing the Cause(s) of the Failure The engineer anticipates the cause of the failure. The engineer is describing what conditions can bring about the failure mode. Locators are not flat and parallel. The handling system causes scratches on a shaft. Inadequate venting and gaging can cause misruns, porosity, and leaks. Inefficient die cooling causes die hot spots. Undersize condition can be caused by heat treat shrinkage, etc. The purpose of a process FMEA (PFMEA) is to analyze or evaluate a process on its ability to perform its functions (part characteristics). Therefore, the initial assumptions in determining causes are: • The design intent meets all customer requirements. • Purchased materials, components, and services comply with specifications. Then and only then, determine causes due to: • Design flaws that cause or lead to process problems • Problems with purchased materials, components, or services Typical causes associated with process FMEA include: Fatigue Poor surface preparation Improper installation Low torque Improper maintenance Inadequate clamping Misuse High RPM Abuse Inadequate venting Unclear instructions Tool wear Component interactions Overheating And so on

SL3151Ch06Frame Page 273 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

273

Estimating the Frequency of Occurrence of Failure The team must estimate the probability that the given failure mode will occur. This team is assessing the likelihood of occurrence, based on their knowledge of the process, using an evaluation scale of 1 to 10. A 1 would indicate a low probability of occurrence, whereas a 10 would indicate a near certainty of occurrence. Estimating the Severity of the Failure In estimating the severity of the failure, the team is weighing the consequence (effect) of the failure. The team uses the same 1 to 10 evaluation scale. A 1 would indicate a minor nuisance, while a 10 would indicate a severe consequence such as “motor inoperative, horn does not blow, engine seizes, no drive, etc.” Identifying Manufacturing Process Controls Manufacturing process controls consist of tests and analyses that detect causes or failure modes during process planning or production. Manufacturing process controls can occur at the specific operation in question or at a subsequent operation. There are three types of process controls, those that: 1. Prevent the cause from happening 2. Detect causes then lead to corrective actions 3. Detect failure modes then lead to corrective actions Manufacturing process controls should be based on process dominance factors. Dominance factors are process elements that generate significant process variation. Dominance factors are the predominant factors that contribute to problems in a process. Most processes have one or two dominant sources of variation. Depending on the source, there are tools that may be used to track these as well as monitor them. Table 6.8 gives a cross reference of the dominance factors and the tools that may be used for tracking them. The following list provides some very common dominance factors: • • • • • • • •

Setup Machine Operator Component or material Tooling Preventive maintenance Fixture/pallet/work holding Environment

Special note: Controls should target the dominant sources of variation. Manufacturing process control examples include:

SL3151Ch06Frame Page 274 Thursday, September 12, 2002 6:09 PM

274

Six Sigma and Beyond

TABLE 6.8 Manufacturing Process Control Matrix Dominance Factor

Attribute Data

Variable Date

Setup

Check sheet Checklist p or c chart Check sheet

X-bar/R chart X-MR chart Run chart X-bar/R chart X-MR chart X-bar/R chart X-MR chart Check sheet Supplier information Tool logs Capability study X-MR chart Time to failure chart Supplier information X-MR chart

Machine

Operator Component/material Tool

Preventive maintenance

Fixture/pallet/work holding

Environment

Check sheet Run chart Check sheet Supplier information Tool logs Check sheet p or c chart Time to failure chart Supplier information

Time to failure chart Check sheet p or c chart Check sheet

Time to failure chart X-bar/R chart X-MR chart Run chart X-MR chart

Statistical Process Control (SPC) • X-bar/R control charts (variable data) • Individual X-moving range charts (variable data) • p; n; u; c charts (attribute data) Non-statistical control • Check sheets, checklists, setup procedures, operational definitions/ instruction sheets • Preventive maintenance • Tool usage logs/change programs (PM) • Mistake proofing/error proofing/Poka Yoke • Training and experience • Automated inspection • Visual inspection It is very important to recognize that inspection is not a very effective control because it is a reactive task. Estimating the Detection of the Failure The detection is directly related to the controls available in the process. So the better the controls, the better the detection. The team in essence is estimating the probability

SL3151Ch06Frame Page 275 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

275

that a potential failure will be detected before it reaches the customer. The team members use the 1 to 10 evaluation scale. A 1 would indicate a very high probability that a failure would be detected before reaching the customer. A 10 would indicate a very low probability that the failure would be detected, and therefore, be experienced by the customer. For instance, a casting with a large hole would be readily detected and would be assessed as a 1. A casting with a small hole causing leakage between two channels only after prolonged usage would be assigned a 10. The team is assessing the chances of finding a defect, given that the defect exists. Calculating the Risk Priority Number The product of the estimates of occurrence, severity, and detection forms a risk priority number (RPN). This RPN then provides a relative priority of the failure mode. The higher the number, the more serious is the mode of failure considered. From the risk priority numbers, a critical items summary can be developed to highlight the top priority areas where actions must be directed. Recommending Corrective Action The basic purpose of an FMEA is to highlight the potential failure modes so that the engineer can address them after this identification phase. It is imperative that the engineer provide sound corrective actions or provide impetus for others to take sound corrective actions. The follow-up aspect is critical to the success of this analytical tool. Responsible parties and timing for completion should be designated in all corrective actions. Strategies for Lowering Risk: (Manufacturing) — High Severity or Occurrence Change the product or process design to: • Eliminate the failure cause or decouple the cause and effect • Eliminate or reduce the severity of the effect (recommend changes in design) Some “tools” to consider: • • • •

Benchmarking Brainstorming Mistake proofing TRIZ, etc.

Evaluate ideas using Pugh concept selection. Some specific examples: • • • • •

Developing a robust design (insensitive to manufacturing variations) Changing process parameters (time, temperature, etc.) Increasing redundancy, adding process steps Altering process inputs (materials, components, consumables) Using mistake proofing (Poka Yoke), reducing handling

SL3151Ch06Frame Page 276 Thursday, September 12, 2002 6:09 PM

276

Six Sigma and Beyond

Strategies for Lowering Risk: (Manufacturing) — High Detection Rating Change the process controls to: • Make failure mode easier to perceive • Detect causes prior to failure mode Some “tools” to consider: • Benchmarking • Brainstorming, etc. Evaluate ideas using Pugh concept selection. Some specific examples: • • • • •

Change testing and inspection procedures/equipment. Improve failure feedback or warning systems. Add sensors/feedback or feed forward systems. Increase sampling and/or redundancy in testing. Alter decision rules for better capture of causes and failures (i.e., more sophisticated tests).

At this stage, now you are ready to enter a brief description of the recommended actions, including the department and individual responsible for implementation, as well as both the target and finish dates, on the FMEA form. If the risk is low and no action is required write “no action needed.” For each entry that has a designated characteristic in the class[ification] column, review the issues that impact cause/occurrence, detection/control, or failure mode. Generate recommended actions to reduce risk. Special RPN patterns suggest that certain characteristics/root causes are important risk factors that need special attention. Guidelines for process control system: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Select the process. Conduct the FMEA on the process. Conduct gage system analysis. Conduct process potential study. Develop control plan. Train operators in control methods. Implement control plan. Determine long-term process capability. Review the system for continual improvement. Develop audit system. Institute improvement actions.

After FMEA:

SL3151Ch06Frame Page 277 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

277

1. Review the FMEA. 2. Highlight the high-risk areas based on the RPN. 3. Identify the critical and major characteristics based on your classification criteria. 4. Ensure that a control plan exists and is being followed. 5. Conduct capability studies. 6. Work on processes that have Cpk of less or equal to 1.33. 7. Work on processes that have Cpk greater than 1.33 to reduce variation and reach a Cpk of greater or equal to 2.0.

MACHINERY FMEA (MFMEA) A machinery FMEA is a systematic approach that applies the traditional tabular method to aid the thought process used by simultaneous engineering teams to identify the machine’s potential failure modes, potential effects, and potential causes of the potential failure modes and to develop corrective action plans that will remove or reduce the impact of the potential failure modes. Generally, the delivery of a MFMEA is the responsibility of the supplier who generates a functional MFMEA for system and subsystem levels. This is in contrast to a DFMEA where the responsibility is still on the supplier but now the focus is to generate transfer mechanisms, spindles, switches, cylinders, exclusive of assembly-level equipment. A typical MFMEA follows a hierarchical model in that it divides the machine into subsystems, assemblies, and lowest replaceable units. For example: Level 1: System level — generic machine Level 2: Subsystem level — electrical, mechanical, controls Level 3: Assembly level — fixtures/tools, material handling, drives Level 4: Component level And so on

IDENTIFY

THE

SCOPE

OF THE

MFMEA

Use the boundary diagram. Once the diagram has been completed, you can focus the MFMEA on the low MTBF and reliability values.

IDENTIFY

THE

FUNCTION

Define the function in terms of an active verb and a noun. Use a functional diagram or the P diagram to find the ideal function. Always focus on the intent of the system, subsystem, or component under investigation.

FAILURE MODE A failure is an event when the equipment/machinery is not capable of producing parts at specific conditions when scheduled or is not capable of producing parts or

SL3151Ch06Frame Page 278 Thursday, September 12, 2002 6:09 PM

278

Six Sigma and Beyond

performing scheduled operations to specifications. Machinery failure modes can occur in three ways: • Component defect (hard failure) • Failure observation (potential) • Abnormality of performance (constitutes the equipment as failed)

POTENTIAL EFFECTS The consequence of a failure mode on the subsystem is described in terms of safety and the big seven losses. (The big seven losses may be identified through warranty or historical data.) Describe the potential effects in terms of downtime, scrap, and safety issues. If a functional approach is used, then list the causes first before developing the effects listing. Associated with the potential effects is the severity, which is a rating corresponding to the seriousness of the effect of a potential machinery failure mode. Typical descriptions are: Downtime • Breakdowns: Losses that are a result of a functional loss or function reduction on a piece of machine requiring maintenance intervention. • Setup and adjustment: Losses that are a result of set procedures. Adjustments include the amount of time production is stopped to adjust process or machine to avoid defect and yield losses, requiring operator or job setter intervention. • Startup losses: Losses that occur during the early stages of production after extended shutdowns (weekends, holidays, or between shifts), resulting in decreased yield or increased scrap and defects. • Idling and minor stoppage: Losses that are a result of minor interruptions in the process flow, such as a process part jammed in a chute or a limit switch sticking, etc., requiring only operator or job setter intervention. Idling is a result of process flow blockage (downstream of the focus operation) or starvation (upstream of the focus operation). Idling can only be resolved by looking at the entire line/system. • Reduced cycle: Losses that are a result of differences between the ideal cycle time of a piece of machinery and its actual cycle time. Scrap • Defective parts: Losses that are a result of process part quality defects resulting in rework, repair, or scrap. • Tooling: Losses that are a result of tooling failures/breakage or deterioration/wear (e.g., cutting tools, fixtures, welding tips, punches, etc.). Safety • Safety considerations: Immediate life or limb threatening hazard or minor hazard.

SL3151Ch06Frame Page 279 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

279

SEVERITY RATING Severity is comprised of three components: • Safety of the machinery operator (primary concern) • Product scrap • Machinery downtime A rating should be established for each effect listed. Rate the most serious effect. Begin the analysis with the function of the subsystem that will affect safety, government regulations, and downtime of the equipment. A very important point here is the fact that a reduction in severity rating may be accomplished only through a design change. A typical rating is shown in Table 6.9. It should be noted that these guidelines may be modified to reflect specific situations. Also, the basis for the criteria may be changed to reflect the specificity of the machine and its real world usage.

CLASSIFICATION The classification column is not typically used in the MFMEA process but should be addressed if related to safety or noncompliance with government regulations. Address the failure modes with a severity rating of 9 or 10. Failure modes that affect worker safety will require a design change. Enter “OS” in the class column. OS (operator safety) means that this potential effect of failure is critical and needs to be addressed by the equipment supplier. Other notations can be used but should be approved by the equipment user.

POTENTIAL CAUSES The potential causes should be identified as design deficiencies. These could translate as: • Design variations, design margins, environmental, or defective components • Variation during the build/install phases of the equipment that can be corrected or controlled Identify first level causes that will cause the failure mode. Data for the development of the potential causes of failure can be obtained from: • Surrogate MFMEA • Failure logs • Interface matrix (focusing on physical proximity, energy transfer, material, information transfer) • Warranty data • Concern reports (things gone wrong, things gone right) • Test reports • Field service reports

Criteria Severity

Very high severity: affects operator, plant, or maintenance personnel safety and/or effects noncompliance with government regulations without warning Hazardous High severity: affects operator, with warning plant or maintenance personnel safety and/or effects noncompliance with government regulations with warning Very high Downtime of 8+ hours or the production of defective parts for over 2 hours High Downtime of 2–4 hours or the production of defective parts for up to 2 hours

Hazardous without warning

Effect Failure occurs every hour

Failure occurs every shift

Failure occurs every day Failure occurs every week

10

9

8

7

Rank

Probability of Failure

R(t) 37%

7

8

9

10

Rank

1 in 80

1 in 24

1 in 8

1 in 1

Alternate Criteria for Occurrence

Low

Very low

Detection

Machinery control will isolate the cause and failure mode after the failure has occurred, but will not prevent the failure from occurring

Team’s discretion depending on machine and situation

Team’s discretion depending on machine and situation

Present design controls cannot detect potential cause or no design control available

Criteria for Detection

7

8

9

10

Rank

280

R(t) = 20%

R(t) = 5%

R(t) < 1 or some MTBF

Criteria for Occurrence

TABLE 6.9 Machinery Guidelines for Severity, Occurrence, and Detection

SL3151Ch06Frame Page 280 Thursday, September 12, 2002 6:09 PM

Six Sigma and Beyond

Downtime of 60–120 min or the production of defective parts for up to 60 min Downtime of 30–60 min with no production of defective parts or the production of defective parts for up to 30 min Downtime of 15–30 min with no production of defective parts Downtime up to 15 min with no production of defective parts

Process parameter variability not within specification limits. Adjustments may be done during production; no downtime and no defects produced Process parameter variability within specification limits; adjustments may be performed during normal maintenance

Moderate

Very minor

None

Minor

Very low

Low

Criteria Severity

Effect

1

2

3

4

5

6

Rank R(t) = 60%

Criteria for Occurrence

Failure occurs every 5 years

Failure occurs every 2 years

R(t) = 98%

R(t) = 95%

Failure occurs R(t) = 85% every 6 months Failure occurs R(t) = 90% every year

Failure occurs R(t) = 78% every 3 months

Failure occurs every month

Probability of Failure

TABLE 6.9 Machinery Guidelines for Severity, Occurrence, and Detection

1

2

3

4

5

6

Rank

1 in 25,000

1 in 10,000

1 in 5000

1 in 2500

1 in 1000

1 in 350

Alternate Criteria for Occurrence

Team’s discretion depending on machine and situation Machinery controls will prevent an imminent failure and isolate the cause Team’s discretion depending on machine and situation

Machinery controls will provide an indicator of imminent failure

Team’s discretion depending on machine and situation

Criteria for Detection

Very high Machinery controls not required; design controls will detect a potential cause and subsequent failure almost every time

High

Medium

Detection

1

2

3

4

5

6

Rank

SL3151Ch06Frame Page 281 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA) 281

SL3151Ch06Frame Page 282 Thursday, September 12, 2002 6:09 PM

282

Six Sigma and Beyond

OCCURRENCE RATINGS Occurrence is the rating corresponding to the likelihood of the failure mode occurring within a certain period of time — see Table 6.8. The following should be considered when developing the occurrence ratings: • Each cause listed requires an occurrence rating. • Controls can be used that will prevent or minimize the likelihood that the failure cause will occur but should not be used to estimate the occurrence rating. Data to establish the occurrence ratings should be obtained from: • • • •

Service data MTBF data Failure logs Maintenance records

SURROGATE MFMEAS Current Controls Current controls are described as being those items that will be able to detect the failure mode or the causes of failure. Controls can be either design controls or process controls. A design control is based on tests or other mechanisms used during the design stage to detect failures. Process controls are those used to alert the plant personnel that a failure has occurred. Current controls are generally described as devices to: • • • •

Prevent the cause/mechanism failure mode from occurring Reduce the rate of occurrence of the failure mode Detect the failure mode Detect the failure mode and implement corrective design action

Detection Rating Detection rating is the method used to rate the effectiveness of the control to detect the potential failure mode or cause. The scale for ranking these methods is based on a 1 to 10 scale — see Table 6.8.

RISK PRIORITY NUMBER (RPN) The RPN is a method used by the MFMEA team to rank the various failure modes of the equipment. This ranking allows the team to attack the highest probability of failure and remove it before the equipment leaves the supplier floor. The RPN typically: • Has no value or meaning (Ratings and RPNs in themselves have no value or meaning. They should be used only to prioritize the machine’s potential

SL3151Ch06Frame Page 283 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

283

design weakness [failure mode] for consideration of possible design actions to eliminate the failures or make them maintainable.) • Is used to prioritize potential design weaknesses (root causes) for consideration of possible design actions • Is the product of severity, occurrence and detection (RPN = S × O × D) Special note on risk identification: Whereas it is true that most organizations using FMEA guidelines use the RPN for identifying the risk priority, some do not follow that path. Instead, they use a three path approach based on: Step 1: severity Step 2: criticality Step 3: detection This means that regardless of the RPN, the priority is based on the highest severity first, especially if it is a 9 or a 10, followed by the criticality, which is the product of severity and occurrence, and then the RPN.

RECOMMENDED ACTIONS • Each RPN value should have a recommended action listed. • The actions are designed to reduce severity, occurrence, and detection ratings. • Actions should address in order the following concerns: • Failure modes with a severity of 9 or 10 • Failure mode/cause that has a high severity occurrence rating • Failure mode/cause/design control that has a high RPN rating • When a failure mode/cause has a severity rating of 9 or 10, the design action must be considered before the engineering release to eliminate safety concerns.

DATE, RESPONSIBLE PARTY • Document the person, department, and date for completion of the recommended action. • Always place the responsible party’s name in this area.

ACTIONS TAKEN/REVISED RPN • After each action has been taken, document the action. • Results of an effective MFMEA will reduce or eliminate equipment downtime. • The supplier is responsible for updating the MFMEA. The MFMEA is a living document. It should reflect the latest design level and latest design actions. • Any equipment design changes need to be communicated to the MFMEA team.

SL3151Ch06Frame Page 284 Thursday, September 12, 2002 6:09 PM

284

Six Sigma and Beyond

REVISED RPN • Recalculate S, O, and D after the action taken has been completed. Always remember that only a change in design can change the severity. Occurrence may be changed by a design change or a redundant system. Detection may be changed by a design change or better testing or better design control. • MFMEA — A team needs to review the new RPN and determine if additional design actions are necessary.

SUMMARY In summary, the steps in conducting the FMEA are as follows: 1. 2. 3. 4. 5. 6. 7. 8.

Select a project and scope. If DFMEA, construct a block diagram. If PFMEA, construct a process flow diagram. Select an entry point based on the block or process flow diagram. Collect the data. Analyze the data. Calculate results (results must be data driven). Evaluate/confirm/measure the results. • Better off • Worse off • Same as before 9. Do it all over again.

SELECTED BIBLIOGRAPHY Chrysler Corporation, Ford Motor Company, and General Motors Corporation, Potential Failure Mode and Effect Analysis (FMEA) Reference Manual, 2nd ed., distributed by the Automotive Industry Action Group (AIAG), Southfield, MI, 1995. Chrysler Corporation, Ford Motor Company, and General Motors Corporation, Advanced Product Quality Planning and Control Plan, distributed by the Automotive Industry Action Group (A.I.A.G.), Southfield, MI, 1995. Chrysler Corporation, Ford Motor Company, and General Motors Corporation, Potential Failure Mode and Effect Analysis (FMEA) Reference Manual, 32nd ed., Chrysler Corporation, Ford Motor Company, and General Motors Corporation. Distributed by the Automotive Industry Action Group (AIAG), Southfield, MI, 2001. The Engineering Society for Advancing Mobility Land Sea Air and Space, Potential Failure Mode and Effects Analysis in Design FMEA and Potential Failure Mode and Effects Analysis in Manufacturing and Assembly Processes (Process FMEA) Reference Manual, SAE: J1739, The Engineering Society for Advancing Mobility Land Sea Air and Space, Warrendale, PA, 1994.

SL3151Ch06Frame Page 285 Thursday, September 12, 2002 6:09 PM

Failure Mode and Effect Analysis (FMEA)

285

The Engineering Society for Advancing Mobility Land Sea Air and Space, Reliability and Maintainability Guideline for Manufacturing Machinery and Equipment, SAE Practice Number M-110, The Engineering Society for Advancing Mobility Land Sea Air and Space, Warrendale, PA, 1999. Ford Motor Company, Failure Mode Effects Analysis: Training Reference Guide, Ford Motor Company — Ford Design Institute. Dearborn, MI, 1998. Kececioglu, D., Reliability Engineering Handbook, Vol. 1–2, Prentice Hall, Englewood Cliffs, NJ, 1991. Stamatis, D.H., Advanced Quality Planning, Quality Resources, New York, 1998. Stamatis, D.H., Failure Mode and Effect Analysis: FMEA from Theory to Execution, Quality Press, Milwaukee, 1995.

SL3151Ch06Frame Page 286 Thursday, September 12, 2002 6:09 PM

SL3151Ch07Frame Page 287 Thursday, September 12, 2002 6:07 PM

7

Reliability

Reliability n — may be relied on; trustworthiness, authenticity, consistency; infallibility, suggesting the complete absence of error, breakdown, or poor performance. In other words, when we speak of a reliable product, we usually expect such adjectives as dependable and trustworthy to apply. But to measure product reliability, we must have a more exact definition. The definition of reliability then, is: the probability that a product will perform its intended function in a satisfactory manner for a specified period of time when operating under specified conditions. Thus, the reliability of a system expresses the length of failure-free time that can be expected from the equipment. Higher levels of reliability mean less failure of the system and consequently less downtime. To measure reliability it is necessary to: • Relate probability to a precise definition of success or satisfactory performance • Specify the time base or operating cycles over which such performance is to be sustained • Specify the environmental or use conditions that will prevail Note: Theoretically, every product has a designed-in reliability function. This reliability function (or curve) expresses the system reliability at any point in time. As time increases the curve must drop, eventually reaching zero.

PROBABILISTIC NATURE OF RELIABILITY We cannot say exactly when a particular product will fail, but we can say what percentage of the products in use will fail by certain times. This is analogous to the reasoning used by insurance companies in defining mortality. We can state reliability in various ways: • The probability that a product will be performing its intended function at 5000 hours of use is 0.95. • The reliability at 5000 hours is 0.95 or 95%. • If we place 1000 units in use, 950 will still be operating with no failures at 5000 hours. Or to cite another example: • The reliability at 8000 hours is 0.80. • The unreliability at 8000 hours is 0.20.

287

SL3151Ch07Frame Page 288 Thursday, September 12, 2002 6:07 PM

288

Six Sigma and Beyond

From a service point of view, we may be interested in repair frequency and then we say that 20% of the units will have to be repaired by 8000 hours. Or the repair per hundred units (R/100) is 20 at 8000 hours. The important point is that the reliability is a metric expressing the probability of maintaining intended function over time and is measurable as a percentage.

PERFORMING THE INTENDED FUNCTION SATISFACTORILY A product fails when it ceases to function in a way that is satisfactory to the customer. Products rarely fail suddenly in the way that a light bulb does. Rather, they deteriorate over time. This eventually leads to unsatisfactory performance from the customer’s standpoint. Unsatisfactory performance can result from: • • • • • •

Excess vibrations Excess noise Intermittent operation Drift Catastrophic failure And many other possibilities

Unsatisfactory performance must be clearly spelled out. The customer’s perspective must be recognized in this process. There will usually be various levels of failure based on the customer’s perceived level of severity. The levels of severity are frequently grouped into two categories such as: • Major • Minor The severity of the failure to the customer must be documented and recognized in a Failure Definition and Scoring Criterion that precisely delineates how each incident on a system or equipment will be handled in regards to reliability and maintainability calculations. Such documents should be developed early in a design and development program so that all concerned are aware of the consequences of incidents that occur during product testing and in field use. The design team must be able to use the failure definition and scoring criterion to address product trade-offs. If the severity of a failure to the customer can be lowered by design changes, the failure definition and scoring criterion should promote such trade-offs.

SPECIFIED TIME PERIOD Products deteriorate with use and even with age when dormant. Longer lengths of usage imply lower reliability. For design purposes, target usage periods must be identified. Typical usage periods are:

SL3151Ch07Frame Page 289 Thursday, September 12, 2002 6:07 PM

Reliability

289

1. Warranty period(s): A warranty is a contract supplied with the product providing the user with a certain amount of protection against product failure. 2. Expected customer life: Customers have a reasonably consistent belief as to how long a product should last. This belief can be determined through a market survey. 3. Durability life: This is a measure of useful life, defining the number of operating hours (or cycles) until overhaul is required.

SPECIFIED CONDITIONS Different environments promote different failure modes and different failure rates for a product. The environmental factors that the product will encounter must be clearly defined. The levels (and rate of change) at which we want to address these environmental factors must also be defined.

ENVIRONMENTAL CONDITIONS PROFILE The environmental profile must include the level and rate of change for each environmental factor considered. Environmental factors include but are not limited to: • • • • • • • • • • • • • • • • • • • • • • • • •

Temperature Humidity Vibration Shock Corrosive materials Immersion Pressures, vacuum Salt spray Dust Cement floors/basements Ice/snow Lubricants Perfumes Magnetic fields Nuclear radiation Weather Contamination Antifreeze Gasoline fumes Rust inhibitors/under coatings Rain Soda pop/hot coffee Sunlight Electrical discharges And so on

SL3151Ch07Frame Page 290 Thursday, September 12, 2002 6:07 PM

290

Six Sigma and Beyond

Not all of these environmental conditions would be appropriate for a particular product. Each product must be considered in its individual operating environment and scenario. Environment must consider the environment induced from operating the product, the environment induced from external factors, and the environment induced by delivering the product to the customer.

RELIABILITY NUMBERS The reliability number attached to a product changes with: • Usage and environmental conditions • Customer’s perception of satisfactory performance At any product age (t) for a population of N products, the reliability at time t denoted by R(t) is R(t) = Number of survivors/N, which is equal to R(t) = 1 – (Number of failures/N) = 1 – Unreliability This is the reliability of this population of products at time t. The real world estimation of reliability is usually much more difficult due to products being sold over time with each having a different usage profile. Calendar time is known but product life on each product is not, while warranty systems monitor and record only failure.

INDICATORS USED

TO

QUANTIFY PRODUCT RELIABILITY

Several metrics are in common use to indicate product reliability. Some of these actually quantify unreliability. Some of the metrics follow: • MTBF — The mean time between failures, also MTTF, MMBF, MCTF. MTBF = 120 hours means that on the average a failure will occur with every 120 hours of operation. • Failure rate — The rate of failures per unit of operating time. λ = 0.05/hour means that one failure will occur with every 20 hours of operation, on the average. • R/100 (or R/1000) — The number of warranty claims per 100 (or 1000) products sold. R/100 = 7 means that there are seven warranty claims for every 100 products sold. • Reliability number — The reliability of the product at some specific time. R = 90% means that 9 out of 10 products work successfully for the specified time.

SL3151Ch07Frame Page 291 Thursday, September 12, 2002 6:07 PM

Reliability

291

RELIABILITY AND QUALITY Customers and product engineers frequently use the terms reliability and quality interchangeably. Ultimately, the customer defines quality. Customers want products that meet or exceed their needs and expectations, at a cost that represents value. This expectation of performance must be met throughout the customer’s expected life for the particular product. Quality is usually recognized as a more encompassing term including reliability. Some quality characteristics are: Psychological • Taste • Beauty, style • Status Technological • Hardness • Vibration • Noise • Materials (bearings, belts, hoses, etc.) Time-oriented • Reliability • Maintainability Contractual • Warranty Ethical • Honesty of repair shop • Experience and integrity of sales force

PRODUCT DEFECTS Quality defects are defined as those that can be located by conventional inspection techniques. (Note: for legal reasons, it is better to identify these defects as nonconformances.) Reliability defects are defined as those that require some stress applied over time to develop into detectable defects. What causes product failure over time? Some possibilities are: • • • • • • • • •

Design Manufacturing Packaging Shipping Storage Sales Installation Maintenance Customer duty cycle

SL3151Ch07Frame Page 292 Thursday, September 12, 2002 6:07 PM

292

Six Sigma and Beyond

CUSTOMER SATISFACTION The ultimate goal of a product is to satisfy a customer from all aspects of cost, performance, reliability, and maintainability. The customer trades off these parameters when making a decision to buy a product. Assuming that we are designing a product for a certain market segment, cost is determined within limits. The tradeoffs are as follows: 1. Performance parameters are the designed-in system capabilities such as acceleration, top speed, rate of metal removal, gain, ability to carry a 5ton payload up a 40 degree grade without overheating, and so on. 2. The reliability of equipment expresses the length of failure-free time that can be expected from the equipment. Higher levels of reliability mean less failure of the equipment and consequently less downtime and loss of use. Although we will attach reliability numbers to products, it should be recognized that the customer’s perspective interprets reliability as the ability of a product to perform its intended function for a given period of time without failure. This concept of failure-free operation is becoming more and more fixed in the mind of the customer. This is true whether the customer is purchasing an automobile, a machine tool, a computer system, a refrigerator, or an automatic coffee maker. 3. Maintainability is defined as the probability that a failed system is restored to operable condition in a specified amount of downtime. 4. Availability is the probability that at any time, the system is either operating satisfactorily or is ready to be operated on demand, when used under stated conditions. The availability might also be looked at as the ability of equipment, under combined aspects of its reliability, maintainability, and maintenance support, to perform its required function at a stated instant of time. This availability includes the built-in equipment features as well as the maintenance support function. Availability combines reliability and maintainability into one measure. There are different kinds of availability that are calculated in different ways — see Von Alven (1964) and ANSI/IEEE (1988). The most popular availabilities are achieved availability and inherent availability. a. Achieved availability includes all diagnostic, repair, administrative, and logistic times. This availability is dependent on the maintenance support system. Achieved availability can be calculated as A = Operating Time/(Operating Time + Unscheduled Time) b. Inherent availability only includes operating time and active repair time addressing the built-in capabilities of the equipment. Inherent availability is calculated as

A=

MTBF MTBF + MTTR

SL3151Ch07Frame Page 293 Thursday, September 12, 2002 6:07 PM

Reliability

293

Infant mortality

Failure rate

Normal life

Wear out

Time FIGURE 7.1 Bathtub curve.

where MTTR = mean time-to-repair and the MTTR is for the active repair time. 5. Active repair time is that portion of downtime when the technicians are working on the system to repair the failure situation. It must be understood that the different availabilities are defined for various time-states of the system. 6. Serviceability is the ease with which machinery and equipment can be repaired. Here repair includes diagnosis of the fault, replacement of the necessary parts, tryout, and bringing the equipment back on line. Serviceability is somewhat qualitative and addresses the ease by which the equipment, as designed, can be diagnosed and repaired. It involves factors such as accessibility to test points, ease of removal of the failed components, and ease of bringing the system back on line.

PRODUCT LIFE

AND

FAILURE RATE

Let us assume that we have released a population of products to the marketplace. The failure rate is observed as the products age. The shape of the failure rate is referred to as a bathtub curve (see Figure 7.1). Here we have overemphasized the different parts of the curve for illustration. This bathtub curve has three distinct regions: 1. Infant mortality period: During the infant mortality period the population exhibits a high failure rate, decreasing rapidly as the weaker products fail. Some manufacturers provide a “burn-in” period for their products to help eliminate infant mortality failures. Generally, infant mortality is associated with manufacturing issues. Examples are: • Poor welds • Contamination • Improper installation • And so on 2. Useful life period: During this period the population of products exhibits a relatively low and constant failure rate. It is explained using the stress – strength inference model for reliability. This model identifies the stress

SL3151Ch07Frame Page 294 Thursday, September 12, 2002 6:07 PM

294

Six Sigma and Beyond

distribution that represents the combined stressors acting on a system at some point in time. The strength distribution represents the piece-to-piece variability of components in the field. The inference area is indicative of a potential failure when stresses exceed the strength of a component. In other words, any failure in this period is a factor of the designed-in reliability. Examples are: • Low safety factors • Abuse • Misapplication • Product variability • And so on 3. Wear out period: At the onset of wear out, the failure rate starts to increase rapidly. When the failure rate becomes high, replacement or major repair must be performed if the product is to be left in service. Wear out is due to a number of forces such as: • Frictional wear • Chemical change • Maintenance practices • Fatigue • Corrosion or oxidation • And so on In conjunction with the bathtub curve there are two more items of concern. The first one is the hazard rate (or the instantaneous failure rate) and the second, the ROCOF plot. The hazard rate is the probability that the product will fail in the next interval of time (or distance or cycles). It is assumed the product has survived up to that time. For example, there is a one in twenty chance that it will crack, break, bend, or fail to function in the next month. Typically, hazard rate is shown as h(t ) =

f (t ) f (t ) = 1 − F (t ) R(t )

where h(t) = hazard rate; f(t) = probability density function [PDF: f(t) = λe–λt]; F(t) = cumulative distribution function [CDF: F(t) = 1 – e–λt; and R(t) = reliability at time t [R(t) = 1 – F(t) = 1 – (1 – e–λt) = e–λt]. The Rate of Change of Failure or Rate of Change of Occurrence of Failure (ROCOF), on the other hand, is a visual tool that helps the engineer to analyze situations where a lot of data over time has been accumulated. Essentially, its purpose is the same as that of the reliability bathtub curve, that is, to understand the life stages of a product or process and take the appropriate action. A typical ROCOF plot (for warranty item) will display an early (decreasing rate) and useful life (constant rate) performance. If wear out is detected, it should be investigated. Knowing what is happening to a product from one region of the bathtub curve to the next helps the engineer specify what failed hardware to collect and aids with calibrating the severity of development tests.

SL3151Ch07Frame Page 295 Thursday, September 12, 2002 6:07 PM

Reliability

295

If the number of failures is small, the ROCOF plot approach may be difficult to interpret. When that happens, it is recommended that a “smoothing” approach be taken. The typical smoothing methodology is to use log paper for the plotting. Obviously, many more ways and more advanced techniques exist. It must be noted here that most statistical software provides this smoothing as an option for the data under consideration. See Volume III for more details on smoothing.

PRODUCT DESIGN AND DEVELOPMENT CYCLE Developing a product that can be manufactured economically and consistently to be delivered to the marketplace in quantity and that will work satisfactorily for the customer takes a well established and precisely controlled design and development cycle. Events must be scheduled to occur at precise times to phase the product into the marketplace. To develop a new internal combustion engine for an automobile takes about a three-year design cycle (down recently from five years), while a new minicomputer takes about 18 months. Although the timing may be different for different companies, the activities comprising a design and development cycle are similar. The following is representative of the activities in a product development cycle: • Market research • Forecast need. • Forecast sales. • Understand who the customer is and how the product will be used. • Set broad performance objectives. • Establish program cost objectives. • Establish technical feasibility. • Establish manufacturing capacity. • Establish reliability and maintainability (R&M) requirements. • Understand governmental regulations. • Understand corporate objectives. • Concept phase • Formulate project team. • Formulate design requirements. • Establish real world customer usage profile. • Develop and consider alternatives. • Rank alternatives considering R&M requirements. • Review quality and reliability history on past products. • Assess feasibility of R&M requirements. • Design phase • Prepare preliminary design. • Perform design calculations. • Prepare rough drawings. • Compare alternatives to pursue. • Evaluate manufacturing feasibility of design approach (design for manufacturability and assembly).

SL3151Ch07Frame Page 296 Thursday, September 12, 2002 6:07 PM

296

Six Sigma and Beyond

•

•

•

• • •

• Complete detailed design. • Perform a design failure mode and effect analysis (FMEA). • Complete detailed design package. • Update FMEA to reflect current design and details. • Develop design verification plan. • Develop R&M model for product. • Estimate product R&M using current design approach. Prototype program • Build components and prototypes. • Write test plan. • Perform component/subsystem tests. • Perform system test. • Eliminate design weaknesses. • Estimate reliability using growth techniques. Manufacturing engineering • Process planning • Assembly planning • Capability analyses • Process FMEA Finalized design • Consider test results. • Consider manufacturing engineering inputs (design for manufacturability/assembly). • Make design changes. Freeze design Release to manufacturing Engineering changes • Manufacturing experience • Field experience

RELIABILITY

IN

DESIGN

The cost of unreliability is: • • • •

High warranty costs Field campaigns Loss of future sales Cost of added field service support

It has been demonstrated in the marketplace that highly reliable products (failure free) gain market share. A very classic example of this is the American automotive market. In the early 1960s, American manufacturers were practically the only game in town with GM capturing some 60% of the market. Since then, progressively and on a yearly basis the market has shifted to the point where Flint (2001) reports that now GM has a shade over 25% without trucks and Saab, Ford 14.7% without Volvo and Jaguar, and Chrysler about 5%. The projections for the 2002 model year are

SL3151Ch07Frame Page 297 Thursday, September 12, 2002 6:07 PM

Reliability

297

not any better with GM capturing only 25%, Ford 15%, and Chrysler 6%. The sad part of the automotive scene is that GM, Ford, and DaimlerChrysler have lost market share, and sales are continually nudging down with no end in sight. That is, as Flint (2001, p. 21) points out, “they are not going to recover that market share, not in the short term, not in the next five to ten years.” The evidence suggests that the mission of a reliability program is to estimate, track, and report the reliability of hardware before it is produced. The reliability of the equipment must be reported at every phase of design and development in a consistent and easy-to-understand format. Warranty cost is an expensive situation resulting from poor manufacturing quality and inadequate reliability. For example, the chairman and chief executive of Ford Motor Company, Jacques Nasser, in the 1st quarter of 2001 leadership cascading meeting made the statement that in 1999, there were 2.1 times as many vehicles recalled as were sold. In 2000, there were six times as many. By way of comparison: In 1994, according to an article in USA Today, the cost of warranty for a Chrysler automobile was as high as $850 per vehicle. From the same article, one could deduce that the cost per vehicle for General Motors was about $350 and for Ford $650. This would be to cover the 36,000 mile warranty in effect at that time. In 2000, the warranty cost for Chrysler was about $1,300, GM about $1,200, and Ford about $850 (Mayne et al., 2001). For each car sold, the manufacturer must collect and retain this expense in a warranty account.

COST

OF

ENGINEERING CHANGES

AND

PRODUCT LIFE CYCLE

The stage of product development/manufacturing and the cost of an engineering change have been estimated many times by many different industries and various trade magazines as a cost that grows by a factor of five to ten as one moves from early design to manufacturing. Typical figures for this high cost are • Prototype stage: $100,000 Therefore, reliability can play an important role in designing products that will satisfy the customer and will prove durable in the real world usage application. The focus of reliability is to design, identify, and detect early potential concerns at a point where it is really cost effective to do so. Reliability must be valued by the organization and should be a primary consideration in all decision making. Reliability techniques and disciplines are integrated into system and component planning, design, development, manufacturing, supply, delivery, and service processes. The reliability process is tailored to fit individual business unit requirements and is based on common concepts that are focused on producing reliable products and systems, not just components.

SL3151Ch07Frame Page 298 Thursday, September 12, 2002 6:07 PM

298

Six Sigma and Beyond

Any organization committed to satisfy the customer’s expectations for reliability (and value) throughout the useful life of its product must be concerned with reliability. For without it, the organization is doomed to fail. The total reliability process includes robustness concepts and methods that are integrated into the organization’s timing schedule and overall business system. Cross-functional teams and empowered individuals are key to the successful implementation of any reliability program. Reliability concepts and methods are generally thought of as a proprietary domain of only the product development department or community. That is not completely true. Reliability may be used anywhere there is a need for design and development work, such as manufacturing and tooling. However, it does not address actions specifically targeted at manufacturing and assembly. This is the reason why under Design for Six Sigma (DFSS), reliability becomes very important from the “get go.” To be sure, reliability currently does not include all the elements of the Advanced Product Quality Plan (APQP), but it is compatible with APQP. It outlines the three quality and reliability phases that all program teams and supporting organizations should go through in the product development process to achieve a more reliable and robust product. The three phases stress useful life reliability, focusing specifically on the deployment of customer-driven requirements, designing in robustness, and verifying that the designs meet the requirements.

RELIABILITY

IN THE

TECHNOLOGY DEPLOYMENT PROCESS

Technology is ever changing on all fronts. Customers expect increased reliability and better quality for a reasonable cost. Reliability may indeed play a major role in bringing technology, customer satisfaction, and lower cost into reality. Let us then try to understand the process of support and the cascading of requirements throughout the Technology Deployment Process (TDP). Understanding the TDP begins with the recognition that this process has three phases and each phase has specific requirements. The three phases are pre deployment process, core engineering process, and quality support. In the pre deployment process, there are three stages with very specific inputs and outputs. In core engineering, the development of generic requirements begins, and in quality support, the “best” reliability practices are developed. 1. Pre-Deployment Process Three stages are involved here. They are: 1. Identify/select new technologies: The main function of this stage is to identify and select technology for reliable and robust products that meet future customer needs or wants. In essence, here we are to develop and understand: • Customer wants process • Competitive analysis • Technology strategy/roadmap

SL3151Ch07Frame Page 299 Thursday, September 12, 2002 6:07 PM

Reliability

299

2. Develop/optimize technology to achieve concept readiness: The main function of this stage is to sufficiently develop and prove through analytical and/or surrogate testing that the technology meets the functional and reliability requirements for customer wants or needs under real world usage conditions. In essence, here we are to generate, understand, and develop readiness through: • Reviewing quality history of similar systems/concepts • Understanding real world usage profile • Defining functional requirements of system • Planning for robustness • Reviewing quality/reliability/durability reports or worksheets 3. Develop/optimize technology to achieve implementation readiness: The main function of this stage is to optimize the technology to meet functional and/or reliability requirements. Additionally, the aim is to demonstrate that the technology is robust and reliable under real world usage conditions. In essence, here we are to further understand the requirements by: • Refining design requirements • Designing for robustness • Verifying the design • Reviewing quality/reliability/durability reports or worksheets 2. Core Engineering Process Develop generic requirements for forward models by providing product lines with generic information on system robust design, such as case studies, system P-diagrams, measurement of ideal functions, etc. In this stage, we also conduct competitive technical information analysis to our potential product lines through test-thebest and reliability benchmarking. Some of the specific tools we may use are: • • • • • • •

System design specification guidelines Real world usage demographics Failure mode and effect analysis Key life testing Fault tree analysis Design verification process And so on

The idea here is to be able to develop common-cause problem resolution, that is, to be able to identify common-cause problems/root causes across the product line(s) and champion corrective action by following reliability disciplines. In essence then, core engineering should: • Prioritize concerns • Identify root causes • Determine/incorporate corrective action

SL3151Ch07Frame Page 300 Thursday, September 12, 2002 6:07 PM

300

Six Sigma and Beyond

• Validate improvements • Champion implementation across product line(s) 3. Quality Support Identify best reliability practices and lead the process standardization and simplification. Develop a toolbox and provide reliability consultation.

RELIABILITY MEASURES — TESTING The purpose of performing a reliability test is to answer the question, “Does the item meet or exceed the specified minimum reliability requirement?” Reliability testing is used to: • Determine whether the system conforms to the specified, quantitative reliability requirements • Evaluate the system’s expected performance in the warranty period and its compliance to the useful life targets as defined by corporate policy • Compare performance of the system to the goal that was established earlier • Monitor and validate reliability growth • Determine design actions based on the outcomes of the test In addition to their other uses, the outcomes of reliability testing are used as a basis for design qualification and acceptance. Reliability testing should be a natural extension of the analytical reliability models, so that test results will clarify and verify the predicted results, in the customer’s environment.

WHAT IS

A

RELIABILITY TEST?

A reliability test is effectively a “sampling” test in that it involves a sample of objects selected from a “population.” From the sample data, some statement(s) are made about the population parameter(s). In reliability testing, as in any sampling test: • The sample is assumed to be representative of the population. • The characteristics of the sample (e.g., sample mean) are assumed to be an estimate of the true value of the population characteristics (e.g., population mean). A key factor in reliability test planning is choosing the proper sample size. Most of the activity in determining sample size is involved with either: 1. Achieving the desired confidence that the test results give the correct information 2. Reducing the risk that the test results will give the wrong information

SL3151Ch07Frame Page 301 Thursday, September 12, 2002 6:07 PM

Reliability

301

WHEN DOES RELIABILITY TESTING OCCUR? Prior to the time that hardware is available, simulation and analysis should be used to find design weaknesses. Reliability testing should begin as soon as hardware is available for testing. Ideally, much of the reliability testing will occur “on the bench” with the testing of individual components. There is good reason for this: The effect of failure on schedule and cost increases progressively with the program timeline. The later in the process that the failure and corrective action are found, the more it costs to correct and the less time there is to make the correction. Some key points to remember regarding test planning: • Develop the reliability test plan early in the design phase. • Update the plan as requirements are added. • Run the formal reliability testing according to the predetermined procedure. This is to ensure that results are not contaminated by development testing or procedural issues. • Develop the test plan in order to get the maximum information with the fewest resources possible. • Increase test efficiency by understanding stress/strength and acceleration factor relationships. This may require accelerated testing, such as AST (Accelerated Stress Test), which will increase the information gained from a test program. • Make sure your test plan shows the relationship between development testing and reliability testing. While all data contribute to the overall knowledge about a system, other functional development testing is an opportunity to gain insight into the reliability performance of your product. Note: A “control sample” should be maintained as a reference throughout the reliability testing process. Control samples should not be subjected to any stresses other than the normal parametric and functional testing.

RELIABILITY TESTING OBJECTIVES When preparing the test plan, keep these objectives in mind: • Test with regard to production intent. Make sure the sample that is tested is representative of the system that the customer will receive. This means that the test unit is representative of the final product in all areas including materials (metals, fasteners, weight), processes (machining, casting, heat treat), and procedures (assembly, service, repair). Of course, consider that these elements may change or that they may not be known. However, use the same production intent to the extent known at the time of the test plan. • Determine performance parameters before testing is started. It is often more important in reliability evaluations to monitor the percentage change in a parameter rather than the performance to specification.

SL3151Ch07Frame Page 302 Thursday, September 12, 2002 6:07 PM

302

Six Sigma and Beyond

• Duplicate/simulate the full range of the customer stresses and environments. This includes testing to the 95th percentile customer. (For most organizations this percentile is the default. Make sure you identify what is the exact percentile for your organization.) • Quantify failures as they relate to the system being tested. A failure results when a system does not perform to customer expectations, even if there is no actual broken part. Remember, • Customer requirements include the specifications and requirements of internal customers and regulatory agencies as well as the ultimate purchaser. • You should structure testing to identify hardware interface issues as they relate to the system being tested. Sudden-Death Testing Sudden-death testing allows you to obtain test data quickly and reduces the number of test fixtures required. It can be used on a sample as large as 40 or more or as small as 15. Sudden-death testing reduces testing time in cases where the lower quartile (lower 25%) of a life distribution is considerably lower than the upper quartile (upper 25%). The philosophy involved in sudden-death testing is to test small groups of samples to a first failure only and use the data to determine the Weibull distribution of the component. The method is as follows: 1. Choose a sample size that can be divided into three or more groups with the same number of items in each group. Divide the sample into three or more groups of equal size and treat each group as if it were an individual assembly. 2. Test all items in each group concurrently until there is a first failure in that group. Testing is then stopped on the remaining units in that group as soon as the first unit fails, hence the name “sudden death.” 3. Record the time to first failure in each group. 4. Rank the times to failure in ascending order. 5. Assign median ranks to each failure based on the sample size equal to the number of groups. Median rank charts are used for this purpose. 6. Plot the times to failure vs. median ranks on Weibull paper. 7. Draw the best fit line. (Eye the line or use the regression model.) This line represents the sudden-death line. 8. Determine the life at which 50% of the first failures are likely to occur (B50 life) by drawing a horizontal line from the 50% level to the suddendeath line. Drop a vertical line from this point down. 9. Find the median rank for the first failure when the sample size is equal to the number of items in each subgroup. Again, refer to the median rank charts. Draw a horizontal line from this point until it intersects the vertical line drawn in the previous step.

SL3151Ch07Frame Page 303 Thursday, September 12, 2002 6:07 PM

Reliability

303

TABLE 7.1 Failure Rates with Median Ranks Failure Order Number

Life Hours

Median Ranks, %

1 2 3 4 5

65 120 155 200 300

12.95 31.38 50.00 68.62 87.06

10. Draw a line parallel to the sudden-death line passing through the intersection point from step 9. This line is called the population line and represents the Weibull distribution of the population. Sudden-death testing is a good method to use to determine the failure distribution of the component. (Note: Only common failure mechanisms can be used for each Weibull distribution. Care must be taken to determine the true root cause of all failures. Failure must be related to the stresses applied during the test.) EXAMPLE Assume you have a sample of 40 parts from the same production run available for testing purposes. The parts are divided into five groups of eight parts as shown below: Group Group Group Group Group

l 2 3 4 5

12345678 12345678 12345678 12345678 12345678

All parts in each group are put on test simultaneously. The test proceeds until any one part in each group fails. At that time, testing stops on all parts in that group. In the test, we experience the following first failures in each group: Group Group Group Group Group

1 2 3 4 5

Part Part Part Part Part

#3 #4 #1 #5 #7

fails fails fails fails fails

at at at at at

120 hours 65 hours 155 hours 300 hours 200 hours

Failure data are arranged in ascending hours to failure, and their median ranks are determined based on a sample size of N = 5. (There are five failures, one in each of five groups.) The chart in Table 7.1 illustrates the data. The median rank percentage for each failure is derived from the median rank (Table 7.2) for five samples. If the life hours and median ranks of the five failures are plotted on Weibull paper, the resulting line is called the sudden-death line. The sudden-death line represents

SL3151Ch07Frame Page 304 Thursday, September 12, 2002 6:07 PM

304

Six Sigma and Beyond

TABLE 7.2 Median Ranks Rank Order 1 2 3 4 5 6 7 8 9 10 Rank Order 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Sample size 1

2

3

4

5

6

7

8

9

10

50.0

29.3 70.7

20.6 50.0 7 9.4

15.9 38.6 61.4 84.1

12.9 31.4 50.0 68.6 87.1

10.9 26.4 42.1 57.9 73.9 89.1

9.4 22.8 36.4 50.0 63.6 77.2 90.6

8.3 20.1 3G.1 44:0 56.0 67.9 79.9 91.7

7.4 18.0 Z8.6 39.3 50.0 60.7 71.4 82.0 92.6

6.7 16.2 25.9 35.5 45.2 54.8 64.5 74.1 83.8 93.3

11

12

13

14

15

16

17

18

19

20

6.1 14.8 23.6 32.4 41.2 50.0 58.8 67.6 76.4 85.2 93.9

5.6 13.6 21.7 29.8 37.9 46.0 54.0 62.1 70.2 78.3 86.4 94.4

5.2 12.6 20.0 27.5 35.0 42.5 50.0 57.5 65.0 72.5 80.0 87.4 94.8

4.8 1 1.7 18.6 25.6 32.6 39.5 46.5 53.5 60.5 67.4 74.4 81.4 88.3 95.2

4.5 10.9 17.4 23.9 30.4 37.0 43.5 50.0 56.5 63.0 69.5 76.1 82.6 89.1 95.5

4.2 10.3 16.4 22.5 28.6 34.7 40.8 46.9 53.1 59.2 65.3 71.4 77.5 83.6 89.7 95.8

4.0 9.7 15.4 21.2 26.9 32.7 38.5 44.2 50.0 55.8 61.5 67.3 73.1 78.8 84.6 90.3 96.0

3.8 9.2 14.6 20.0 25.5 30.9 36.4 41.8 47.3 52.7 58.2 63.6 69.1 74.5 80.0 85.4 90.8 96.2

3.6 8.7 13.8 19.0 24.2 29.3 34.5 39.7 44.8 50.0 55.2 60.3 65.5 70.7 75.8 81.0 86.2 91.3 96.4

3.4 8.3 13.1 18.1 23.0 27.9 32.8 37.7 42.6 47.5 52.5 57.4 62.3 67.2 72.1 77.0 81.9 86.9 91.7 96.6

Sample Size

the cumulative distribution that would result if five assemblies failed, but it actually represents five measures of the first failure in eight of the population. The median life point on the sudden-death line (point at which 50% of the failures occur) will correspond to the median rank for the first failure in a sample of eight, which is 8.30%. The population line is drawn parallel to the sudden-death line through a point plotted at 8.30% and at the median life to first failure as determined above. This estimate of the population’s minimum life is just as reliable as the one that would have been obtained if all 40 parts were tested to failure.

SL3151Ch07Frame Page 305 Thursday, September 12, 2002 6:07 PM

Reliability

305

Accelerated Testing Accelerated testing is another approach that may be used to reduce the total test time required. Accelerated testing requires stressing the product to levels that are more severe than normal. The results that are obtained at the accelerated stress levels are compared to those at the design stress or normal operating conditions. We will look at examples of this comparison during this section. We use accelerated testing to: • Generate failures, especially in components that have long life under normal conditions • Obtain information that relates to life under normal conditions • Determine design/technology limits of the hardware Accelerated testing is accomplished by reducing the cycle time, such as by: • Compressing cycle time by reducing or eliminating idle time in the normal operating cycle • Overstressing There are some pitfalls in using accelerated testing: • Accelerated testing can cause failure modes that are not representative. • If there is little correlation to “real” use, such as aging, thermal cycling, and corrosion, then it will be difficult to determine how accelerated testing affects these types of failure modes.

ACCELERATED TEST METHODS There are many test methods that can be used for accelerated testing. This section covers: • • • •

Constant-stress testing Step-stress testing Progressive-stress testing AST/PASS testing

Before we discuss the methods, keep in mind that any product may be subjected to multiple stresses and combinations of stresses. The stresses and combinations are identified very early in the design phase. When accelerated tests are run, ensure that all the stresses are represented in the test environment and that the product is exposed to every stress.

CONSTANT-STRESS TESTING In constant-stress testing, each test unit is run at constant high stress until it fails or its performance degrades. Several different constant stress conditions are usually

SL3151Ch07Frame Page 306 Thursday, September 12, 2002 6:07 PM

306

Six Sigma and Beyond

employed, and a number of test units are tested at each condition. Some products run at constant stress, and this type of test represents actual use for those products. Constant stress will usually provide greater accuracy in estimating time to failure. Also, constant-stress testing is most helpful for simple components. In systems and assemblies, acceleration factors often differ for different types of components.

STEP-STRESS TESTING In step-stress testing, the item is tested initially at a normal, constant stress for a specified period of time. Then the stress is increased to a higher level for a specified period of time. Increases continue in a stepped fashion. The main advantage of step-stress testing is that it quickly yields failure, because increasing stress ensures that failures occur. A disadvantage is that failure modes that occur at high stress may differ from those at normal use conditions. Quick failures do not guarantee more accurate estimates of life or reliability. A constantstress test with a few failures usually yields greater accuracy in estimating the actual time to failure than a shorter step-stress test; however, we may need to do both to correlate the results so that the results of the shorter test can be used to predict the life. (Always remember that failures must be related to the stress conditions to be valid. Other test discrepancies should be noted and repaired and the testing continued.)

PROGRESSIVE-STRESS TESTING Progressive-stress testing is step-stress testing carried to the extreme. In this test, the stress on a test unit is continuously increased, rather than being increased in steps. Usually, the accelerating variable is increased linearly with time. Several different rates of increase are used, and a number of test units are tested at each rate of increase. Under a low rate of increase of stress, specimens tend to live longer and to fail at lower stress because of the natural aging effects or cumulative effects of the stress on the component. Progressive-stress testing has some of the same advantages and disadvantages as step-stress testing.

ACCELERATED-TEST MODELS The data from accelerated tests are interpreted and analyzed using different models. The model that is used depends upon the: • Product • Testing method • Accelerating variables The models give the product life or performance as a function of the accelerating stress. Keep these two points in mind as you analyze accelerated test data: 1. Units run at a constant high stress tend to have shorter life than units run at a constant low stress.

SL3151Ch07Frame Page 307 Thursday, September 12, 2002 6:07 PM

Reliability

307

2. Distribution plots show the cumulative percentage of the samples that fails as a function of time. In fact, over time the smoothing of the curve in the shape of an “S” is indeed the estimate of the actual cumulative percentage failing as a function of time. Two common models — although appropriate for component level testing — that deal specifically with accelerated tests are: 1. Inverse Power Law Model 2. Arrhenius Model Inverse Power Law Model The inverse power law model applies to many failure mechanisms as well as to many systems and components. This model assumes that at any stress, the time to failure is Weibull distributed. Thus: • The Weibull shape parameter β has the same value for all the stress levels. • The Weibull scale parameter θ is an inverse power function of the stress. The model assumes that the life at rated stress divided by the life at accelerated stress is equal to the quantity, accelerated stress divided by rated stress, raised to the power n, where: n = acceleration factor determined from the slope of the S-N diagram on the log-log scale. Using the above information, we can say that: θu = θa[Accelerated stress/Rated stress]n where θu = life at the rated (usage) stress level; θa = life at the accelerated stress level; and n = acceleration factor determined from the slope of the S-N diagram on the log-log scale. EXAMPLE Let us assume we tested 15 incandescent lamps at 36 volts until all items in the sample failed. A second sample of 15 lamps was tested at 20 volts. Using these data, we will determine the characteristic life at each test voltage and use this information to determine the characteristic life of the device when operated at 5 volts. From the accelerated test data:

θ20 volts = 11.7 hours θ36 volts = 2.3 hours Since we know these two factors, we can determine the acceleration factor, n. We have the following relationship:

SL3151Ch07Frame Page 308 Thursday, September 12, 2002 6:07 PM

308

Six Sigma and Beyond

[Life at rated stress/life at accelerated stress] = [Accelerated stress/rated stress]n This relationship becomes

[θ20 volts/θ36 volts] = [36 volts/20 volts]n Substituting the values for theta 20 v and theta 36 v we have

11.7hrs  36v  = 2.3hrs  20 v 

n

Therefore,

n = 2.767 Now we can use the following equation to determine the characteristic life at 5 volts:

θu = θa [Accelerated stress/Rated stress]n  36  θ5 v = θ36 v [Accelerated stress/Rated stress] = 2.3    5 n

2.767

= 542 hours

The characteristic life at 5 volts is 542 hours.

The reader must be very careful here because not all electronic parts or assemblies will follow the inverse power law model. Therefore, its applicability must usually be verified experimentally before use. Arrhenius Model The Arrhenius relationship for reaction rate is often used to account for the effect of temperature on electrical/electronic components. The Arrhenius relationship is as follows:  − Ea  Reaction rate = A exp    K BT 

where: A = normalizing constant; KB = Boltzman’s constant (8.63 × 10–5 ev/degrees K); T = ambient temperature in degrees Kelvin; and Ea = activation energy type constant (unique for each failure mechanism). In those situations where it can be shown that the failure mechanism rate follows the Arrhenius rate with temperature, the following Acceleration Factor (AF) can be developed:

SL3151Ch07Frame Page 309 Thursday, September 12, 2002 6:07 PM

Reliability

309

 − Ea  Rateuse = A exp    K BTuse    − Ea Rateaccelerated = A exp    K BTaccelerated 

 − Ea  A exp    K BTa  Acceleration Factor = AF = Ratea/Rateu =  − Ea  A exp    K BTu   −E  1 E  1 1  1  AF = exp  a  −   = exp  a  −    K B  Ta Tu    K B  Tu Ta   where Ta = acceleration test temperature in degrees Kelvin and Tu = actual use temperature in degrees Kelvin. EXAMPLE Assume we have a device that has an activation energy of 0.5 and a characteristic life of 2750 hours at an accelerated operating temperature of 150°C. We want to find the characteristic life at an expected use temperature of 85°C. (Remember that the conversion factor for Celsius to Kelvin is: °K = °C + 273 — You may want to review Volume II.) Therefore:

Ta = 150 + 273 = 423°K and Tu = 85 + 273 = 358°K The Ea = 0.5. Our calculations would look like:

  1 .5 1  − AF = exp    −5  8.63x10  358 423   AF = exp [2.49] = 12. Therefore, the acceleration factor is 12. To determine life at 85°C, multiply the acceleration factor times the characteristic life at the accelerated test level of 150°C.

Characteristic life at 85°C = (12) (2750 hours) = 33,000 hours

SL3151Ch07Frame Page 310 Thursday, September 12, 2002 6:07 PM

310

Six Sigma and Beyond

AST/PASS HALT (Highly Accelerated Life Test) and HASS (Highly Accelerated Stress Screens) are two types of accelerated test processes used to simulate aging in manufactured products. The HALT/HASS process was invented by Dr. Gregg Hobbs in the early 1980s. It has since been used with much success in various military and commercial applications. The HALT/HASS methods and tools are still in the development phase and will continue to evolve as more companies embrace the concept of accelerated testing. Many companies use this type of testing, which they call AST (Accelerated Stress Test) and PASS (Production Accelerated Stress Screen). The goal of accelerated testing is to simulate aging. If the stress-strength relationships are plotted, the design strength and field stress are distributed around means. Let us assume the stress and strength distributions are overlapped (the right tail of the stress curve is overlapped with the left tail of the strength curve). When that happens, there is an opportunity for the product to fail in the field. This area of overlap is called interference. Many products, including some electronic products, have a tendency to grow weaker with age. This is reflected in a greater overlap of the curves, thus increasing the interference area. Accelerated testing attempts to simulate the aging process so that the limits of design strength are identified quickly and the necessary design modifications can be implemented.

PURPOSE

OF

AST

AST is a highly accelerated test designed to fail the target component or module. The goal of this process is to cause failure, discover the root cause, fix it, and retest it. This process continues until the “limit of technology” is reached and all the components of one technology (i.e., capacitors, diodes, resistors) fail. Once a design reaches its limit of technology, the tails of the stress-strength distribution should have minimal overlap. The AST method uses step-stress techniques to discover the operating and destruct limits of the component or module design. This method should be used in the pre-prototype and/or pre-bookshelf phase of the product development cycle or as soon as the first parts are available. Let us look at an example: We want to discover the operating and destruct limits of a component/module design for minimum temperature. The unit is placed in a test chamber, stabilized at –40°C, then powered up to verify the operation. The unit is then unpowered, the temperature lowered to –45°C and the unit allowed to stabilize at that temperature. It is then powered on and verified. This process is repeated as the temperature is lowered by 5° increments. At –70°C, the unit fails. The unit is warmed to –65°C to see if it recovers. Normally, it will recover. The temperature of –65°C is said to be its operational limit. The test continues to determine the destruct limit. The limit is lowered to –75°C, stabilized, powered to see if it operates, then returned it to –65°C to see if it recovers. If when this unit is taken down to –95°C and returned to –65°C, it does not recover, the minimum temperature destruct limit for this module is determined

SL3151Ch07Frame Page 311 Thursday, September 12, 2002 6:07 PM

Reliability

311

to be –95°C. The failed module is then analyzed to determine the root cause of the failure. The team must then determine if the failure mode is the limit of technology or if it is a design problem that can be fixed. Experience has shown that 80% of the failures are design problems accelerated to failure using the AST or similar accelerated stress test methods.

AST PRE-TEST REQUIREMENTS Before AST is run on a product, the product development team should verify that: • The component/module meets specification requirements at minimum and maximum temperature. • The vibration evaluation test (sine-sweep) is complete. • Data are available for review by the reliability engineer. • A copy of all schematics is available for review. The product development team will provide the component/module monitoring equipment used during AST and will work with the reliability engineer to define what constitutes a “failure” during the test.

OBJECTIVE

AND

BENEFITS

OF

AST

The objective of AST is to discover the operational and destruct limits of a design and to verify how close these limits are to the technological limits of the components and materials used in the design. It also verifies that the component/module is strong enough to meet the requirements of the customer and product application. These requirements must be balanced with reasonable cost considerations. The benefits of AST include: • Easier system and subsystem validation due to: • Elimination of component- /module-related failures • Verification of worst-case stress analysis and derating requirements • A list of failure modes and corrections to be shared with the design team and incorporated into future designs • Products that allow the manufacturing team to use PASS and to eliminate the in-process “build and check” types of tests The failure modes from the AST and PASS are used by the manufacturing team to ensure that they do not see any of these problems in their products.

PURPOSE

OF

PASS

PASS is incorporated into a process after the design has been first subjected to AST. The purpose of PASS is to take the process flaws created in the component/module from latent (invisible) to patent (visible). This is accomplished by severely stressing a component enough to make the flaws “visible” to the monitoring equipment. These

SL3151Ch07Frame Page 312 Thursday, September 12, 2002 6:07 PM

312

Six Sigma and Beyond

flaws are called outliers, and they result from process variation, process changes, and different supplier sources. The goal of PASS is to find the outliers, which will assist in the determination of the root cause and the correction of the problem before the component reaches the customer. This process offers the opportunity for the organization to eliminate module conditioning and burn-in. PASS development is an iterative process that starts when the pre-pilot units become available in the pre-pilot phase of the product development cycle. The initial PASS screening test limits are the AST operational limits and will be adjusted accordingly as the components/modules fail and the root cause determinations indicate whether the failures are limits of technology or process problems. The PASS also incorporates findings from process failure mode and effect analysis (PFMEA) regarding possible “significant” process failure modes that must be detected if present. When PASS development is complete, a strength-of-PASS test is performed to verify that the PASS has not removed too much useful life from the product. A sample of 12 to 24 components is run through 10 to 20 PASS cycles. These samples are then tested using the design verification life test. If the samples fail this test, the screen is too strong. The PASS will be adjusted based on the root cause analysis, and the strength-of-PASS will be rerun.

OBJECTIVE

AND

BENEFITS

OF

PASS

The objective of PASS is to precipitate all manufacturing defects in the component/module at the manufacturing facility, while still leaving the product with substantially more fatigue life after screening than is required for survival in the normal customer environment. The benefits of PASS include: • Accelerated manufacturing screens • Reduced facility requirements • Improved rate of return on tester costs

CHARACTERISTICS OF A RELIABILITY DEMONSTRATION TEST Eight characteristics are important in reliability demonstration testing. These are: 1. Specified reliability, Rs: This value is sometimes known as the “customer reliability.” Traditionally, this value is represented as the probability of success (i.e., 0.98); however, other measures may be used, such as a specified MTBF. 2. Confidence level of the demonstration test: While customers desire a certain reliability, they want the demonstration test to prove the reliability at a given confidence level. A demonstration test with a 90% confidence level is said to “demonstrate with 90% confidence that the specified reliability requirement is achieved.”

SL3151Ch07Frame Page 313 Thursday, September 12, 2002 6:07 PM

Reliability

313

3. Consumer’s risk, β: Any demonstration test runs the risk of accepting bad product or rejecting good product. From the consumer’s point of view, the risk is greatest if bad product is accepted. Therefore, the consumer wants to minimize that risk. The consumer’s risk is the risk that a test can accept a product that actually fails to meet the reliability requirement. Consumer’s risk can be expressed as: β = 1 – confidence level 4. Probability distribution: This is the distribution that is used for the number of failures or for time to failure. These are generally expressed as normal, exponential, or Weibull. 5. Sampling scheme 6. Number of test failures to allow 7. Producer’s risk, α : From the producer’s standpoint, the risk is greatest if the test rejects good product. Producer’s risk is the risk that the test will reject a product that actually meets the reliability requirement. 8. Design reliability, Ra: This is the reliability that is required in order to meet the producer’s risk, α, requirement at the particular sample size chosen for the test. Small test sample sizes will require a high design reliability in order to meet the producer’s risk objective. As the sample size increases, the design reliability requirement will become smaller in order to meet the producer’s risk objective.

THE OPERATING CHARACTERISTIC CURVE The relationship between the probability of acceptance and the population reliability can be shown with an operating characteristic (OC) curve. An operating characteristic curve can also be used to show the relationship between the probability of acceptance and MTBF or failure rate. Given then an OC curve, one may calculate the: 1. Producer’s risk, α 2. Consumer’s risk, β 3. Probability of acceptance at any other population reliability or MTBF or failure rate Obviously, a specific OC curve will apply for each test situation and will depend on the number of pieces tested and the number of failures allowed.

ATTRIBUTES TESTS If the components being tested are merely being classified as acceptable or unacceptable, the demonstration test is an attributes test. Attributes tests: • May be performed even if a probability distribution of the time to failure is not known • May be performed if a probability distribution such as normal, exponential, or Weibull is assumed by dichotomizing the life distribution into acceptable and unacceptable time to failure

SL3151Ch07Frame Page 314 Thursday, September 12, 2002 6:07 PM

314

Six Sigma and Beyond

• Are usually simpler and cheaper to perform than variables tests • Usually require larger sample sizes to achieve the same confidence or risks as variables tests

VARIABLES TESTS Variables tests are used when more information is required than whether the unit passed or failed, for example, “What was the time to failure?” The test is a variables test if the life of the items under test is: • Recorded in time units • Assumed to have a specific probability distribution such as normal, exponential, or Weibull

FIXED-SAMPLE TESTS When the required reliability and the test confidence/risk are known, statistical theory will dictate the precise number of items that must be tested if a fixed sample size is desired.

SEQUENTIAL TESTS A sequential test may be used when the units are tested one at a time and the conclusion to accept or reject is reached after an indeterminate number of observations. In a sequential test: 1. The “average” number of samples required to reach a conclusion will usually be lower than in a fixed-sample test. This is especially true if the population reliability is very good or very poor. 2. The required sample size is unknown at the beginning of the test and can be substantially larger than that in the fixed-sample test in certain cases. 3. The test time required is much longer because samples are tested one at a time (in series) rather than all at the same time (in parallel), as in fixedsample tests. Now that you are familiar with the four test types, let us look at the test methods. Note that the four test types are not mutually exclusive. We can have fixed-sample or sequential-attributes tests as well as fixed-sample or sequential-variables tests.

RELIABILITY DEMONSTRATION TEST METHODS Attributes tests can be used when: • The accept/reject criterion is a go/no-go situation. • The probability distribution of times to failure is unknown. • Variables tests are found to be too expensive.

SL3151Ch07Frame Page 315 Thursday, September 12, 2002 6:07 PM

Reliability

315

SMALL POPULATIONS — FIXED-SAMPLE TEST USING THE HYPERGEOMETRIC DISTRIBUTION When items from a small population are tested and the accept/reject decision is based on attributes, the hypergeometric distribution is applicable for test planning. The definition of successfully passing the test will be that an item survives the test. The parameter to be evaluated is the population reliability. The estimation of the parameter is based on a fixed sample size and testing without repair. The method to use is described below: 1. 2. 3. 4.

Define the criteria for success/failure. Define the acceptance reliability, RS. Specify the confidence level or the corresponding consumer’s risk, β. Specify, if desired, producer’s risk, α. (Producer’s risk can be used to calculate the design reliability target, Rd, needed in order to meet the α requirements.)

The process consists of a trial-and-error solution of the hypergeometric equation until the conditions for the probability of acceptance are met. The equation that is used is: f

Pr(x ≤ f) =

∑ x =0

 N (1 − R)  NR   x   n − x    N  n  

where Pr(x < –f) = probability of acceptance; f = maximum number of failures to be allowed; x = observed failures in sample; R = reliability of population; N = population size; and n = sample size.  N N!  n = n N −n! !  

(

)

LARGE POPULATION — FIXED-SAMPLE TEST USING THE BINOMIAL DISTRIBUTION When parts from a large population are tested and the accept/reject decision is based on attributes, the binomial distribution can be used. Note that for a large N (one in which the sample size will be less than 10% of the population), the binomial distribution is a good approximation for the hypergeometric distribution. The binomial attribute demonstration test is probably the most versatile for use on product components for several reasons: 1. The population is large. 2. The time-to-failure distribution for the parts is probably unknown. 3. Pass/fail criteria are usually appropriate.

SL3151Ch07Frame Page 316 Thursday, September 12, 2002 6:07 PM

316

Six Sigma and Beyond

As with the hypergeometric distribution, the procedure begins by identification of: 1. Specified reliability, Rs 2. Confidence level or consumer’s risk, β 3. Producer’s risk, α (if desired) The process consists of a trial-and-error solution of the binomial equation until the conditions for the probability of acceptance are met. The equation that is used is: f

Pr(x ≤ f) =

 n

∑  x (1 − R) ( R) x

n− x

x =0

where Pr(x < f) = probability of acceptance; f = maximum number of failures to be allowed; x = observed failures in sample; R = reliability of population; and n = sample size.

LARGE POPULATION — FIXED-SAMPLE TEST USING THE POISSON DISTRIBUTION The Poisson distribution can be used as an approximation of both the hypergeometric and the binomial distributions if: The population, N, is large compared to the sample size, n. The fractional defective in the population is small (Rpopulation > 0–9). The process consists of a trial-and-error solution using the following equation or Poisson tables, Rs, Rd, and various sample sizes until the conditions of α and β are satisfied. f

Pr(x ≤ f) =

∑ x =0

λxpoie

− λ poi

x!

where Pr(x ≤ f) = probability of acceptance; f = maximum number of failures to be allowed; x = observed failures in sample; λpoi = (n) (1 – R) (The reader should note that the λpoi is the Poisson density and does not relate to failure rate); r = reliability of population; and n = sample size.

SUCCESS TESTING Success testing is a special case of binomial attributes testing for large populations where no failures are allowed. Success testing is the simplest method for demonstrating a required reliability level at a specified confidence level. In this test case, n items are subjected to a test for the specified time of interest, and the specified

SL3151Ch07Frame Page 317 Thursday, September 12, 2002 6:07 PM

Reliability

317

reliability and confidence levels are demonstrated if no failures occur. The method uses the following relationship: R = (1 – C)1/n = (β)1/n where R = reliability required; n = number of units tested; C = confidence level; and β = consumer’s risk. The necessary sample size to demonstrate the required reliability at a given confidence level is: n=

SEQUENTIAL TEST PLAN

FOR THE

ln(1 − C ) ln R

BINOMIAL DISTRIBUTION

The sequential test is a hypothesis testing method in which a decision is made after each sample is tested. When sufficient information is gathered, the testing is discontinued. In this type of testing, sample size is not fixed in advance but depends upon the observations. Sequential tests should not be used when the exact time or cost of the test must be known beforehand or is specified. This type of test plan may be useful when the: 1. Accept/reject criterion for the parts on test is based on attributes 2. Exact test time available and sample size to be used are not known or specified The test procedure consists of testing parts one at a time and classifying the tested parts as good or defective. After each part is tested, calculations are made based on the test data generated to that point. The decision is made as to whether the test has been passed or failed or if another observation should be made. A sequential test will result in a smaller average number of parts tested when the population tested has a reliability close to either the specified or design reliability. The method to use is described below: Determine Rs, Rd, α.β Calculate the accept/reject decision points using: β 1−β and 1−α α As each part is tested, classify it as either a failure or success. Evaluate the following expression for the binomial distribution, f

 1 − Rs   Rs  L =     1 − Rd   Rd 

s

SL3151Ch07Frame Page 318 Thursday, September 12, 2002 6:07 PM

318

Six Sigma and Beyond

where F = total number of failures and S = total number of successes.

If

If L >

1−β , the test is failed. α

If L


1−β , the test is failed. α

If L
0); and θ = characteristic life (θ ≥ δ). For a given reliability

1

 1 β t = θ + (θ – δ) × ln( )  R  and the B10 life is

1

 1 β B10 = θ + (θ – δ) × ln( )  0.90 

DESIGN OF EXPERIMENTS IN RELIABILITY APPLICATIONS Certainly we can use DOE in passive observation of the covariates in the tested components. We can also use DOE in directed experimentation as part of our reliability improvement. Covariates are usually called factors in the experimentation framework. Two main technical problems arise in the reliability area, however, when standard methods of experimental design are employed. 1. Failure time data are rarely normally distributed, so standard analysis tools that rely on symmetry, e.g., normal plots, do not work too well. 2. Censoring. The first problem can be overcome by considering a transformation of the fail times to make them approximately normal — the log transformation is usually a good choice. The exact form of the fail time distribution is not important because we are looking for effects that improve reliability, rather than exact predictions of the reliability itself. The second problem of censoring is a little bit trickier but can be dealt with by iteration as follows: 1. Choose a basic model to fit to the data. 2. Fit the model to the data, treating the censor times as failure times. 3. Using this model, make a conditional prediction for the unobserved fail times for each censored observation. The prediction is conditional because the actual failure time must be consistent with the censoring mechanism. 4. Replace censor times with the fail time predictions from step 3. 5. Go back to step 2.

SL3151Ch07Frame Page 336 Thursday, September 12, 2002 6:07 PM

336

Six Sigma and Beyond

Eventually this process will converge, i.e., the predictions for the fail times of the censorings will stop changing from one iteration to the next. If necessary, the process can be tried with several model choices for step 1. In fact, the algorithm of the five steps leads to the same results as maximum likelihood estimation.

RELIABILITY IMPROVEMENT THROUGH PARAMETER DESIGN Two special categories of covariates in any parameter design are design parameters (or control factors) and error variables (or noise factors). The terms in parenthesis are the equivalent terms within the context of robustness, which we already have discussed in Volume V of this series. The achievement of higher reliability can also be viewed as an improvement to robustness. Robustness is defined as reduced sensitivity to noise factors. In most industries, noise factors have five main categories: 1. 2. 3. 4. 5.

Piece to piece variation Changes to component characteristics over time Customer duty cycle Environmental conditions Interfacing (environment created by neighboring components in the system)

Typically, noises in categories 3, 4, and 5 can induce noises in category 2. If the function of the component can be made robust to noises in category 2, then the component will, by definition, be more reliable. Often, noise category 1 contributes to infant mortality, category 2 to degradation, and categories 3, 4, and 5 to useful life problems. Recognizing this pattern of noises, we can relate them to the bathtub curve (see Figure 7.1) for the hazard function. Often, knowing the type of failure rate that is acting on our component can give a clue as to the offending noise factor and hence lead to a root cause analysis of the failure mechanism. Components can be made robust to noises by experimenting with control factors. The idea (as in robustness generally) is to look for interactions between control and noise factors. The reliability connection is made if there is a “time lag” between the extremes of the noise space, denoted N– and N+, say — see Figure 7.6. Note that the functional measure is not failure time, but some ideal function of the system. C1 and C2 represent two settings of a control factor. A design with C2 is more robust to noise than one with C1 and is therefore more reliable. Note: A closely related area to robustness in reliability studies is Accelerated Degradation Testing (ADT), which is closely associated with Accelerated Life Testing (ALT). A parameter design layout in reliability applications follows the pattern for parameter design studies, as in the example shown in Figure 7.7.

SL3151Ch07Frame Page 337 Thursday, September 12, 2002 6:07 PM

Reliability

337

Functional measure C2

C1 N-

N+ Time

FIGURE 7.6 Control factors and noise interactions.

Control Factors Configuration

A

B

C ...

G

2 3 4 5 6 7 8

+ + + +

+ + + +

+ + + +

+ + + +

Noise Factors time N(new) ylY2Y3Y4YsY6Y7Y8-

N+ (old) Y1+ Y2+ Y3+ Y4+ Ys+ Y6+ Y7+ Y8+

FIGURE 7.7 An example of a parameter design in reliability usage.

The idea of experimental layouts of this type is to look for interactions between control factors and noise factors, which lead to configurations with minimum difference between the y values.

DEPARTMENT OF DEFENSE RELIABILITY AND MAINTAINABILITY — STANDARDS AND DATA ITEMS Table 7.5 provides very useful information about reliability and maintainability (R&M) standards and data items used in reliability.

SL3151Ch07Frame Page 338 Thursday, September 12, 2002 6:07 PM

338

Six Sigma and Beyond

TABLE 7.5 Department of Defense Reliability and Maintainability — Standards and Data Items Standard General Design Standards MIL-STD-454M MIL-HDBK-727 MIL-STD-810E MIL-STD-1629A MIL-STD-1686A MIL-E-4158E-(USAF) MIL-E-5400T MIL-HDBK-251 MIL-HDBK-263A MIL-HDBK-338A Reliability Standards MIL-STD-721C MIL-STD-756B MIL-STD-781 D

Explanation

Standard General Requirements for Electronic Equipment Design Guidance for Producibility Environmental Test Methods & Engineering Guidelines Procedures for Performing a Failure Mode Effects & Criticality Analysis Electrostatic Discharge Control Program for Protection of Electrical & Electronic Parts, Assemblies & Equipment General Specification for Ground Electronic Equipment General Specification for Aerospace Electronic Equipment Reliability/Design Thermal Applications Electrostatic Discharge Handbook for Protection of Electrical & Electronic Parts, Assemblies & Equipment Electronic Reliability Design Handbook

DoD-HDBK-344-(USAF)

Definitions of Terms for Reliability & Maintainability Reliability Modeling & Prediction Reliability Testing for Engineering Development Qualification & Production Reliability Program Systems & Equipment Development & Production Reliability Program Requirements for Space & Launch Vehicles Failure Reporting Analysis & Corrective Action System Environmental Stress Screening Process for Electronic Equipment Quality Program Requirements Reliability Growth Management Reliability Prediction of Electronic Equipment Reliability Test Methods, Plans & Environments for Engineering Development, Qualification & Production Environmental Stress Screening of Electronic Equipment

Maintainability Standards MIL-STD-470B MIL-STD-471A MIL-STD-2084-(AS) MIL-STD-2165 MIL-HDBK-472

Maintainability Program for Systems & Equipment Maintainability Demonstration General Requirements for Maintainability Testability Program for Electronic Systems & Equipment Maintainability Prediction

Major Parts Standards MIL-STD-198E MIL-STD-199E MIL-STD-202E MIL-STD-701N MIL-STD-750C

Selection & Use of Capacitors Selection & Use of Resistors Test Methods for Electronic & Electrical Component Parts Lists of Standard Semiconductor Devices Test Methods for Semiconductor Devices

MIL-STD-785B MIL-STD-1543B-(USAF) MIL-STD-2155-(AS) MIL-STD-2164-(EC) MIL-0–9858A MIL-HDBK-189 MIL-HDBK-217F MIL-HDBK-781

SL3151Ch07Frame Page 339 Thursday, September 12, 2002 6:07 PM

Reliability

339

TABLE 7.5 (continued) Department of Defense Reliability and Maintainability — Standards and Data Items MIL-STD-790E MIL-STD-883D MIL-STD-965A MIL-STD-983 MIL-STD-1546A-(USAF) MIL-STD-1547A-(USAF) MIL-STD-1556B MIL-STD-1562W MIL-STD-1772B MIL-S-19500H + OPL MIL-M-38510J + QPL MIL-H-38534A + QML MIL-1–38535A + QML MIL-HDBK-339-(USAF) MIL-HDBK-780A MIL-BUL-103J

Reliability Assurance Program for Electronic Part Specifications Test Methods & Procedures for Microelectronics Parts Control Program Substitution List for Microcircuits Parts, Materials & Processes Control Program for Space & Launch Vehicles Electronic Parts, Materials & Processes for Space & Launch Vehicles Government/Industry Data Exchange Program (GIDEP) Contractor Participation Requirements Lists of Standard Microcircuits Certification Requirements for Hybrid Microcircuit Facility & Lines General Specification for Semiconductor Devices General Specification for Microcircuits General Specification for Hybrid Microcircuits General Specification for Integrated Circuits (Microcircuits) Manufacturing Custom LSI Circuit Development & Acquisition for Space Vehicles Standardized Military Drawings List of Standardized Military Drawings (SMDs)

Reliability Analysis Center Publications DSR Discrete Semiconductor Device Reliability FMD Failure Mode/Mechanism Distributions FTA Fault Tree Analysis MFAT-1 Microelectronics Failure Analysis Techniques — A Procedural Guide MFAT-2 GaAs Characterization & Failure Analysis Techniques NONOP-1 Nonoperating Reliability Data NPRD Nonelectronic Parts Reliability Data NPS-1 Analysis Techniques for Mechanical Reliability PRIM A Primer for DoD Reliability, Maintainability, Safety and Logistics Standards RDSC-1 Reliability Sourcebook RMST Reliability and Maintainability Software Tools SOAR-2 Practical Statistical Analysis for the Reliability Engineer SOAR-4 Confidence Bounds for System Reliability SOAR-6 ESD Control in the Manufacturing Environment SOAR-7 A Guide for Implementing Total Quality Management SOAR-8 Process Action Team (PAT) Handbook VZAP Electrostatic Discharge Susceptibility Data Computer Formats NPRD-P NRPS VZAP-P

Nonelectronic Parts Reliability Data (IBM PC database) Nonoperating Reliability Prediction Software (Includes NONOP-1) VZAP Data (IBM PC database)

SL3151Ch07Frame Page 340 Thursday, September 12, 2002 6:07 PM

340

Six Sigma and Beyond

TABLE 7.5 (continued) Department of Defense Reliability and Maintainability — Standards and Data Items Rome Laboratory Technical Reports Rome Laboratory (formerly Rome Air Development Center) has published hundreds of useful R&M technical reports that are available from the Defense Technical Information Center and the National Technical Information Service. Call RAC for a list. [Address at publication time: Reliability Analysis Center * 201 Mill Street * Rome, NY. 13440–6916 * Telephone: 315.337.0900] Data Item Descriptions MIL-STD-756 Reliability Modeling and Prediction DI-R-7081 B Mathematical Model(s) B Predictions Report(s) DI-R-7082 B Block Diagrams & Math. Models Report DI-R-7094 B Predict. & Doc. of Support. Material DI-R-7095 B Report for Explor. Advanced Develop. DI-R-7100 MIL-STD-781 Reliability Test Methods, Plans, and Environments for engineering development, Qualification and Production DI-RELI-80247 Thermal Survey Report DI-RELI-80248 Vibration Survey Report DI-RELI-80249 ESS Report DI-RELI-80250 B Test Plan DI-RELI-80251 B Test Procedures DI-RELI-80252 B Test Report DI-RELI-80253 Failed Item Analysis Report DI-RELI-80254 Corrective Action Plan DI-RELI-80255 Failure Summary and Analysis Report MIL-STD-785 Reliability Program for Systems and Equipment Development and Production and MIL-STD-1543 Reliability Program Requirements for Space and Launch Vehicles DI-R-7079 R Program Plan DI-R-7084 Elect. Parts/Circuits Tol. Analysis Report DI-R-7086 FMECA Plan DI-A-7088 Conference Agenda DI-A-7089 Conference Agenda DI-OCIC-80125 ALERT/SAFE ALERT DI-OCIC-80126 Response to ALERT/SAFE ALERT DI-RELI-80249 ESS Report DI-RELI-80250 Test Plan DI-RELI-80251 Test and Demo. Procedures DI-RELI-80252 Test Reports DI-RELI-80253 Failed Item Analysis Report DI-RELI-80255 Report, Failure Summary and Analysis DI-RELI-80685 Critical Item List DI-RELI-80686 Allocat., Assess. & Analysis Report DI-RELI-80687 Report, FMECA

SL3151Ch07Frame Page 341 Thursday, September 12, 2002 6:07 PM

Reliability

341

TABLE 7.5 (continued) Department of Defense Reliability and Maintainability — Standards and Data Items MIL-STD-2155 FRACA System DI-E-2178 Computer Software Trouble Report DI-R-21597 FRACA System Plan DI-R-21598 Failure Report DI-R-21599 Develop. & Product. Failure Summary Report MIL-STD-2164 ESS Process for Electronic Equipment DI-ENVR-80249 Environmental Stress Screening Report DOD-HDBK-344 ESS of Electronic Equipment DI-ENVR-80249 Environmental Stress Screening Report MIL-STD-810 Environmental Test Methods and Engineering Guidelines DI-ENVR-80859 Environmental Management Plan DI-ENVR-80860 Life Cycle Environmental Profile DI-ENVR-80861 Environmental Design Test Plan DI-ENVR-80862 Operational Environment Verif. Plan DI-ENVR-80863 Environmental Test Report MIL-STD-1629 Procedures for Performing a FMECA DI-R-7085 FMECA Report DI-R-7086 FMECA Plan MIL-STD-1686 ESD Control Program for Protection of Electrical and Electronic Parts, Assemblies and Equipment DI-RELI-80669 ESD Control Program Plan DI-RELI-80670 Reporting Results of ESD Sensitivity Tests of Electrical & Electronic Parts DI-RELI-80671 Handling Procedure for ESD Sensitive Items MIL-STD-1546 Parts, Materials, and Processes Control Program for Space and Launch Vehicles DI-A-7088 Conference Agenda DI-A-7089 Conference Minutes DI-MI SC-80526 Parts Control Program Plan DI-MISC-80072 Program Parts Selection List (PPSL) DI-MISC-80071 Part Approval Requests MIL-STD-1556 GIDEP Contractor Participation Requirements DI-QCIC-80125 ALERT/SAFE-ALERT DI-QCIC-80126 Response to an ALERT/SAFE-ALERT DI-QCIC-80127 GIDEP Annual Progress Report

SL3151Ch07Frame Page 342 Thursday, September 12, 2002 6:07 PM

342

Six Sigma and Beyond

TABLE 7.5 (continued) Department of Defense Reliability and Maintainability — Standards and Data Items MIL-STD-470 Maintainability Program for Systems and Equipment DI-R-2129 M Demo. Plan (MIL-STD-470A, Task 301 only) DI-R-7085 FMECA Report DI-MNTY-80822 Program Plan DI-MNTY-80823 M Status Report DI-MNTY-80824 Data Collect., Anal. & Correct. Action System DI-MNTY-80825 M Modeling Report DI-MNTY-80826 M Allocations Report M Predictions Report DI-MNTY-80827 M Analysis Report DI-MNTY-80828 M Design Criteria Plan DI-MNTY-80829 Inputs to the Detailed Maintenance Plan & LSA DI-MNTY-80830 M Testability Demo. Test Plan DI-MNTY-80831 M Testability Demo. Test Report DI-MNTY-80832 MIL-STD-471 Maintainability Demonstration DI-R-2129 M Demonstration Plan DI-MNTY-80831 M Testability Demo. Test Plan DI-MNTY-80832 M Testability Demonstration Report DI-MNTY-81188 Verif., Demo., Assess. & Evaluation Plan DI- QCIC-81187 Quality Assessment Report MIL-STD-2165 Testability Program for Electronic Systems and Equipments DI-E-5423 Design Review Data Package DI-T-7198 Testability Program Plan DI-T-7199 Testability Analysis Report DI-MNTY-80824 Data Collect., Anal. & Correct. Act. System Plan DI-MNTY-80831 M/Testability Demo. Test Plan DI-MNTY-80832 M/Testability Demo. Report MIL-HDBK-472 Maintainability Prediction DI-MNTY-80827 M Predictions Report Note: Only data items specified in the Contract Data Requirements List (CDRL) are deliverable.

REFERENCES Anon., Warranty Cost Issue Hurts Chrysler, USA Today, Oct. 24, 1994, p. 3B. ANSI/IEEE Standard 100–1988, 4th ed., IEEE Standard Dictionary of Electrical and Electronic Terms, The Institute of Electrical and Electronic Engineers, Inc., New York, 1988. Flint, J., It Is Time To Get Realistic, WARD’S AUTOWORLD, Oct. 2001, p. 21. Mayne, E. et al., Quality Crunch, Ward’s AUTOWORLD, July 2001, pp. 14–18. VonAlven, W.H., Ed., Reliability Engineering, Prentice Hall, Inc., Englewood Cliffs, NJ, 1964.

SL3151Ch07Frame Page 343 Thursday, September 12, 2002 6:07 PM

Reliability

343

SELECTED BIBLIOGRAPHY Aitken, M., A note on the regression analysis of censored data, Technometrics, 23, 161–163, 1981. Box, G.E.P. and Meyer, R.D., Finding the active factors in fractionated screening experiments, Journal of Quality Technology, 25, 94–105, 1993. Cox, D.R. and Oakes, D., Analysis of Survival Data, Chapman Hall, London, 1984. Grove, D.M. and Davis, T.P., Engineering, Quality, and Experimental Design, Longman, Harlow, England, 1992. Hamada, M. and Wu, C.F.J., Analysis of censored data from highly fractionated experiments. Technometrics, 33, 25–3, 1991. Hamada, M. and Wu, C.F.J., Analysis of designed experiments with complex aliasing, Journal of Quality Technology, 23, 130–137, 1992. Kalbfleisch, J.D. and Prentice, R.L., The Statistical Analysis of Failure Time Data, Wiley, New York, 1980. Kapur, K.C. and Lamberson, L.R., Reliability in Engineering Design, Wiley, New York, 1977. Kececioglu, D., Reliability Engineering Handbook, Vols. 1 and 2, Prentice Hall, Englewood Cliffs, NJ, 1991. Lawless, J. F., Statistical Models and Methods for Lifetime Data, Wiley, New York, 1982. McCormick, N.J., Reliability and Risk Analysis, Academic Press, New York, 1981. Nelson, W., Theory and applications of hazard plotting for censored failure data, Technometrics, 14, 945–966, 1972. Schmee, J. and Hahn, G., A simple method of regression analysis with censored data. Technometrics, 21, 417–432, 1979. Smith, R.L., Weibull regression models for reliability data, Reliability Engineering and System Safety, 34, 55–57, 1991.

SL3151Ch07Frame Page 344 Thursday, September 12, 2002 6:07 PM

SL3151Ch08Frame Page 345 Thursday, September 12, 2002 6:07 PM

8

Reliability and Maintainability

As the world moves towards building more competitive products, it is important to put additional emphasis on reliability and maintainability (R&M), which support reduction of inventories and “build to schedule” targets. The Quality Systems Requirements, Tooling & Equipment (TE) Supplement to QS-9000 was developed by Chrysler, Ford, General Motors, and Riviera Die & Tool to enhance quality systems while eliminating redundant requirements, facilitating consistent terminology, and reducing costs. It is important that everyone involved in the design or purchase of machinery be aware of this supplement and their responsibilities as outlined in the QS-9000 process. It is also important that everyone understand that the TE supplement defines machinery as tooling and equipment combined. Machinery is a generic term for all hardware, including necessary operational software, which performs a manufacturing process. The TE goal is to improve the quality, reliability, maintainability, and durability of products through development and implementation of a fundamental quality management system. The supplement communicates additional common system requirements unique to the manufacturers of tooling and equipment as applied to the QS-9000 requirements. This particular chapter will emphasize the reliability and maintainability areas. Quality operating systems (QOS) and durability are equally important subjects but are beyond the scope of this work. The reader is encouraged to review Volume IV — the material on machine acceptance.

WHY DO RELIABILITY AND MAINTAINABILITY? Due to a lack of confidence in the performance of our equipment, we have traditionally purchased excessive facilities and tooling in order to meet production objectives. It is estimated that approximately 73% of the total cost in a program development through launching, in the automotive industry for example, is in this area. Additionally, capital spent on “insurance-type” spare tooling hidden for unplanned breakdowns shows a lack of confidence in production equipment. Operational effects of production shortfall and the inability to predict downtime are countless. They include unplanned overtime, unplanned and increasing maintenance requirements and costs, and excessive work in process around constraint operations. The R&M process builds confidence in predicting performance of machinery, and, through this process, we can improve the expected and demonstrated levels of machinery performance. Properly predicting and improving performance contributes to lower total cost and improved profits for the organization.

345

SL3151Ch08Frame Page 346 Thursday, September 12, 2002 6:07 PM

346

Six Sigma and Beyond

The R&M process consists of five phases that form a continuous loop. The five phases are: (1) concept; (2) design and development; (3) machinery build and installation; (4) machinery operation, continuous improvement, performance analysis; and (5) conversion concept of next cycle. As the loop continues, each generation of machinery improves. In this chapter we will concentrate on the first three phases of the loop, not because they are more important, but because they are the major focus of this planning effort of the design for six sigma (DFSS) campaign. The last two phases should be well documented in each organization for they are facility dependent.

OBJECTIVES The emphasis of all R & M is focused on three objectives: Reliability — The probability that machinery and equipment can perform continuously, without failure, for a specified interval of time (when operating under stated conditions) Maintainability — A characteristic of design, installation, and operation, usually expressed as the probability that a machine can be retained in, or restored to, specified operable conditions within a specified interval of time (when maintenance is performed in accordance with prescribed procedures) Durability — Ability to perform intended function over a specified period (under normal use with specified maintenance) without significant deterioration

MAKING RELIABILITY AND MAINTAINABILITY WORK Machinery reliability and maintainability should be considered an integral part of all facilities and tooling (F&T) purchases. However, the appropriate degree of time and effort dedicated to R&M engineering must be individually applied for each unique application and purchase situation. Each project engineering manager should consider the value proposition of applying varying degrees of R&M engineering for the unique circumstances surrounding each equipment purchase. For example, we may choose to apply a large amount of R&M engineering resources to a project that includes a large quantity of single design machines. The value proposition would show that investing up-front resources on a single design that can be leveraged beyond a single application would offer a large payoff. We would also consider applying high-level R&M engineering to equipment critical to a continuous operation. On the other hand, we may choose to apply a minimal level of R&M engineering on a purchase of equipment that has a mature design and minimally demonstrated field problems. Some of the issues to consider when determining appropriate levels of R&M engineering for a project include:

SL3151Ch08Frame Page 347 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

347

1. Review the availability of existing machines in the organization that may be idle. This is a good opportunity for reusability. 2. How many units are we ordering with identical or leverageable design? 3. What is the condition of the existing machinery that will be rehabilitated? 4. What is the status of the operating conditions? Are they extremely demanding? 5. What is the cycle plan for the machinery? Does it require continuous or intermittent duty? For how many years is the equipment expected to produce? 6. Where is the machinery in the manufacturing process? Is it a constraint (bottleneck) operation? 7. How well documented and complete is the root cause analysis for the design? Will it decrease up-front work? 8. How much data exist to support known design problems?

WHO’S RESPONSIBLE? Full realization of R&M benefits requires consistent application of the process. Simultaneous engineering (SE) teams, together with the plants and the supply base, must align their efforts and objectives to provide quality machinery designed for R&M. Reliability and maintainability engineering is the responsibility of everyone involved in machinery design, as much as the collection and maintenance of operational data are the responsibility of those operating and maintaining the equipment day to day. The R&M process places responsibility on the groups possessing the skills or knowledge necessary to efficiently and accurately complete a given set of tasks. It turns out that much of the expertise is in the supply base, and as such, the suppliers must take the lead role and responsibility in R&M efforts. The R&M process encourages the organization and suppliers to lock into budget costs based on Life Cycle Costing (LCC) analysis of options and cost targets. Warranty issues should be considered in the LCC analysis so that design helps decrease excessive warranty costs after installation. The focus places responsibility for correcting design defects on the machinery designers. Facility and tooling producers who practice R&M will ultimately reduce the cost (such as warranty) of their product and will become more competitive over time. Further, suppliers that practice R&M will qualify as QS-9000 certified, preferred, global sourcing partners. Engineers and program managers who practice and encourage R&M will reduce operational costs over time. In doing so, they will meet manufacturing and cost objectives for their projects or programs.

TOOLS There are many R&M tools. The ones mentioned here are required in the Design and Development Planning (4.4.2) section of the TE Supplement. Many others beyond the few that are addressed here are available and can improve reliability.

SL3151Ch08Frame Page 348 Thursday, September 12, 2002 6:07 PM

348

Six Sigma and Beyond

Mean Time Between Failure (MTBF) is defined as the average time between failure occurrences. It is simply the sum of the operating time of a machine divided by the total number of failures. For example, if a machine runs for 100 hours and breaks down four times, the MTBF is 100 divided by 4 or 25 hours. As changes are made to the machine or process, we can measure the success by comparing the new MTBF with the old MTBF and quantify the action that has been taken. Mean Time to Repair (MTTR) is defined as the average time to restore machinery or equipment to its specified conditions. This is accomplished by dividing the total repair time by the number of failures. It is important to note that the MTTR calculation is based on repairing one failure and one failure only. The length of time it takes to repair each failure directly affects up-time, up-time %, and capacity. For example, if a machine runs 100 hours and has eight failures recorded with a total repair time of four hours, the MTTR for this machine would be four hours divided by eight failures or .5 hours. This is the mean time it takes to repair each failure. Fault Tree Analysis (FTA) is an effect-and-cause diagram. It is a method used to identify the root causes of a failure mode using symbols developed in the defense industry. The FTA is a great prescriptive method for determining the root causes associated with failures and can be used as an alternative to the Ishikawa Fish Bone Diagram. It compliments the Machinery Failure Mode and Effects Analysis (MFMEA) by representing the relationship of each root cause to other failure-mode root causes. Some feel the FTA is better suited than the FMEA to providing an understanding of the layers and relationships of causes. An FTA also aids in establishing a troubleshooting guide for maintenance procedures. It is a top down approach. Life Cycle Costs (LCC) are the total costs of ownership of the equipment or machinery during its operational life. A purchased system must be supported during its total life cycle. The importance of life cycle costs related to R&M is based on the fact that up to 95% of the total life cycle costs are determined during the early stages of the design and development of the equipment. The first three phases of the equipment’s life cycle are typically identified as non-recurring costs. The remaining two phases are associated with the equipment’s support costs.

SEQUENCE AND TIMING The R&M process is a generic model of logically sequenced events that guides the simultaneous engineering team through the main drivers of good design for R&M engineering. The amount of time budgeted for each activity or task should vary depending on the circumstances surrounding the equipment or processes in design. However, regardless of the unique conditions, all of the steps in the R&M process need to be considered in their logical sequence and applied as needed. In Table 8.1, we identify different activities that you may consider in the first three phases of the R&M process. These phases are divided into main areas for consideration; then, various activities are listed for each area. This list is not complete, but it focuses the reader on the type of activities that should occur during each time period. This list also helps identify the sequence in which these activities may be completed, depending on the project.

SL3151Ch08Frame Page 349 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

349

TABLE 8.1 Activities in the First Three Phases of the R&M Process Concept

Design/Development

Build and Installation

Bookshelf data manufacturing process selection R&M and production needs analysis

R&M planning Process design for R&M machinery FMEA — design review

Equipment run-off Operation of machinery

To determine timing for the R&M process, you may use the following procedure: 1. Determine deadline dates to meet production requirements. 2. Check relevance of R&M activities with regard to achieving program/project targets. 3. Plan relevant R&M activities by working backwards from deadline dates, estimating time required for completion of each activity. 4. Set appropriate start dates for each activity/stage based on requirements and timing. 5. Determine and assign responsibility for stage-based deliverables. 6. Continually track progress of your plan, within and at the conclusion of each stage.

CONCEPT BOOKSHELF DATA Activities associated with the bookshelf data stage include: 1. 2. 3. 4. 5. 6. 7. 8. 9.

Identify good design practices. Collect machinery things gone right/things gone wrong (TGR/TGW). Document successful machinery R&M features. Collect similar machinery history of mean time between failures (MTBF). Collect similar standardized component history of mean time between failures (MTBF). Collect similar machinery history of mean time to repair (MTTR). Collect similar machinery history of overall equipment effectiveness (OEE). Collect similar machinery history of reliability growth. Collect similar machinery history of root cause analyses.

At this point it is important to ask and answer this question: Have we collected all of the relevant historical data from similar operations or designs and documented them for use during the process selection and design stages?

SL3151Ch08Frame Page 350 Thursday, September 12, 2002 6:07 PM

350

Six Sigma and Beyond

MANUFACTURING PROCESS SELECTION Activities associated with the manufacturing process selection stage include: 1. Identify general life cycle costs to drive the manufacturing process selection. 2. Establish OEE targets including availability, quality, and performance efficiency numbers that drive the manufacturing process selection. 3. Establish broad R&M target ranges that drive the manufacturing process selection. 4. Establish manufacturing assumptions based on cycle plan, including volumes and dollar targets. 5. Identify simultaneous engineering (SE) partners for project. 6. Select manufacturing process based on demonstrated performance and expected ability to meet established targets. 7. Search for other surplus equipment to be considered for reuse. 8. If surplus machinery has not been identified for reuse, identify a supplier, based on manufacturing process selection (evaluate R&M capability). 9. Generate detailed life cycle costing analysis on selected manufacturing process. At this point it is important to ask and answer these questions: Have broad, high level R&M targets been set to drive detailed process trade-off decisions? Is the life cycle cost analysis complete for the selected manufacturing process? Do the projections support the budget per the affordable business structure?

R&M

AND

PREVENTIVE MAINTENANCE (PM) NEEDS ANALYSIS

Activities associated with the R&M and PM needs analysis stage include: 1. Establish a clear definition of failure by using all known operating conditions and unique circumstances surrounding the process. 2. Establish R&M requirements for the unique operating conditions surrounding the chosen manufacturing process. 3. Establish/issue R&M engineering requirements for the project to the designers of the machinery. 4. Identify PM requirements for maintainability. At this point it is important to ask and answer this question: Have specific R&M targets been set to support the unique operating conditions and PM program objectives?

DEVELOPMENT AND DESIGN R&M PLANNING Activities associated with the R&M planning stage include:

SL3151Ch08Frame Page 351 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

351

1. Conduct process concept review. 2. Identify design effects for other related equipment (automation, integration, processing, etc.). 3. Standardize fault diagnostics (controls, software, interfaces, level of diagnosis, etc.). 4. Develop R&M/PM plan (process/machinery FMEA, mechanical/electrical derating, materials compatibility, thermal analyses, finite element analysis to support machine condition signature analysis, R&M predictions, R&M simulations, design for maintainability, etc.). 5. Establish R&M/PM testing requirements (burn-in testing, voltage cycling, probability ratio sequential testing, design of experiments for process optimization, environmental stress screening, life testing, test-analyze-fix, etc.). At this point it is important to ask and answer these questions: Does the R&M plan address each project target? Is the R&M plan sufficient to meet project targets?

PROCESS DESIGN

FOR

R&M

Activities associated with the process design for R&M stage include: 1. 2. 3. 4.

Conduct process design review. Develop process flow chart. Develop process simulation model. Conduct process design simulation for multiple scenarios by analyzing operational effects of various R&M design trade-offs. 5. Develop life cycle costing analysis on process-related equipment. 6. Review process FMEA. 7. Complete final process review and simultaneous engineering team input. At this point it is important to ask and answer this question: Is the process FMEA complete, and have causes of potentially common failure modes been addressed and redesigned?

MACHINERY FMEA DEVELOPMENT Activities associated with the machinery FMEA development stage include: 1. Develop plant floor computer data collection system (activity tracking, downtime, reliability growth curves). 2. Establish machinery data feedback plan (crisis maintenance, MTBF, MTTR, tool lives, OEE, production report, etc.). 3. Verify completion of machinery FMEA on all critical machinery. Confirm design actions, maintenance burdens, things gone wrong, root cause analyses, etc. 4. Develop fault diagnostic strategy (built in test equipment, rapid problem diagnosis, control measures).

SL3151Ch08Frame Page 352 Thursday, September 12, 2002 6:07 PM

352

Six Sigma and Beyond

5. Review equipment and material handling layouts (panels, hydro, coolant systems). At this point it is important to ask and answer these questions: Is the machinery FMEA complete, and have causes of potentially common failure modes been addressed and redesigned? Is the data collection plan complete?

DESIGN REVIEW Activities associated with the design review stage include: 1. Conduct machinery design review (field history, machinery FMEA, test or build problems, R&M simulation and reliability predictions, maintainability, thermal/mechanical/electrical analyses, etc.). 2. Provide R&M requirements to tier two suppliers (levels, root cause analyses, standardized component applications, testing, etc.). At this point it is important to ask and answer this question: Have the R&M plan requirements been incorporated in the machinery design?

BUILD AND INSTALL EQUIPMENT RUN-OFF Activities associated with the equipment run-off stage include: 1. Conduct machinery run-off (perform root cause analysis, Failure, Reporting Analysis, and Corrective Action System [FRACAS], complete testing, verify R&M and TPM requirements, validate diagnostic logic and data collection). 2. Complete preventative maintenance/predictive maintenance manuals and review maintenance burden. At this point it is important to ask and answer this question: Has the plant maintenance department devised a maintenance plan based on expected machine performance?

OPERATION

OF

MACHINERY

Activities associated with the operation of machinery stage include: 1. 2. 3. 4. 5.

Implement and utilize machinery data feedback plan. Implement and utilize FRACAS. Evaluate PM program. Update FMEA and reliability predictions. Conduct reliability growth curve development and analysis.

At this point it is important to ask and answer this question: Have design practices been documented for use by the next generation design teams? (Also note that as

SL3151Ch08Frame Page 353 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

353

the machinery begins to operate, the continuous improvement cycle phases begin to lead the R&M effort in phases four and five.)

OPERATIONS AND SUPPORT After the equipment has been installed and the run-off has been performed, the Durability phase of the cycle begins. The PM program now begins to utilize the R&M team member more as a team leader than a participant. Durability, as defined in the TE supplement, is the “ability to perform intended function over a specified period under normal use (with specified maintenance, without significant deterioration).” As the machinery begins to acquire additional operation hours, PM personnel identify issues and take corrective action. These issues and corrections are fed back to FMEA personnel and R&M planners as lessons learned for the next generation of machinery. Whether these corrections involve the design of the machinery or the maintenance schedule/tasks, each must be incorporated into the continuous improvement loop.

CONVERSION/DECOMMISSION Conversion is one of the key elements of the investment efficiency loop. The R&M process for reuse of equipment is very similar to the purchase of new equipment except that you have more limitations on the concept of the new process. The data are collected and phase one is repeated, often, with more specific direction as the current equipment may limit some of the other concepts. While decommission may be the process of equipment disposal, it is necessary to verify and record R&M data from this equipment to help identify the best design practices. It is also important to make note of those design practices that did not work as well as planned. As plans for decommission become firm, it is important to generate forecasts for equipment availability. These forecasts should then be entered into a database for future forecasted and available machinery and equipment. Maintenance data, including condition, operation description, and reason for availability should be included. This will assist engineers evaluating surplus machinery and equipment for reuse in their programs.

TYPICAL R&M MEASURES R&M MATRIX Perhaps the most important document in the R&M process is the R&M matrix. This matrix identifies the requirements of the customer on a per phase basis. Three major categories of tasks are usually identified. They are: R&M programmatic tasks Engineering tasks R&M continuous improvement

SL3151Ch08Frame Page 354 Thursday, September 12, 2002 6:07 PM

354

Six Sigma and Beyond

RELIABILITY POINT MEASUREMENT This may be expressed by:

R(t ) = e

 −t     MTBF 

where R(t) = reliability point estimate during a constant failure rate period; e = natural logarithm which is 2.718281828…; t = schedule time or mission time of the equipment or machinery; and MTBF = mean time between failure. Special note: This calculation may be performed only when the machine has reached the bottom of the bathtub curve. EXAMPLE A water pump is scheduled (mission time) to operate for 100 hours. The MTBF for this pump is also rated at 100 hours and the MTTR is 2 hours. The probability that the pump will not fail during the mission is:

R(t ) = e

 −t     MTBF 

= R(t ) = e

 −100     100 

= .37 or 37%.

This means that the pump will have a 37% chance of not breaking down during the 100-hour mission time. Conversely, the unreliability of the pump can be calculated as:

R = 1 – R = 1 – .37 = .63 or 63%. This means that the pump has a 63% chance of failing during the 100 hour mission.

MTBE Mean time between event can be calculated as: MTBE = Total Operating Time/N where Total Operating Time = the total scheduled production time when machinery or equipment is powered and producing parts and N = the total number of downtime events, scheduled and unscheduled. EXAMPLE The total operating time for a machine is 550 hours. In addition, the machine experiences 2 failures, 2 tool changes, 2 quality checks, 1 preventive maintenance meeting, and 5 lunch breaks. What is the MTBE?

SL3151Ch08Frame Page 355 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

355

MTBE = Total Operating Time/N = 550/12 = 45.8 hours

MTBF Mean time between failure is the average time between failure occurrences and is calculated as: MTBF = Operating Time/N where Operating Time = scheduled production time and N = total number of failures observed during the operating period. EXAMPLE If machinery is operating for 400 hours and there are eight failures, what is the MTBF? MTBF = Operating Time/N = 400/8 = 50 hours. (Special note: Sometimes C (cycles) is substituted for T. In that case, we calculate the MCBF. The steps are identical to those of the MTBF calculation.)

FAILURE RATE Failure rate estimates the number of failures in a given unit of time, events, cycles, or number of parts. It is the probability of failure within a unit of time. It is calculated as: Failure rate = 1/MTBF EXAMPLE The failure rate of a pump that experiences one failure within an operating time period of 2000 hours is:

Failure rate = 1/MTBF = 1/2000 = .0005 failures per hour. This means that there is a .0005 probability that a failure will occur with every hour of operation.

MTTR Mean time to repair is a calculation based on one failure and one failure only. The longer each failure takes to repair, the more the equipment’s cost of ownership goes up. Additionally, MTTR directly effects uptime, uptime percent, and capacity. It is calculated as:

MTTR =

∑t N

SL3151Ch08Frame Page 356 Thursday, September 12, 2002 6:07 PM

356

Six Sigma and Beyond

where

∑ t = total repair time and N = total number of repairs.

EXAMPLE A pump operates for 300 hours. During that period there were four failure events recorded. The total repair time was 5 hours. What is the MTTR?

MTTR =

∑t N

= 5/4 = 1.25 hours

AVAILABILITY Availability is the measure of the degree to which machinery or equipment is in an operable and committable state at any point in time. Availability is dependent upon (a) breakdown loss, (b) setup and adjustment loss, and (c) other factors that may prevent machinery from being available for operation when needed. When calculating this metric, it is assumed that maintenance starts as soon as the failure is reported. (Special note: Think of the measurement of R&M in terms of availability. That is, MTBF is reliability and MTTR is maintainability.) Availability is calculated as: Availability = MTBF/(MTBF + MTTR) EXAMPLE What is the availability for a system that has an MTBF of 50 hours and an MTTR of 1 hour?

Availability = MTBF/(MTBF + MTTR) = 50/(50 + 1) = .98 or 98%

OVERALL EQUIPMENT EFFECTIVENESS (OEE) Overall equipment effectiveness (OEE) is a measure of three variables. They are: 1. Availability = percent of time a machine is available to produce 2. Performance efficiency = actual speed of the machine as related to the design speed of the machine 3. Quality rate = percent of resulting parts that are within specifications A good OEE is considered to be 85% or higher.

LIFE CYCLE COSTING (LCC) Life cycle costing (LCC) is the total cost over the life of the machine or equipment. It is calculated based on the following: LCC = Acquisition costs (A) + Operating costs (O) + Maintenance costs (M) ± Conversion and or decommission costs (c)

SL3151Ch08Frame Page 357 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

357

TABLE 8.2 Cost Comparison of Two Machines Costs Acquisition costs (A) Operating costs (O) Maintenance costs (M) Conversion and/or decommission costs (C) Total LCC

Machine A

Machine B

$2,000.00 $9,360.00 $7,656.00

$1,520.00 $10,870.00 $9,942.00

$19,016.00

$22,332.00

EXAMPLE What is the LCC for the two machines shown in Table 8.2 and which one is a better deal? The reader should notice that before the decision is made all costs should be evaluated. In this case, machine A has a higher acquisition cost than machine B, but it turns out that machine A has a lower LCC than machine B. Therefore, machine A is the better deal.

TOP 10 PROBLEMS

AND

RESOLUTIONS

This list allows the designer to see the major sources of downtime associated with the current equipment. Once the list items are identified, a root cause analysis or problem resolution should be conducted on each of the failures. If the design is known, the designer can then modify the design to reflect the changes. (Sometimes the top ten problems are based on historical data and must be adjusted to reflect current design considerations.)

THERMAL ANALYSIS This analysis is conducted to help the designer to develop the appropriate and applicable heat transfer (Table 8.3). The actual analysis is conducted by following these six steps: 1. 2. 3. 4. 5.

Develop a list of all electrical components in the enclosure. Identify the wattage rating for each component located in the enclosure. Sum the total wattage for the enclosure. Add in any external heat generating sources. Calculate the surface area of the enclosure that will be available for cooling. 6. Calculate the thermal rise above ambient. EXAMPLE The electrical enclosure is 5 ft. tall by 4 ft. deep. The surface area for this enclosure is calculated as follows:

SL3151Ch08Frame Page 358 Thursday, September 12, 2002 6:07 PM

358

Six Sigma and Beyond

TABLE 8.3 Thermal Calculation Values Thermal Calculation Values Component Name

Quantity

Individual Wattage Maximum

Total Wattage

Internal Relay A18 contactor A25 contactor PS27 power supply Monochrome monitor

4 1 2 1 1

2.5 1.7 2 71 85 Subtotal wattage

10.0 1.7 4.0 71.0 85.0 171.7

External Servo transformer

1

450 Subtotal wattage Total enclosure wattage

63.0 63.0 234.7

Note: The servo transformer is mounted externally and next to the enclosure. Therefore, only 14% of the total wattage is estimated to radiate into the enclosure

Front and Back = 5 ft. × 4ft. × 2 = 40 sq. ft. Sides = 2 ft. × 5 ft. × 2 = 20 sq. ft. Enclosure top = 2ft. × 4ft. = 8 sq. ft Bottom is ignored due to the fact that heat rises.

Total surface area = 40 + 20 + 8 = 68 sq. ft. To calculate the thermal rise (∆T) we use the following formula:

Thermal rise (∆T) = Thermal resistance (θCA) cabinet to ambient × Power (W) θCA = 1/(Thermal conductivity × Cooling area) The thermal conductivity value is found in the catalog of the National Electrical Manufacturing Association (NEMA).

θCA = 1/(.25 W/degree F) × (square footage) θCA = 1/.25 × 68 = .0588 Thus, .25 W/degree F is the thermal conductivity value for a NEMA 12 enclosure. If the equipment inside the enclosure generates 234.7 watts, then the thermal rise is

SL3151Ch08Frame Page 359 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

359

∆T = θCA × wattage = .0588 × 234.7 = 13.8°F. If the ambient temperature is 100°F, then the enclosure temperature will reach 113.8°F. If the enclosure temperature is specified as 104°F, then the design exceeds the specification by approximately 9.8°F. The enclosure must be increased in size, the load must be reduced, or active cooling techniques need to be applied. (Special note: Remember that a 10% rise in temperature decreases the reliability by about 50%. Also the method just mentioned in this example is not valid for enclosures that have other means of heat dissipation such as fans, or for those made of heavier metal or if the material were changed. This specific calculation assumes that the heat is being radiated through convection to the outside air.)

ELECTRICAL DESIGN MARGINS Design margins in electrical engineering of the equipment are referred to as derating. On the other hand, mechanical design margins are referred to as safety margins. A rule of thumb for derating is about 20% for electrical components. However, the actual calculation is % derating = 1 −

IT IS

where IT = total circuit current draw and IS = total supply current. EXAMPLE During a design review, the question arose as to whether the 24 V power supply for a motor was adequately derated. The power supply takes 480 VAC three phase with a 2 A circuit breaker and has a rated output of 10 A. An examination of the system reveals that 24 V power is delivered to the load through three circuit breakers (A = .477 A, B = .73 A, and C = 5.53 A. The total for the three circuits is therefore 6.737 A.) When these circuit breakers are combined, 11 A of current flow to the load. This situation may not happen, but further investigation is required.

% derating = 1 −

IT 6.737 = 1− = 32.63% IS 10.0

This means that in this case the power supply will not be overloaded and the circuit breakers are generously oversized. In other words, the circuit breakers should not be tripped due to false triggers.

SAFETY MARGINS (SM) For mechanical components, SM are generally defined as the amount of strength of a mechanical component relating to the applied stress. A rule of thumb for SM with a normally distributed stress load relationship is that the safety margin should always be greater or equal to three. However, the actual calculation for the MS is

SL3151Ch08Frame Page 360 Thursday, September 12, 2002 6:07 PM

360

Six Sigma and Beyond

SM =

U STRENGTH − U STRESS Sv 2 + Lv 2

Where SM = safety margin; USTRENGTH = mean strength; ULOAD = mean load; Lv2 = load variance; and Sv2 = strength variance. EXAMPLE A robot’s arm has a mean strength of 80 kg. The maximum allowable stress applied by the end of arm tooling is 50 kg. The strength variance is 8 kg and the stress variance is 7 kg. What is the SM?

SM =

U STRENGTH − U STRESS Sv + Lv 2

2

=

80 − 50 82 + 72

= 2.822

(A low SM may indicate the need to assign another size robot or redesign the tooling material.)

INTERFERENCE Once the SM is calculated, it can be used to calculate the interference and reliability of the components under investigation. Interference may be thought of as the overlap between the stress and the strength distributions. In more formal terms, it is the probability that a random observation from the load distribution exceeds a random observation from the strength distribution. To calculate interference, we use the SM equation and substitute the z for the SM distribution:

Z=

U STRENGTH − U STRESS Sv 2 + Lv 2

EXAMPLE If we use the answer from the previous example (z = 2.822), we can use the z table (in this case the area under the z = 2.822 is .0024). This means that there exists a .0024 or .24% probability of failure. Reliability, on the other hand, may be calculated as

R = 1 – interference or R = 1 – α R = 1 – .0024 = .9976 or 99.76%. This means that even though the strength and the load have a very low (.24%) probability of failure, the reliability of the system is very high with a 99.76%.

SL3151Ch08Frame Page 361 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

361

TABLE 8.4 Guidelines for the Duane Model β

Recommended Actions

0 to .2

No priority is given to reliability improvement; failure data not analyzed; corrective action taken for important failure modes, but with low priority Routine attention to reliability improvement; corrective action taken for important failure modes Priority attention to reliability improvement; normal (typical stresses) environment utilization; well-managed analysis and corrective action for important failure modes Eliminating failures takes top priority; immediate analysis and corrective action for all failures

.2 to .3 .3 to .4 .4 to .6

CONVERSION

OF

MTBF

TO

FAILURE RATE

AND

VICE VERSA

The relationship between these two metrics is

MTBF =

1 1 and FR = FR MTBF

RELIABILITY GROWTH PLOTS This plot is an effective method to track continual improvement for R&M as well as to predict reliability growth of machinery from one machine to the other. The steps to generate this plot are: Step 1. Collect data on the machine and calculate the cumulative MTBF value for the machine. Step 2. Plot the data on log–log paper. (An increasing slope indicates a reliability growth flatness, which indicates that the machine has achieved its inherent level of MTBF and cannot get any better) Step 3. Calculate the slope, using regression analysis or best fit line. Once the slope (the beta value) is calculated, we can apply the Duane model interpretation. The guidelines (Table 8.4) for the interpretation are

MACHINERY FMEA Machinery FMEA is a systematic approach that applies the tabular method to aid the thought process used by simultaneous engineering teams to identify the machine’s potential failure modes, potential effects, and potential causes and to develop corrective action plans that will remove or reduce the impact of the failure modes. Perhaps the most important use of the machinery FMEA is to identify and correct all safety issues. A more detailed discussion will be given in Chapter 6.

SL3151Ch08Frame Page 362 Thursday, September 12, 2002 6:07 PM

362

Six Sigma and Beyond

KEY DEFINITIONS IN R&M The following terms are commonly encountered in R&M: Accelerated life testing — Verification of machine and equipment design relationship much sooner than if operated typically. Intended especially for new technology, design changes, and ongoing development. Derating — The practice of limiting stresses that may be applied to a component to levels below the specified maxima in order to enhance reliability. Derating values of electrical stress are expressed as ratios of applied stress to rated maximum stress. The applied stress is taken as the maximum likely to be applied during worst-case operating conditions. Thermal derating is expressed as a temperature value. Design of experiments (DOE) — A technique that focuses on identifying factors that affect the level or magnitude of a product/process response, examining the response surface, and forming the mathematical prediction model. Design review — A review providing in-depth detail relative to the evolving design supported by drawings, process flow descriptions, engineering analyses, reliability design features, and maintainability design considerations. Dry run — The rehearsal or cycling of machinery, normally with the intent of not processing the work piece, to verify function, clearances, and construction stability. Durability — Ability to perform intended function over a specified period under normal use with specified maintenance, without significant deterioration. Equipment — The portion of process machinery that is not specific to a component or sub assembly. Failure — An event when machinery/equipment is not available to produce parts under specified conditions when scheduled or is not capable of producing parts or performing scheduled operations to specifications. For every failure, an action is required. Failure mode and effects analysis (FMEA) — A technique to identify each potential failure mode and its effect on machinery performance. Failure reporting, analysis, and corrective action system (fracas) — An orderly system of recording and transmitting failure data from the supplier’s plant to the end users fits into a unitary database. The database allows identification of pattern failures and rapid resolution of problems through rigorous failure analysis. Fault tree analysis (FTA) — A top down approach to failure analysis starting with an undesirable event and determining all the ways it can happen. Feasibility — A determination that a process, design, procedure, or plan can be successfully accomplished in the required time frame. Finite element analysis (FEA) — A computational structure analysis technique that quantifies a structure’s response to applied loading conditions. Total productive maintenance (TPM) — Natural cross-functional groups working together in an optimal balance to improve the overall effectiveness

SL3151Ch08Frame Page 363 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

363

of their equipment and processes within their work areas. TPM implementation vigorously benchmarks, measures, and corrects all losses resulting from inefficiencies. Life cycle — The sequence through which machinery and equipment pass from conception through decommission. Life cycle costs (LCC) — The sum of all cost factors incurred during the expected life of machinery. Machine condition signature analysis (MCSA) — An application that applies mechanical signature (vibration) analysis techniques to characterize machinery and equipment on a systems level to significantly improve reliability and maintainability. Machinery — Tooling and equipment combined. A generic term for all hardware (including necessary operational software) that performs a manufacturing process. Maintainability — A characteristic of design, installation, and operation, usually expressed as the probability that a machine can be retained in, or restored to, specified operable condition within a specified interval of time when maintenance is performed in accordance with prescribed procedures. Mean time between failures (MTBF) — The average time between failure occurrences. The sum of the operating time of a machine divided by the total number of failures. Predominantly used for repairable equipment. Mean time to failure (MTTF) — The average time to failure for a specific equipment design. Used predominantly for non-repairable equipment. Mean time to repair (MTTR) — The average time to restore machinery or equipment to specified conditions. Overall equipment effectiveness (OEE) — Percentage of the time the machinery is available (Availability) × how fast the machinery is running relative to its design cycle (Performance efficiency) × percentage of the resulting product within quality specifications (Yield). Perishable tooling — Tooling which is consumed over time during a manufacturing operation. Plant floor information system (PFIS) — An information gathering system used on the plant floor to gather data relating to plant operations including maintenance activities. Predictive maintenance (PdM) — A portion of scheduled maintenance dedicated to inspection for the purpose of detecting incipient failures. Preventative maintenance (PM) — A portion of scheduled maintenance dedicated to taking planned actions for the purpose of reducing the frequency or severity of future failures, including lubrication, filter changes, and part replacement dictated by analytical techniques and predictive maintenance procedures. Probability ratio sequential testing (PRST) — A reliability qualification test to demonstrate if the machinery/equipment satisfies a specified MTBF requirement and is not lower than an acceptable MTBF (MIL-STD-781). Process — Any operation or sequence of operations that contributes to the transformation of raw material into a finished part or assembly.

SL3151Ch08Frame Page 364 Thursday, September 12, 2002 6:07 PM

364

Six Sigma and Beyond

Product — In relation to tooling and equipment suppliers, the term “product” refers to the end item produced (e.g., machine, tool, die, etc.). Production — In relation to tooling and equipment suppliers, the term “production” refers to the process required to produce the product. R&M plan — A reliability and maintainability (R&M) plan shall establish a clear implementation strategy for design assurance techniques, reliability testing and assessment, and R&M continuous improvement activities during the machinery/equipment life cycle. R&M targets — The range of values that MTBF and MTTR are expected to fall between plus an improvement factor that leads to MTBF and MTTR requirements. Reliability — The probability that machinery and equipment can perform continuously, without failure, for a specified interval of time when operating under stated conditions. Reliability growth — Machine reliability improvement as a result of identifying and eliminating machinery or equipment failure causes during machine testing and operations. Root cause analysis (RCA) — A logical, systematic approach to identifying the basic reasons (causes, mechanisms, etc.) for a problem, failure, nonconformance, process error, etc. The result of root cause analysis should always be the identification of the basic mechanism by which the problem occurs and a recommendation for corrective action. Simultaneous engineering (SE) — Product engineering that optimizes the final product by the proper integration of requirements, including product function, manufacturing and assembly processing, service engineering, and disposal. Things gone right/things gone wrong (TGR/TGW) — An evolving program-level compilation of lessons learned that capture successful and unsuccessful manufacturing engineering activity and equipment/performance for feedback to an organization and its suppliers for continuous improvement. Tooling — The portion of the process machinery that is specific to a component of sub assembly.

DFSS AND R&M R&M’s goal is to make sure that the machinery/tool delivered to the customer meets or exceeds its requirements. DFSS, on the other hand, is the methodology that controls the process for satisfying the customer’s expectations early on in the product development cycle. This is very important since in R&M the reliability matrix actually attempts to quantify the initial product vision with the customer’s requirements. Having said that, we must also recognize that quite often in product development we do not have all the answers. In fact, quite often we are on a fuzzy front end. This is where DFSS offers its greatest contribution. That is, with the process knowledge of DFSS, the engineer not only will be aware but also will make sure that the appropriate design fits within both the customer’s and the organization’s goals.

SL3151Ch08Frame Page 365 Thursday, September 12, 2002 6:07 PM

Reliability and Maintainability

365

DFSS may be applied in an original design, which involves elaborating original solutions for a given task; adaptive design, which involves adapting a known system to a changed task or evolving a significant subsystem of a current product; variant design, which involves varying parameters of certain aspects of a product to develop a new or more robust design; and redesign, which implies any of the items just mentioned. A redesign is not a variant design, rather it implies that a product already exists that is perceived to fall short in some criteria, and a new solution is needed. The new solution can be developed through any of the above approaches. In fact, it is often difficult to argue against the maxim that all design is redesign (Otto and Wood, 2001).

REFERENCES Otto, K. and Wood, K., Product Design, Prentice Hall, Upper Saddle River, NJ, 2001.

SELECTED BIBLIOGRAPHY Anon., Reliability and Maintainability Guideline for Manufacturing Machinery and Equipment, M-110.2, 2nd ed., Society of Automotive Engineers, Inc., Warrendale, PA and National Center for Manufacturing Sciences, Inc., Ann Arbor, MI, 1999. Anon., ISO/TS16949. International Automotive Task Force. 2nd ed. AIAG. Southfield, MI, 2002. Automotive Industry Action Group, Potential Failure Mode and Effect Analysis, 3rd ed., Chrysler Corp., Ford Motor Co., and General Motors. Distributed by AIAG, Southfield, MI, 2001. Blenchard, B.S., Logistics Engineering and Management, 3rd ed., Prentice Hall, Englewood Cliffs, NJ, 1986. Chrysler, Ford, and GM, Quality System Requirements: QS-9000, distributed by Automotive Industry Action Group, Southfield, MI, 1995. Chrysler, Ford, and GM, Quality System Requirements: Tooling and Equipment Supplement, distributed by Automotive Industry Action Group, Southfield, MI, 1996. Creveling, C.M., Tolerance Design: A Handbook for Developing Optimal Specifications, Addison Wesley Longman, Reading, MA, 1997. Hollins, B. and Pugh, S., Successful Product Design, Butterworth Scientific. London, 1990. Kapur, K.C. and Lamberson, L.R., Reliability in Engineering Design, Wiley, New York, 1977. Nelson, W., Graphical analysis of system repair data, Journal of Quality Technology, 20, 24–35, 1988. Stamatis, D.H., Implementing the TE Supplement to QS-9000, Quality Resources, New York, 1998.

SL3151Ch08Frame Page 366 Thursday, September 12, 2002 6:07 PM

SL3151Ch09Frame Page 367 Thursday, September 12, 2002 6:05 PM

9

Design of Experiments SETTING THE STAGE FOR DOE

Design of Experiments (DOE) is a way to efficiently plan and structure an investigatory testing program. Although DOE is often perceived to be a problem-solving tool, its greatest benefit can come as a problem avoidance tool. In fact, it is this avoidance that we emphasize in design for six sigma (DFSS). This chapter is organized into nine sections. The user who is looking for a basic DOE introduction in order to participate with some understanding in a problemsolving group is urged to study and understand the first two sections or go back and review Volume V of this series. The remaining sections discuss more complex topics including problem avoidance in product and process design, more advanced experimental layouts, and understanding the analysis in more detail.

WHY DOE (DESIGN

OF

EXPERIMENTS) IS

A

VALUABLE TOOL

DOE is a valuable tool because: 1. DOE helps the responsible group plan, conduct, and analyze test programs more efficiently. 2. DOE is an effective way to reduce cost. Usually the term DOE brings to mind only the analysis of experimental data. The application of DOE necessitates a much broader approach that encompasses the total process involved in testing. The skills required to conduct an effective test program fall into three main categories: 1. Planning/organizational 2. Technical 3. Analytical/statistical The planning of the experiment is a critical phase. If the groundwork laid in the planning phase is faulty, even the best analytic techniques will not salvage the disaster. The tendency to run off and conduct tests as soon as a problem is found, without planning the outcome, should be resisted. The benefits from up-front planning almost always outweigh the small investment of time and effort. Too often, time and resources are wasted running down blind alleys that could have been avoided. Section 2 of this chapter contains a more detailed discussion of planning and the techniques used to ensure a well-planned experiment. 367

SL3151Ch09Frame Page 368 Thursday, September 12, 2002 6:05 PM

368

Six Sigma and Beyond

TABLE 9.1 One Factor at a Time The group tests configurations containing the following combinations of the factors: Level of Factor (1 and 2 Indicate the Different Levels)

Results

Test Number

A

B

C

D

E

F

G

a

b

1 2 3 4 5 6 7 8

1 2 1 1 1 1 1 1

1 1 2 2 2 2 2 2

1 1 1 2 1 1 1 1

1 1 1 1 2 2 2 2

1 1 1 1 1 2 1 1

1 1 1 1 1 1 2 1

1 1 1 1 1 1 1 2

271.4 215.0 275.3 235.2 296.6 305.2 278.8 251.9

266.3 211.2 271.1 231.5 301.6 301.1 275.3 254.3

DOE can be a powerful tool in situations where the effect on a measured output of several factors, each at two or more levels, must be determined. In the traditional “one factor at a time” approach, each test result is used in a small number of comparisons. In DOE, each test is used in every comparison. A simplified example follows. EXAMPLE A problem-solving brainstorming group suspects 7 factors (named A, B, C, D, E, F, and G), each at two levels (level 1 and level 2), of influencing a critical, measurable function of the design. The group wants to determine the best settings of these factors to maximize the measured test results — see Table 9.1. Two evaluations (a and b) are run at each test configuration rather than a single evaluation in order to attain a higher confidence in the difference between factor levels (this assumes no need for a “tie breaker”). The group makes comparisons as shown in Table 9.2. Sixteen total tests are run, and four tests are used to determine the difference between levels for each factor. The best combination of factors is (1, 2, 1, 2, 2, 1, 1) for factors A through G. However, using DOE the group runs test configurations as shown in Table 9.3. The group makes comparisons as shown in Table 9.4. Eight total tests are run, and eight tests are used to determine the difference between levels for each factor. This can be done because each level of every factor equally impacts the determination of the average response at all levels of all of the other factors (i.e., of the four tests run at A = 1, two were run at B = 1 and two were run at B = 2; this is also true of the four tests run at A = 2). This relationship is called orthogonality. This concept is very important, and the reader should work through the relationships between the levels of at least two other factors to better understand the use of orthogonality in this testing matrix. The best level is [1, (1 or 2), 1, 2, (1 or 2), 1, 1] for A through G. Factors B and E are not significant and may be set to the least expensive level.

SL3151Ch09Frame Page 369 Thursday, September 12, 2002 6:05 PM

Design of Experiments

369

TABLE 9.2 Test Numbers for Comparison Test Numbers Used to Determine: Factor A B C D E F G

Level 1 1a, 1a, 3a, 3a, 5a, 6a, 6a,

Difference Level 1 – Level 2

Level 2

1b 1b 3b 3b 5b 6b 6b

2a, 3a, 4a, 5a, 6a, 7a, 8a,

2b 3b 4b 5b 6b 7b 8b

55.8 –4.4 39.9 –25.7 –4.3 26.1 50.1

TABLE 9.3 The Group Runs Using DOE Configurations Level of Factor (1 and 2 Indicate the Different Levels)

Test Number

A

B

C

D

E

F

G

Result

1 2 3 4 5 6 7 8

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

270.7 223.8 158.2 263.1 129.3 175.1 195.4 194.6

TABLE 9.4 Comparisons Using DOE Test Numbers Used to Determine Factor A B C D E F G

Level 1

Level 2

1, 1, 1, 1, 1, 1, 1,

5, 3, 3, 2, 2, 2, 2,

2, 2, 2, 3, 3, 4, 4,

3, 5, 7, 5, 6, 5, 6,

4 6 8 7 8 8 7

6, 4, 4, 4, 4, 3, 3,

7, 7, 5, 6, 5, 5, 5,

8 8 6 8 7 8 8

Difference Level 1 – Level 2 55.4 –3.1 39.7 –25.8 –3.3 26.3 49.6

SL3151Ch09Frame Page 370 Thursday, September 12, 2002 6:05 PM

370

Six Sigma and Beyond

TABLE 9.5 Comparison of the Two Means Number of Tests

Estimate at the Best Levels

Confidence Interval at 90% Confidence

16 8

301.1 299.6

± 3.7 ± 3.3

One factor at a time DOE

For a comparison of the two methods, see Table 9.5. Half as many tests are required using a DOE approach and the estimate at each level is better (four tests per factor level versus two). This is almost like getting something for nothing. The only thing that is required is that the group plan out what is to be learned before running any of the tests. The savings in time and testing resources can be significant. Direct benefits include reduced product development time, improved problem correction response, and more satisfied customers. And that is exactly what DFSS should be aiming at.

This approach to DOE is also very flexible and can accommodate known or suspected interactions and factors with more than two levels. A properly structured experiment will give the maximum amount of information possible. An experiment that is less well designed will be an inefficient use of scarce resources.

TAGUCHI’S APPROACH Here it is appropriate to summarize Dr. Taguchi’s approach, which is to minimize the total cost to society. He uses the “Loss Function” (Section 4) to evaluate the total cost impact of alternative quality improvement actions. In Dr. Taguchi’s view, we all have an important societal responsibility to minimize the sum of the internal cost of producing a product and the external cost the customer incurs in using the product. The customer’s cost includes the cost of dissatisfaction. This responsibility should be in harmony with every company’s objectives when the long-term view of survival and customer satisfaction is considered. Profits may be maximized in the short run by deceiving today’s customers or trading away the future. Traditionally, the next quarter’s or next year’s “bottom line” has been the driving force in most corporations. Times have changed, however. Worldwide competition has grown, and customers have become more concerned with the total product cost. In this environment, survival becomes a real issue, and customer satisfaction must be a part of the cost equation that drives the decision process. Dr. Taguchi uses the signal-to-noise (S/N) ratio as the operational way of incorporating the loss function into experimental design. Experiment S/N is analogous to the S/N measurement developed in the audio/electronics industry. S/N is used to ensure that designs and processes give desired responses over different conditions of uncontrollable “noise” factors. S/N is introduced in Section 4 and developed in examples in later sections.

SL3151Ch09Frame Page 371 Thursday, September 12, 2002 6:05 PM

Design of Experiments

371

There are three basic types of product design activity in Dr. Taguchi’s approach: 1. System design 2. Parameter design 3. Tolerance design System design involves basic research to understand nature. System design involves scientific principles, their extension to unknown situations, and the development of highly structured basic relationships. Parameter and tolerance design involves optimizing the system design using empirical methods. Taguchi’s methods are most useful in parameter and tolerance designs. The rest of this chapter will discuss these applications. Parameter design optimizes the product or process design to reach the target value with minimum possible variability with the cheapest available components. Note the emphasis on striving to satisfy the requirements in the least costly manner. Parameter design is discussed in Section 8. Tolerance design only occurs if the variability achieved with the least costly components is too large to meet product goals. In tolerance design, the sensitivity of the design to changes in component tolerances is investigated. The goal is to determine which components should be more tightly controlled and which are not as crucial. Again, the driving force is cost. Tolerance design is discussed in Section 9. Problem resolution might appear to be another type of product design. If targets are set correctly, however, and parameter and tolerance design occur, there will be little need for problem resolution. When problems do arise, they are attacked using elements of both parameter and tolerance design, as the situation warrants.

MISCELLANEOUS THOUGHTS A tremendous opportunity exists when the basic relationships between components are defined in equation form in the system design phase. This occurs in electrical circuit design, finite element analysis, and other situations. In these cases, once the equations are known, testing can be simulated on a computer and the “best” component values and appropriate tolerances obtained. It might be argued that the true best values would not be located using this technique; only the local maxima would be obtained. The equations involved are generally too complex to solve to the true best values using calculus. Determining the local best values in the region that the experienced design engineer considers most promising is generally the best available approach. It definitely has merit over choosing several values and solving for the remaining ones. The cost involved is computation time, and the benefit is a robust design using the widest possible tolerances. Those readers who have some experience in classical statistics may wonder about the differences between the classical and Taguchi approaches. Although there are some operational differences, the biggest difference is in philosophical emphasis — see Volume V of this series. Classical statistics emphasizes the producer’s risk. This means a factor’s effect must be shown to be significantly different

SL3151Ch09Frame Page 372 Thursday, September 12, 2002 6:05 PM

372

Six Sigma and Beyond

from zero at a high confidence level to warrant a choice between levels. Taguchi uses percent contribution as a way to evaluate test results from a consumer’s risk standpoint. The reasoning is that if a factor has a high percent contribution, more often than not it is worth pursuing. In this respect, the Taguchi approach is less conservative than the classical approach. Dr. Taguchi uses orthogonal arrays extensively in his approach and has formulated them into a “cookbook” approach that is relatively easy to learn and apply. Classical statistics has several different ways of designing experiments including orthogonal arrays. In some cases, another approach may be more efficient than the orthogonal array. However, the application of these methods may be complex and is usually left to statisticians. Dr. Taguchi also approaches uncontrollable “noise” differently. He emphasizes developing a design that is robust over the levels of noise factors. This means that the design will perform at or near target regardless of what is happening with the uncontrollable factors. Classical statistics seeks to remove the noise factors from consideration by “blocking” the noise factors. In certain cases, the approaches Taguchi recommends may be more complicated than other statistical approaches or may be questioned by classical statisticians. In these cases, alternative approaches are presented as supplemental information at the end of the appropriate section. Additional analysis techniques are also presented in section supplements. The reader is encouraged to thoroughly analyze the data using all appropriate tools. Incomplete analysis can result in incorrect conclusions.

PLANNING THE EXPERIMENT The purpose of this section is to: 1. Impress upon the reader the importance of planning the experiment as a prerequisite to achieving successful results 2. Present some tools to use and points to consider during the planning phase 3. Demonstrate DOE applications via simple examples

BRAINSTORMING The first steps in planning a DOE are to define the situation to be addressed, identify the participants, and determine the scope and the goal of the investigation. This information should be written down in terms that are as specific as possible so that everyone involved can agree on and share a common understanding and purpose. The experts involved should pool their understanding of the subject. In a brainstorming session, each participant is encouraged to offer an opinion of which factors cause the effect. All ideas are recorded without question or discussion at this stage. To aid in the organization of the proposed factors, a branching (fishbone) format is often used, where each main branch is a main aspect of the effect under investigation (e.g., material, methods, machine, people, measurement, environment). The construction of a cause-and-effect (fishbone or Ishikawa) diagram in a brainstorming session provides a structured, efficient way to ensure that pertinent ideas are collected

SL3151Ch09Frame Page 373 Thursday, September 12, 2002 6:05 PM

Design of Experiments

Cooling

373

Engine Control Hardware Calibration Spark Scatter Poor Ground Injectors F/A Stuck Too Great Ratio Broken Spark Advance Contaminated Range Internal Harness & Too Small to Veh Connectors Intermittents EMI Fuel Improper Air Ports Flow Connector CBs Fit Piston Ring Scuff/Power Bolt Torque Loss Bore Fit Distortion Compression Grinding Height Piston Piston Rings Design Timing Suppliers Bore Buffs Camshaft Finish Finish Compression Ratio Assembly Engine Manufacturing Hardware

FIGURE 9.1 An example of a partially completed fishbone diagram.

and considered and that the discussion stays on track. An example of a partially completed cause-and-effect diagram is shown in Figure 9.1. After the participants have expressed their ideas on possible causes, the factors are discussed and prioritized for investigation. Usually, a three-level (high, moderate, and low) rating system is used to indicate the group consensus on the level of suspected contribution. Quite often, the rating will be determined by a simple vote of the participants. In situations where several different areas of contributing expertise are represented, participants’ votes outside of their areas of expertise may not have the importance of the expert’s vote. Handling this situation becomes a management challenge for the group leader and is beyond the scope of this document — the reader may need to review Volume II of this series. During the brainstorming and prioritization process, the participants should consider the following: 1. The situation — What is the present state of affairs and why are we dissatisfied? 2. The goal — When will we be satisfied (at least in the short term)? 3. The constraints — How much time and resources can we use in the investigation? 4. The approach — Is DOE appropriate right now or should we do other research first? 5. The measurement technique and response — What measurement technique will be used and what response will be measured?

CHOICE

OF

RESPONSE

The choice of measurement technique and response is an important point that is sometimes not given much thought. The obvious response is not always the best.

SL3151Ch09Frame Page 374 Thursday, September 12, 2002 6:05 PM

374

Six Sigma and Beyond

Factor 2 = Low

Response Factor 2 = High

Low Level

High Level Factor 1

FIGURE 9.2 An example of interaction.

As an example, consider the gap between two vehicle body panels. At first thought, that gap could be used as the response in a DOE aimed at achieving a target gap. However, the gap can be a symptom of more basic problems with the: • Width of the panels • Location holes in the panels • Location of the attachment points on the body frame All of these must be right for the gap to be as intended. If the goal of the experiment is to identify which of these has the biggest impact on the gap, the choice of the gap as a response is appropriate. If the purpose is to minimize the deviation from the target gap, the gap may not be the right response. A more basic investigation of the factors that contribute to the underlying cause is required. Do not confuse the symptom with the underlying causes. This thought process is very similar to the thought process used in SPC and failure mode and effect analysis (FMEA) and draws heavily upon the experience of experts to frame the right question. In DOE, the choice of an improper response could result in an inconclusive experiment or in a solution that might not work as things change due to interactions between the factors. An interaction occurs when the change in the response due to a change in the level of a factor is different for the different levels of a second factor. An example is shown in Figure 9.2. The choice of the proper response characteristic will usually result in few interactions being significant. Since there is a limitation as to how much information can be extracted from a given number of experiments, choosing the right response will allow the investigation of the maximum number of factors in the minimum number of tests without interactions between factors blurring the factor main effect. Interactions will be discussed in more detail in Section 3. The proper setup of an

SL3151Ch09Frame Page 375 Thursday, September 12, 2002 6:05 PM

Design of Experiments

375

experiment is not only a statistical task. Statistics serve to focus the technical expertise of the participating experts into the most efficient approach. In summary, the response should: 1. Relate to the underlying causes and not be a symptom 2. Be measurable (if possible, a continuous response should be chosen) 3. Be repeatable for the test procedure The prioritization process continues until the most critical factors that can be addressed within the resources of the test program are identified. The next step is to determine: 1. Are the factors controllable or are some of them “noise” beyond our control? 2. Do the factors interact? 3. What levels of each factor should be considered? 4. How do these levels relate to production limits or specs? 5. Who will supply the parts, machines, and testing facilities, and when will they be available? 6. Does everyone agree on the statement of the problem, goal, approach, and allocation of roles? 7. What kind of test procedure will be used? When all of these questions have been answered, the person who is acting as the statistical resource for the group can translate the answers into a hypothesis and experimental setup to test the hypothesis. The following example illustrates how the process can work: EXAMPLE A particular bracket has started to fail in the field with a higher than expected frequency. Timothy, the design engineer, and Christine, the process engineer, are alerted to the problem and agree to form a problem-solving team to investigate the situation. Timothy reviews the design FMEA, while Christine reviews the process FMEA. The information relating to the previously anticipated potential causes of this failure and SPC charts for the appropriate critical characteristics are brought to the first meeting. The team consists of Timothy, Christine, Cary (the machine operator), Stephen (the metallurgist), and Eric (another manufacturing engineer who has taken a DOE course and has agreed to help the group set up the DOE). In the first meeting, the group discussed the applicable areas from the FMEAs, reviewed the SPC charts, and began a cause-and-effect listing for the observed failure mode. At the conclusion of the meeting, Timothy was assigned to determine if the loads on the bracket had changed due to changes in the components attached to it; Christine was asked to investigate if there had been any change to the incoming material; Stephen was asked to consider the testing procedure that should be used to duplicate field failure modes and the response that should be measured, and all

SL3151Ch09Frame Page 376 Thursday, September 12, 2002 6:05 PM

376

Six Sigma and Beyond

Machine

Operator/Machine Interface C8

Material C1

C4

C9

C5

C6

C2

C3

C7 Bracket Breaks

C16

C12

C14

C17

C15

Process

C10 C13 C11 Design

FIGURE 9.3 Example of cause-and-effect diagram.

TABLE 9.6 The Test Matrix for the Seven Factors Test Number 1 2 3 4 5 6 7 8

Levels for Each Suspected Factor for Each of Eight Tests C1

C2

C7

C11

C13

C15

C16

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

of the group members were asked to consider additions to the cause-and-effect list. At the second meeting, the participants reported on their assignments and continued constructing the cause-and-effect (C & E) diagram. Their cause-and-effect diagram is shown in Figure 9.3 with the specific causes shown as “C1, C2, …” rather than the actual descriptions that would appear on a real C & E diagram. The group easily reached the consensus that seven of the potential causes were suspected of contributing to the field problem. Eric agreed to set up the experiment assuming two levels for each factor, and the others determined what those levels should be to relate the experiment to the production reality. Eric returned to the group and announced that he was able to use an L8 orthogonal array to set up the experiment and that eight tests were all that were needed at this time. The test matrix for the seven suspected factors is shown in Table 9.6. Eric explained that this matrix would allow the group to determine if a difference in test responses existed for the two levels of each factor and would prioritize the

SL3151Ch09Frame Page 377 Thursday, September 12, 2002 6:05 PM

Design of Experiments

377

TABLE 9.7 Test Results

10 13 15 17 14 16 19 21

C2

–

–

–

1 2 LEVEL

C16

1 2 LEVEL

–

–

– – – – – – –

–

–

1 2 LEVEL

C15

– – – – – – 1 2 LEVEL

– – – – – –

–

– – – – – –

–

–

–

1 2 LEVEL

C13 18 17 16 15 14 13

C11

C7

– – – – – –

–

1 2 3 4 5 6 7 8

– – – – – –

– – – – – – 1 2 LEVEL

R E S P O N S E

Result

C1 18 17 16 15 14 13

–

R E S P O N S E

Test Number

1 2 LEVEL

FIGURE 9.4 Plots of averages (higher responses are better). within-factor differences. Since the two levels of each factor represented an actual situation that existed in production during the time the failed parts were produced, this information could be used to correct the problem. By now, Stephen had identified a test procedure and response that seemed to fit the requirements outlined in this section.

Two weeks were required to gather all the material and parts for the experiment and to run the experiment. The test results are shown in Table 9.7. While Eric entered the data into the computer for analysis, Timothy and Christine plotted the data to see if anything was readily apparent. The factor level plots are shown in Figure 9.4.

SL3151Ch09Frame Page 378 Thursday, September 12, 2002 6:05 PM

378

Six Sigma and Beyond

Part 1 5 factors

Part 2 2 factors

Part 3 5 factors

Part 4 3 factors

Part 5 6 factors

FIGURE 9.5 A linear example of a process with several factors. When Eric finished with the computer, he reported that of all the variability observed in the data, 53.65% was due to the change in factor C2; 33.38% was due to the change in factor C1; and 11.92% was due to the change in factor C11. The remaining 1.04% was due to the other factors and experimental error. The large percentage variability contribution, coupled with the fact that the differences between the levels of the three factors are significant from an engineering standpoint, indicate that these three factors may indeed be the culprits. The computer analysis indicated that the best estimate for a test run at C1 = 2, C2 = 2, and C11 = 2 is 21.4. One of the eight tests in the experiment was run at this condition and the result was 21. Two confirmatory tests were run and the results were 11 and 20. The group then moved into a second phase of the investigation to identify what the specs limits should be on C1, C2, and C11. In the second round of testing, eight tests were required to investigate three levels for each of the three factors. The setup for the second round of testing involved an advanced procedure (idle column method) that will be presented later in this chapter, so the example will be concluded for now.

In summary, the group in the example took the following actions: 1. 2. 3. 4. 5. 6. 7. 8.

Gathered appropriate backup data Called together the right experts Made a list of the possible causes for the problem Prioritized the possible causes Determined the proper test procedure and response to be measured Reached agreement prior to running any tests Approached the investigation in a structured manner Asked and addressed one question at a time

Obviously, there are many ways to approach a particular DOE. In a situation where testing or material is very expensive, the most efficient experimental layout must be used. In the following sections, techniques are introduced that help the experimenter optimize the experimental design. Additional opportunities to optimize the experiment should be examined. Consider the situation where there is a five-part process. A brainstorming group has constructed a cause-and-effect diagram for a particular process problem. The number of suspected factors for each part of the process is shown in Figure 9.5. The obvious approach would be to set up the experiment with 21 factors. An alternative approach would be to consider only seven factors for the first round of testing. These would be the six factors within part 5 plus one factor for the best and worst input to part 5. If the difference in input to part 5 is significant, then the investigation is expanded upstream. The decision to approach a problem in this manner is dependent upon the beliefs of the experts. If the experts have a strong

SL3151Ch09Frame Page 379 Thursday, September 12, 2002 6:05 PM

Design of Experiments

379

TABLE 9.8 An Example of Contrasts Factor C1 C2 C7 C11 C13 C15 C16

Average at Level One

Average at Level Two

13.75 13.25 15.18 14.50 15.50 15.50 15.50

17.50 18.00 15.50 16.75 15.75 15.75 15.75

Contrast (Level 2 Avg. – Level 1 Avg.) 3.75 4.75 –0.25 2.25 0.25 0.25 0.25

prior belief that a factor in part 1, for instance, is significant, then a different approach should be used. This approach is also dependent upon the structure of the situation. The above example is presented to illustrate the point that the experimenter should be alert for ways to test more efficiently and effectively.

MISCELLANEOUS THOUGHTS An additional useful method of looking at the data is to plot the contrasts on normal probability paper. For a two-level factor, the contrast is the average of all the tests run at one level subtracted from the average of the tests run at the other level. For the example in this section, the contrasts are shown in Table 9.8. These contrasts are plotted on normal probability paper versus median ranks. The values for median ranks are available in many statistics and reliability books and are used in Weibull reliability plotting. For this example, the normal contrast plot is shown in Figure 9.6. To plot the contrasts on normal paper, the contrasts are ranked in numerical order, here from –0.25 (C7) to 4.75 (C2). The contrasts are then plotted against the median ranks or, in this case, against the rank number shown on the left margin of the plot. Factors that are significant have contrasts that are relatively far from zero and do not lie on a line roughly defined by the rest of the factors. These factors can lie off the line on the right side (level 2 higher) or on the left side (level 1 higher). In the example, two separate lines seem to be defined by the contrasts. This could be due to either of these situations: • C1, C2, and C11 are significant and the others are not. • There may be one or more bad data points that occur when C1, C2, and C11 are at one level and the other factors are set at the other level. In this example, C1, C2, C11, and C16 were at level 2 and the other factors were set at level 1 for run number eight. Depending upon the situation, it would be worthwhile to either rerun that test or to investigate the circumstances that accompanied that the test (e.g., was the test hard to run because of the factor settings or

SL3151Ch09Frame Page 380 Thursday, September 12, 2002 6:05 PM

380

Six Sigma and Beyond

C2

Numerical Rank Corresponding To Median Rank Probability

7

6

C1

C11 5

C16 4

3

C13 2

C15 C7 1 0

-1

1

2

3

4

5

Contrast

FIGURE 9.6 Contrasts shown in a graphical presentation.

did something else change that was not in the experiment?). In the example, this combination of factors represented the best observed outcome, and the confirmation runs supported the results of the original test. Plotting contrasts is a way of better understanding the data. It helps the experimenter visualize what is happening with the data. Sometimes, information that might be lost in a table of data will be crystal clear on a plot.

SETTING UP THE EXPERIMENT This section discusses: 1. 2. 3. 4.

The choice of the number of levels for each factor Fitting a linear graph to the experiment Special applications to reduce the number of tests How to handle noise factors in an experiment

CHOICE

OF THE

NUMBER

OF

FACTOR LEVELS

To review: A factor is a unique component or characteristic about which a decision will be made.

SL3151Ch09Frame Page 381 Thursday, September 12, 2002 6:05 PM

Design of Experiments

381

A factor level is one of the choices of the factor to be evaluated (e.g., if the screw speed of a machine is the factor to be investigated, two factor levels might be 1200 and 1400 rpm). Investigating a larger number of levels for a factor requires more tests than investigating a smaller number of levels. There is usually a trade-off required concerning the amount of information needed from the experiment to be very confident of the results and the time and resources available. If testing and material are cheap and time is available, evaluate many levels for each factor. Usually, this is not the case, and two or three levels for each factor are recommended. An exception to this occurs when the factor is non-continuous, and several levels are of interest. Examples of this type of factor include the evaluation of a number of suppliers, machines, or types of material. This situation will be discussed later in this section. The first round of testing is usually designed to screen a large number of factors. To accomplish this in a small number of tests, two levels per factor are usually tested. The choice of the levels depends upon the question to be addressed. If the question is “Have we specified the right spec limits?” or “What happens to the response in the clear worst possible situation?” then the choice of what the levels should be clear. A more complicated question to address is “How will the distribution in production affect the response?” As suppliers become capable of maintaining low variability about a target value, testing at the spec limits will not give a good answer to this question. There are at least two approaches that can be used: 1. Test at the production limits, as a worst case. 2. Test at other points that put less emphasis on the tails of the distribution where few parts are produced and more emphasis on the bulk of the distribution. It is a difficult choice to pick two points to represent an entire distribution. If this approach is being used, a rule of thumb is to choose a level that encompasses approximately 70% of that distribution (mean ± 1 standard deviation). The main point of this discussion is that the choice of levels is an integral part of the experimental definition and should be carefully considered by the group setting up the experiment. The second and subsequent rounds of testing are usually designed to investigate particular factors in more detail. Generally, three levels per factor are recommended. Using two levels allows the estimation of a linear trend between the points tested. The testing of three levels gives an indication of non-linearity of the response across the levels tested. This non-linearity can be used in determining specification limits to optimize the response. Although this concept will be explored in more detail in a later section on tolerance design, its application can be illustrated as follows: First round of testing — Level B of factor 1 gives a response that is more desirable than that given by level A. See Figure 9.7. Second round of testing — Level B gives a response that is more desirable than those given by either C or D. However, the differences are not great. Spec limits are set at C and D with B as the nominal. See Figure 9.8.

SL3151Ch09Frame Page 382 Thursday, September 12, 2002 6:05 PM

382

Six Sigma and Beyond

Response

Levels of Factor 1

B

A

FIGURE 9.7 First round testing.

Response

Levels of Factor 1

C

B

D

FIGURE 9.8 Second round testing.

In a manner similar to the two-level-per-factor situation, the choice of the specific three levels to be tested depends upon the question under investigation. Testing at three levels can be used by the experimenter to focus on a particular area of the possible factor settings to optimize the response over as large a range as possible. If three levels of a factor are used to gain understanding for an entire distribution, a rule of thumb is to choose the levels at the mean and mean ± 1.225 standard deviations that encompass approximately 78% of the distribution. These rules of thumb will be used in tolerance design.

LINEAR GRAPHS After the number of levels has been determined for each factor, the next step is to decide which experimental setup to use. Dr. Taguchi uses a tool called “linear graphs” to aid the experimenter in this process. Linear graphs are provided in the Appendix of Volume V for several situations. Typical designs, however, are: 1. All factors at two levels (L4, L8, L12, L16, L32) 2. All factors at three levels (L9, L27) 3. A mix of two- and three-level factors (L18, L36)

SL3151Ch09Frame Page 383 Thursday, September 12, 2002 6:05 PM

Design of Experiments

DEGREES

OF

383

FREEDOM

In the orthogonal array designation, the number following the L indicates how many testing setups are involved. This number is also one more than the degrees of freedom available in the setup. Degrees of freedom are the number of pair-wise comparisons that can be made. In comparing the levels of a two-level factor, one comparison is made and one degree of freedom is expended. For a three-level factor, two comparisons are made as follows: first, compare A and B, then compare whichever is “best” with C to determine which of the three is “best.” Two degrees of freedom are expended in this comparison. Once the number of levels for each factor is determined, the degrees of freedom required for each factor are summed. This sum plus one becomes the bottom limit to the orthogonal array choice. The degrees of freedom for an interaction are determined by multiplying the degrees of freedom for the factors involved in the interaction. A two-level factor interacting with a two-level factor requires one degree of freedom (df) (1 × 1 = 1). A three-level factor interacting with a three-level factor requires 4 df (2 × 2 = 4). A three-level factor interacting with a two-level factor requires 2 df (2 × 1 = 2). Although the test response should be chosen to minimize the occurrence of interactions, there will be times when the experts know or strongly suspect that interactions occur. In these cases, linear graphs allow the interaction to be readily included in the experiment. If more than one test is run for each test setup, the total df is the total number of tests run minus one. The dfs used for assigning factors remain the same as without the repetition. The other dfs are used to estimate the non-repeatability of the experiment.

USING ORTHOGONAL ARRAYS

AND

LINEAR GRAPHS

In an orthogonal array, the number of rows corresponds to the number of tests to be run and, in fact, each row describes a test setup. The factors to be investigated are each assigned to a column of the array. The value that appears in that column for a particular test (row) tells to what level that factor should be set for that test. As an example, consider an L4 test setup — Table 9.9. If factor A was assigned to column 1 and factor B was assigned to column 2, then test number 3 would be set up with A at level 2 and B at level 1. The sum of the degrees of freedom required for each column (a two-level column requires 1 df; a three-level column requires 2 df) equals the sum of the available dfs in the setup. Another property of the arrays is that orthogonally is maintained among the columns. Orthogonally, mentioned earlier, is the property that allows each level of every factor to equally impact the average response at each level of all other factors. Using the L4 as an example, for the test where column 1 (factor A) is at level one, column 2 (factor B) is tested at the low level and at the high level an equal number of times. This is also column 1 at level 2. In fact, orthogonality is maintained for all three columns. The reader is invited to study the L4 and verify this statement.

SL3151Ch09Frame Page 384 Thursday, September 12, 2002 6:05 PM

384

Six Sigma and Beyond

TABLE 9.9 L4 Setup Column

1

Row

1

2

3

1 2 3 4

1 1 2 2

1 2 1 2

1 2 2 1

3

2

FIGURE 9.9 Linear graph for L4.

Generally, near the orthogonal array are line-and-dot figures that look a little like “stick” drawings. These are linear graphs. The dots represent the factors that can be assigned to the orthogonal array, and the lines represent the possible interaction of the two dots joined by the line. The numbers next to the dots and lines correspond to the column numbers in the orthogonal array. For example, the linear graph for the L4 is shown in Figure 9.9. The interpretation of this linear graph is that if a factor is assigned to column 1 and a factor is assigned to column 2, column 3 can be used to evaluate their interaction. If the interaction is not suspected of influencing the response, another factor can be assigned to column 3. If no other factor remains, column 3 is left unassigned and becomes an estimator of experimental error or non-repeatability. This will be explained in more detail later in this chapter. The interrelationships between the columns are such that there are many ways of writing the linear graphs.

COLUMN INTERACTION (TRIANGULAR) TABLE Also shown in Volume V near the orthogonal array is the column interaction table for that particular array. This table shows in which column(s) the interaction would be located for every combination of two columns. The linear graphs have been constructed using this information. The L8 column interaction table is shown in Table 9.10. The interaction between two factors can be assigned by finding the intersection in the column interaction table of the orthogonal array columns to which those factors have been assigned. As an example, suppose that a factor was assigned to column 3 and another factor was assigned to column 5. If the brainstorming group suspects that the interaction of these two factors is a significant influence and includes that interaction in the analysis, that interaction must be assigned to column 6 in the orthogonal array. (Note that the interaction of two two-level factors [one degree of freedom each] can be assigned to one column which as one degree of freedom [1 × 1 = 1]).

SL3151Ch09Frame Page 385 Thursday, September 12, 2002 6:05 PM

Design of Experiments

385

TABLE 9.10 The L8 Interaction Table Column Column 1 2 3 4 5 6

FACTORS

WITH

1

2

3

4

5

6

7

3

2 1

5 6 7

4 7 6 1

7 4 5 2 3

6 5 4 3 2 1

THREE LEVELS

The orthogonal arrays, linear graphs, and column interaction tables for factors with three levels are similar to the two-level situation. Since a three-level factor requires two degrees of freedom, the three-level orthogonal array columns use two of the available dfs. The interaction of two three-level factors requires 4 dfs (2 × 2). In the linear graphs and column interaction table, and interaction is shown with two column numbers. If an interaction is being investigated, it must be assigned to two columns. The L9 orthogonal array, linear graph, and column interaction table are presented in Figure 9.10.

INTERACTIONS

AND

HARDWARE TEST SETUP

The orthogonal array specifies the hardware setup for each test. To set up the hardware for a particular test in the orthogonal array, the experimenter should disregard the interaction columns and use only the columns assigned to single factors. If an interaction is included in the experiment, its level will be based solely upon the levels of the interacting factors. The interaction will come into consideration during the analysis of the data. An example will demonstrate the use of the linear graph and the layout of a simple experiment. EXAMPLE A brainstorming group has constructed a cause-and-effect diagram and determined that four factors (A through D) are suspected of being contributors to the problem. In addition, two interactions are suspected (B × D and C × D). The group has decided to use two levels for each factor. The experiment is laid out as follows: 1. Determine the df requirement. Four dfs are required for the main factors (one for each two level factor). Two dfs are required for the interactions (one for each interaction of two level factors). Six dfs are required in total.

SL3151Ch09Frame Page 386 Thursday, September 12, 2002 6:05 PM

386

Six Sigma and Beyond

Orthogonal Array Row 1 2 3 4 5 6 7 8 9

1 1 1 1 2 2 2 3 3 3

Linear Graph

Column 2 3 1 1 2 2 3 3 1 2 2 3 3 1 1 3 2 1 3 2

1 4 1 2 3 3 1 2 2 3 1

3, 4

2

Column Interaction Table Column 1 2 3

1

Column 2 3 3 4

2 4 1 4

4 2 3 4 3 1 2

1. A******

A

D

B

C 2. B*********** 3. 3 1

5 6

2 4 7

FIGURE 9.10 The orthogonal array (OA), linear graph (LG), and column interaction for L9. 2. Determine a likely orthogonal array. Since 6 dfs + 1 = 7 tests minimum and all factors have two levels, the L8 array is a likely place to start. 3. Draw the linear graph required for the experiment. The linear graph required for the experiment is shown in Figure 9.10A. 4. Compare the linear graph(s) of the orthogonal array to the linear graph required for the experiment. One of the linear graphs for the L8 that could fit is shown in Figure 9.10B. 5. Assign factors to the orthogonal array columns. Make the column assignments shown in Figure 9.10C.

SL3151Ch09Frame Page 387 Thursday, September 12, 2002 6:05 PM

Design of Experiments

387

C********** Column

1

2

3

4

5

6

7

Factor

D

B

BxD

C

CxD

A

unassigned

where, B x D indicates the interaction between B and D.

D***** 1 7 2

5

3

4 6

E***** Column 1 2 Four Level Factor 1 1 1 1

2

2

2

1

3

2

2

4

F***** Test Number 1 2 3 4 5 6 7 8

1 0 0 0 0 0 0 0 0

2 0 0 0 0 0 0 0 0

3 1 1 2 2 3 3 4 4

Columns 4 1 2 1 2 1 2 1 2

5 1 2 1 2 2 1 2 1

6 1 2 2 1 1 2 2 1

7 1 2 2 1 2 1 1 2

FIGURE 9.10 (continued)

CHOICE

OF THE

TEST ARRAY

For a particular experiment, the test response should be chosen to minimize interaction, and the smallest orthogonal array that fits the situation should be used. The emphasis should be on assigning factors to as many columns as possible. This allows the question posed by the situation to be answered using a minimum number of tests.

SL3151Ch09Frame Page 388 Thursday, September 12, 2002 6:05 PM

388

Six Sigma and Beyond

G********* 1

3

5 7

2

6

4

H***** Column 2

1

1 2 3 4 5 6 7 8

1 2 1 2 1 2 1 2

1 1 2 2 1 1 2 2

1 1 1 1 2 2 2 2

Eight Level Factor

4

I ********

2

3, 4 5 6, 7 1 9, 10 8 12, 13

11 FIGURE 9.10 (continued)

SL3151Ch09Frame Page 389 Thursday, September 12, 2002 6:05 PM

Design of Experiments

389

Whether an interaction exists or not is an important issue that must be addressed in setting up the experiment. If an interaction does exist and provision is not made for it in the experimental setup, its effect becomes “mixed up” or confounded with the effect of the factor assigned to the column where the interaction would be assigned. The analysis will not be able to separate the two. This is an important reason why confirmatory runs are necessary. Confirmatory runs should be made with the nonsignificant factors set to their different levels, just to make sure. Another way to minimize the effect of interactions is to use an L12, L18, or L36 orthogonal array. These arrays have a special property that some, or all, of the interactions between columns are spread across all columns more or less equally instead of being concentrated in a column. This property can be used by the experimenter to rank the contribution of factors without worrying about interactions. There are times when this can be a valuable tool for the experimenter. The linear graphs for those arrays tell which interactions can be estimated and which cannot.

FACTORS

WITH

FOUR LEVELS

A factor with four levels can easily be assigned to a two-level orthogonal array. A four-level factor requires 3 dfs. Since a two-level column has 1 df, three two-level columns are used for the four-level factor. The three columns chosen must be represented in the linear graph by two dots and the connecting interaction line. One of the L8 linear graphs is shown in Figure 9.10D. The line enclosing the column 1, 2, 3 designators indicates that these columns will be used for a four-level factor. The particular level of the four-level factor for each run can be determined by taking any two of the three columns that are to be combined and assigning the four combinations to the four levels of the factor. As an example, consider columns 1 and 2 (see Figure 9.10E). Although column 3 is not used in determining the level of the four-level factor, its df is used and no other factor can be assigned to it. In the orthogonal array, one of the columns used for the four-level factor is set to the levels of the four-level factor and the other two columns are set to zero for each test. For the L8 example, the modified array would be Figure 9.10F.

FACTORS

WITH

EIGHT LEVELS

In a similar manner, a factor with eight levels requires 7 dfs and takes up seven twolevel columns. The particular columns are chosen by taking a closed triangle in the linear graph and the interactions column of one of the points of the triangle with the opposite base. One example is shown in Figure 9.10G. The column interaction table indicates that the interaction of columns 1 and 6 will be in column 7. The actual factor level for each test is determined by looking at the combinations of the three columns that make up the corners of the triangle (see Figure 9.10H). None of the seven columns which are used for the eight-level factor can be assigned to another factor. In the orthogonal array, one of the columns used for the eight-level factor is set to the levels of the eight-level factor and the other six columns are set to zero for each test.

SL3151Ch09Frame Page 390 Thursday, September 12, 2002 6:05 PM

390

Six Sigma and Beyond

TABLE 9.11 An L9 with a Two-Level Column

FACTORS

WITH

Columns

Test Number

1

2

3

4

1 2 3 4 5 6 7 8 9

1 1 1 2 2 2 1 1 1

1 2 3 1 2 3 1 2 3

1 2 3 2 3 1 3 1 2

1 2 3 3 1 2 2 3 1

NINE LEVELS

A factor with nine levels is handled in a similar manner to a four-level factor. The nine-level factor requires 8 dfs, which are available in four three-level columns. Two three-level columns and their two interaction columns are used. One of the L27 linear graphs is shown in Figure 9.10I. The line enclosing the column 1, 2, 3, 4 designators indicates that these four columns will be used for the nine-level factor. The level of the nine-level factor to be used in a particular test can be determined by taking any two of the four columns that are to be combined and assigning their nine combinations to the nine levels of the factor. This is left to the reader to demonstrate. In the orthogonal array, one of the columns used for the nine-level factor is set to the levels of the nine-level factor and the other three columns are set to zero.

USING FACTORS

WITH

TWO LEVELS

IN A

THREE-LEVEL ARRAY

Dummy Treatment Often, the situation calls for a mix of factors with two and three levels. A two-level factor can be assigned to a three-level column by using one of the two levels as the third level in the test determination. Consider using a two-level factor in an L9 array — see Table 9.11. In column 1, the second set of 1s (in experiments 7, 8, and 9) is the dummy treatment. In the analysis, the average at level one of the factor assigned to column 1 is determined with more accuracy than the average at level two since more tests are run at level one. The level that is of more interest to the experimenter should be the one used for the dummy treatment. Combination Method Two two-level factors can be assigned to a single three-level column. This is done by assigning three of the four combinations of the two two-level factors to the three-level

SL3151Ch09Frame Page 391 Thursday, September 12, 2002 6:05 PM

Design of Experiments

391

TABLE 9.12 Combination Method Factor A

Factor B

Three-Level Column

1 1 2

1 2 1

1 2 3

factor and not testing the fourth combination. As an example, two two-level factors are assigned to a three-level column as in Table 9.12. Note that the combination A2B2 is not tested. In this approach, information about the AB interaction is not available, and many ANOVA (analysis of variance) computer programs are not able to break apart the effect of A and B. A way of doing that manually will be presented later.

USING FACTORS

WITH

THREE LEVELS

IN A

TWO-LEVEL ARRAY

A factor with three levels requires 2 dfs. Although it would seem that two two-level columns combined would give the required dfs, the interaction of those two columns is confounded with the three-level factor. The approach used to assign one threelevel factor to a two-level array is to construct a four-level column and use the dummy treatment approach to assign the three-level factor to the four-level column. Assigning more than one three-level factor to a two-level array uses a variation of this approach. Recall that in constructing a four-level column, three two-level columns are used. These three must be shown in the linear graph as two dots connected by an interaction line. Any two of these columns are used to determine the level to be tested. The third column’s df is used up in assigning a four-level factor. In assigning a three-level factor, the third column’s df is not used for the level three factor since it require only 2 dfs. However, the third column is confounded with the three-level factor and should not be assigned to another factor. That column is said to be “idle.” When two or more three-level factors are assigned to a two-level array, the three-level factors can share the same idle column. An example of assigning two three-level factors to an L8 array is shown in Figure 9.11. Here column 1 would be idle (a factor cannot be assigned to column 1), columns 2 and 3 would be used to determine the levels of a three-level factor, columns 4 and 5 would be used to determine the levels of the second three-level factor, and columns 6 and 7 are available for two-level factors. The modified orthogonal array for this experiment is shown in Table 9.13 (level 2 is the dummy treatment in both cases). The idle column approach cannot be used with four-level factors. If it were attempted, insufficient degrees of freedom would exist and the four-level factors would be confounded.

OTHER TECHNIQUES There are other techniques for setting up an experiment that will be mentioned here but will not be discussed in detail. The user is invited to read the chapter on pseudofactor design in Quality Engineering — Product and Process Design Optimization,

SL3151Ch09Frame Page 392 Thursday, September 12, 2002 6:05 PM

392

Six Sigma and Beyond 1 (idle) 7

3

5

2

4

6

FIGURE 9.11 Three-level factors in a L8 array.

TABLE 9.13 Modified L8 Array Columns

Test Number

1

2

3

4

5

6

7

1 2 3 4 5 6 7 8

1 1 1 1 2 2 2 2

0 0 0 0 0 0 0 0

1 1 2 2 3 3 2 2

0 0 0 0 0 0 0 0

1 2 1 2 2 3 2 3

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

by Yuin Wu and Dr. Willie Hobbs Moore or to consult with a statistician to use these techniques. Nesting of Factors Occasionally, levels of one factor have meaning only at a particular level of another factor. Consider the comparison of two types of machine. One is electrically operated and the other is hydraulically operated. The voltage and frequency of the electrical power source and the temperature and formulation of the hydraulic fluid are factors that have meaning for one type of machine but not the other. These factors are nested within the machine level and require a special setup and analysis which is discussed in the reference given above. Setting Up Experiments with Factors with Large Numbers of Levels Experiments with factors with large numbers of levels can be assigned to an experimental layout using combinations of the techniques that have been covered in this booklet.

SL3151Ch09Frame Page 393 Thursday, September 12, 2002 6:05 PM

Design of Experiments

393

TABLE 9.14 An L8 with an L4 Outer Array L8

Test No.

1

2

3

4

5

6

7

1 2 3 4 5 6 7 8

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

L4

1

2

2

1

(on side) →

1 1

2 1

1 2

2 2

Test Results

x1 x5

x2 x6

x3 x7

x4 x8

. . .

. . .

. . .

. . .

x29

x30

x31

x32

Note: The x values refer to experimental test results.

INNER ARRAYS

AND

OUTER ARRAYS

Factors are generally divided into three basic types: 1. Control factors are the factors that are to be optimized to attain the experimental goal. 2. Noise factors represent the uncontrollable elements of the system. The optimum choice of control factor levels should be robust over the noise factor levels. 3. Signal factors represent different inputs into the system for which system response should be different. For example, if several micrometers were to be compared, the standard thickness to be measured would be levels of a signal factor. The optimum micrometer choice would be the one that operated best at all the standard thicknesses. Signal factors are discussed in more detail on pages 430–441. Control and noise factors are usually handled differently from one another in setting up an experiment. Control factors are entered into an orthogonal array called an inner array. The noise factors are entered into a separate array called an outer array. These arrays are related so that every test setup in the inner array is evaluated across every noise setup in the outer array. As an example, consider an L8 inner (control) array with an L4 outer (noise) array, as shown in Table 9.14. The purpose of this relationship is to equally and completely expose the control factor choices to the uncontrollable environment. This ensures that the optimum factor will be robust. A signal-to-noise (S/N) ratio can be calculated for each of the control factor array test situations. This allows the experimenter to identify the control factor level choices that meet the target response consistently.

SL3151Ch09Frame Page 394 Thursday, September 12, 2002 6:05 PM

394

Six Sigma and Beyond

RANDOMIZATION

OF THE

EXPERIMENTAL TESTS

In the orthogonal arrays, each test setup is identified by a test number. Generally, the tests should not be run in the order of test number. If the tests were run in that order, all the tests with the factor assigned to column one at level one would be run before any of the tests with that factor at level two. A quick glance at an orthogonal array will confirm this relationship. In fact, the columns toward the left of the array change less often than the columns toward the right of the array. If an uncontrolled noise factor changes during the testing process, the effect of that noise factor could be mixed in with one or more of the factor effects. This could result in an erroneous conclusion. The possibility of this occurring can be minimized by randomizing the order of the experiment runs. If the order of the tests is randomized, the effect of the changing uncontrolled noise factor will be more or less spread evenly over all the levels of the controlled factors and although the experimental error will be increased, the effects of the controlled factors will still be identifiable. Randomization can be done as simply as writing the test numbers on slips of paper and drawing them out of a hat. There are two situations where randomization may not be possible or where its importance is lessened. 1. If it is very expensive, difficult, or time-consuming to change the level of a factor, all tests at one level of a factor may have to be run before the level of that factor can be changed. In this case, noise factors should be chosen for the outer array that represent the possible variation in uncontrolled environment as much as possible. 2. If the noise factors in the outer array are properly chosen, the confident experimenter may elect to dispense with randomization. In most cases, the purpose of the experiment is to learn more about the situation, and the experimenter does not have complete confidence. Therefore, the test order should be randomized whenever the circumstances permit.

MISCELLANEOUS THOUGHTS Dr. Taguchi stresses evaluating as many main factors as possible and filling up the available columns. If it turns out that the experimental design will result in unassigned columns, some column assignment schemes are better than others in a few situations. The rationale behind these choices is that they minimize the confounding of unsuspected two-factor interactions with the main factors. A detailed discussion is beyond the scope of this chapter. The user is invited to read Chapter 12 of Statistics for Experiments, by G. Box, W. Hunter, and J.S. Hunter to learn more about this concept. Consider an L8 for which there are to be four two-level factors assigned. This implies that there will be three columns that will not be assigned to a main factor. There are 35 ways in which the four factors can be assigned to the seven columns. The recommended assignment is to use columns 1, 2, 4, and 7 for the main factors. The interactions to be evaluated, the linear graphs, and the column interaction table

SL3151Ch09Frame Page 395 Thursday, September 12, 2002 6:05 PM

Design of Experiments

395

TABLE 9.15 Recommended Factor Assignment by Column Number of Factors 4 5 6 7 8 9 10 11 12 13 14 15 a

L8 Array

L16 Array

1, 2, 4, 7

1, 1, 1, 1, 1,

a a a

— — — — — — — —

2, 2, 2, 2, 2,

4, 4, 4, 4, 4,

8 8, 8, 7, 7,

L32 Array

15 11, 13 8, 11, 13 8, 11, 13, 14 a a a a a a a

1, 2, 4, 8, 16 1, 2, 4, 8, 16, 31 1, 2, 4, 8, 15, 16, 23 1, 2, 4, 8, 15, 16, 23, 27 1, 2, 4, 8, 15, 16, 23, 27, 29 1, 2, 4, 8, 15, 16, 23, 27, 29, 30 1, 2, 4, 7, 8, 11, 13, 14, 16, 19, 21 1, 2, 4, 7, 8, 11, 13, 14, 16, 19, 21, 1, 2, 4, 7, 8, 11, 13, 14, 16, 19, 21, 1, 2, 4, 7, 8, 11, 13, 14, 16, 19, 21, 1, 2, 4, 7, 8, 11, 13, 14, 16, 19, 21,

22 22, 25 22, 25, 26 22, 25, 26, 28

No recommended assignment scheme.

determine if the recommended column assignments are usable for a particular experiment. The recommended column assignments are given in Table 9.15. Some of the linear graphs may be found in the Appendix of Volume V. However, the user will find that the linear graphs in other books and reference materials may not make these assignments available. There are many equally valid ways that linear graphs for the larger arrays can be constructed from the column interaction table. It is not feasible for any one book to list all the possibilities. An excellent source is Taguchi and Konishi (1987). In many cases, the brainstorming group may not have a good feel for whether interactions exist or not. In these cases, two alternatives are usually considered: 1. Design an experiment that allows all two-factor interactions to be estimated. 2. Design an experiment in which no factor is assigned to a column that also contains the interaction of two other factors, although pairs of two-factor interactions may be assigned to the same column. The recommended factor assignments given in Table 9.15 are examples of this approach. The second approach is based on the assumption that few of the interactions will be significant and that later testing can be used to investigate them in more detail. The reader is urged to seek statistical assistance in approaching this type of experiment. Sometimes, the response is not related to the input factors in a linear fashion. Testing each factor at two levels allows only a linear relationship to be defined and, in this more complex situation, can give misleading results. A detailed statistical analysis tool called response surface methodology can be used to investigate the complex relationship of the input factors to the response in these cases.

SL3151Ch09Frame Page 396 Thursday, September 12, 2002 6:05 PM

396

Six Sigma and Beyond

All of this seems to indicate that DOEs must be lengthy and complicated when interactions or nonlinear relationships are suspected. In most situations, time and resources are not available to run a large experiment. Sometimes, a transformation of the measured data or of a quantitative input factor can allow a linear model to fit within the region covered by the input factors. The linear model requires fewer data points than a curvilinear model and is easier to interpret. Unfortunately, unless multiple observations are made at each inner array setup, the choice of transformation is guided mainly by the experience of the experimenter or by trying several transformations and seeing which one fits best. The choice of the proper transformation to use is related to the choice of the proper response. As an example, two common measures of fuel usage are “miles per gallon” and “liters per kilometer.” With the multiplication of a constant, these two measures are inverses of each other. A model that is linear in mi/g will be definitely non-linear in l/km. Which measurement is correct? There is no easy answer. The experimenter should evaluate several different transformations to determine the best model. Some transformations that are useful are: y = Y1/2useful for count data (Poisson distributed) such as the number of flaws in a painted surface y = log(Y) or ln(Y)useful for comparing variances y = Y–1/2 y = 1/Y When there are several observations at each inner array test setup either through replication or through testing with and outer array, another guide to choosing the right transformation can be used. For the ANOVA to work correctly, the variances at all test points should be equal. The observed variances should be compared as follows:

( )

1. Calculate the average X and the standard deviation(s) for each inner array test setup. 2. Take the log or ln of each X and s. 3. Plot log s (y-axis) versus log X (x-axis) and estimate the slope. 4. Use the estimated slope as a rough guide to determine which transformation to use: Slope 0.0 0.5 1.0 1.5 2.0

Transformation no transformation y = Y1/2 y = log(Y) or 1n(Y) y = Y–1/2 y = 1/Y

It should be noted that the addition or subtraction of a constant before plotting will not affect the standard deviation but will affect the relative spacing of the log X and hence the slope of the line. This approach can be used to improve the fit of the transformation. With the widespread use of computers, data analysis of this type

SL3151Ch09Frame Page 397 Thursday, September 12, 2002 6:05 PM

Design of Experiments

397

Loss In $

Scrap Rework

Spec Limit

Spec Limit

FIGURE 9.12 Traditional approach.

should be easy and should be pursued as a means to get the most information out of the data. Examples of this approach will be given later in the chapter. The reader is invited to refer to Statistics for Experiments by G. Box, W. Hunter, and J.S. Hunter to learn more about the use of transformations in analyzing data.

LOSS FUNCTION AND SIGNAL-TO-NOISE This section discusses: 1. The Taguchi loss function and its cost-oriented approach to product design 2. A comparison of the loss function and the traditional approach to calculating loss 3. The use of the loss function in evaluating alternative actions 4. A comparison of the loss function and Cpk and the appropriate use of each 5. The relationship of the loss function and the signal-to-noise (S/N) calculation that Dr. Taguchi uses in design of experiments

LOSS FUNCTION

AND THE

TRADITIONAL APPROACH

In the traditional approach — see Figure 9.12 — to considering company loss, parts produced within the spec limits perform equally well, and parts outside of the spec limits are equally bad. This approach has a fallacy in that it assumes that parts produced at the target and parts just inside the spec limit perform the same and that parts just inside and just outside the spec limits perform differently. Statistical Process Control (SPC) and process capability calculations (Cpk) have brought to the manufacturing floor an awareness of the importance of reducing process variability and centering around the target. However, the question still remains, “How can this thought process carry over into product and process decision?” The loss function provides a way of considering customer satisfaction in a quantitative manner during the development of a product and its manufacturing process. The loss function is the cornerstone of the Taguchi philosophy. The basic premise of the loss function is that there is a particular target value for each critical

SL3151Ch09Frame Page 398 Thursday, September 12, 2002 6:05 PM

398

Six Sigma and Beyond

characteristic that will best satisfy all customer requirements. Parts or systems that are produced farther away from the target will not satisfy the customer as well. The level of satisfaction decreases as the distance from the target increases. The loss function approximates the total cost to society, including customer dissatisfaction, of producing a part at a particular characteristic value. Taken for a whole production run, the total cost to society is based on the variability of the process and the distance of the distribution mean to the target. Decisions that affect process variability and centering or the range over which the customer will be satisfied can be evaluated using the common measurement of loss to society. The loss function can be used when considering the expenditure of resources. Customer dissatisfaction is very difficult to quantify and is often ignored in the traditional approach. Its inclusion in the decision process via the loss function highlights a gold mine in customer-perceived quality and repeat purchases that would be hidden otherwise. This gold mine is often available at a relatively minor expense applied to improving the product or process. Note: Use of the loss function implies a total system that starts with the determination of targets that reflect the greatest level of customer satisfaction. Calculation of losses using nominals that were set using other methods may yield erroneous results.

CALCULATION

OF THE

LOSS FUNCTION

Dr. Taguchi uses a quadratic equation to describe the loss function. A quadratic form was chosen because: 1. It is the simplest equation that fulfills the requirement of increasing as it moves away from the target. 2. Taguchi believes that, historically, costs behave in this fashion. 3. The quadratic form allows direct conversion from signal-to-noise ratios and decomposition used in analysis of experimental results. The general form for the loss function is:

(

L( x) = k x − m

)

2

where L(x) is the loss associated with producing a part at “x” value; k is a unique constant determined for each situation; x is the measured value of the characteristic; and m is the target of the characteristic. When the general form is extended to a production of “n” items, the average loss is:

( )∑ ( x − m)

L= k/n

2

SL3151Ch09Frame Page 399 Thursday, September 12, 2002 6:05 PM

Design of Experiments

399

Ao Cost

m–∆

m+∆

m

FIGURE 9.13 Nominal the best.

This can be simplified to:

(

)

2  L = k σ2 + µ − m   

where σ 2 the population piece-to-piece variance; µ is the population mean; and (µ – m) is the offset of the population mean from the target. In the Nominal-the-Best (NTB) situation shown in Figure 9.13, A0 is the cost incurred in the field by the customer or warranty when a part is produced ∆ from the target. ∆ is the point at which 50% of the customers would have the part repaired or replaced. A0 and ∆ define the shape of the loss function and the value of “k.” The loss resulting from producing a part at m – ∆ is:

(

) (

L m−∆ =k m−∆−m

)

2

A0 = k∆2

k = A0 / ∆2 In general, the loss per piece is:

()

(

L x = A0 / ∆2 * x − m

)

2

The loss for the population is:

(

L = A0 / ∆2 * σ2 + offset 2

)

EXAMPLE A particular component is manufactured at an internal supplier, shipped to an assembly plant, and assembled into a vehicle. If this component deviates from its target

SL3151Ch09Frame Page 400 Thursday, September 12, 2002 6:05 PM

400

Six Sigma and Beyond of 300 units by 10 or more, the average customer will complain, and the estimated warranty cost will be $150.00. In this case,

k = $150.00/(10 units)2 = $1.50 per unit2 SPC records indicate that the process average is 295 units and the variability is eight units2. The present total loss is:

(

)

2  L = k σ2 + µ − 300   

(

)

2  = $1.50 82 + 295 − 300   

= $133.50 per part Fifty thousand parts are produced per year. The total yearly loss (and opportunity for improvement) is $6.7 million.

Situation 1 It is estimated that a redesign of the system would make the system more robust, and the average customer would complain if the component deviated by 15 units or more from 300. In this case:

(

k = $150 / 15 units

)

2

= $0.67 per unit 2 The total loss would be:

(

)

2  L = $0.6782 + 295 − 300   

= $59.63 per part The net yearly improvement due to redesigning the system would be:

(

)

Improvement = $1.33.50 − $59.63 * 5000 = $3, 693, 500 This cost should be balanced against the cost of the redesign.

SL3151Ch09Frame Page 401 Thursday, September 12, 2002 6:05 PM

Design of Experiments

401

Situation 2 It is estimated that a new machine at the component manufacturing plant would improve the mean of the distribution to 297 units and improve the process variability to 6 units2. In this case, the total loss would be:

(

)

2  L = $1.50 62 + 297 − 300   

= $67.50 per part The net yearly improvement due to using the new machine would be:

(

)

Improvement = $1.33.50 − $67.50 * 50, 000 = $3, 300, 000 This cost should be balanced against the cost of the new machine.

From these situations, it is apparent that the quality of decisions using the loss function is heavily dependent upon the quality of the data that goes into the loss function. The loss function emphasizes making a decision based on quantitative total cost data. In the traditional approach, decisions are difficult because of the unknowns and differing subjective interpretations. The loss function approach requires investigation to remove some of the unknowns. Subjective interpretations become numeric assumptions and analyses, which are easier to discuss and can be shown to be based on facts. In the smaller-the-better (STB) situation illustrated in Figure 9.14, the loss function reduces to: L = k [1/n ∑ x 2]

A0 Cost

X0 FIGURE 9.14 Smaller the better.

SL3151Ch09Frame Page 402 Thursday, September 12, 2002 6:05 PM

402

Six Sigma and Beyond

For the larger-the-better (LTB) situation illustrated in Figure 9.15, the loss function reduces to: L = k [1/n ∑1/x 2]

FIGURE 9.15 Larger the better.

COMPARISON

OF THE

LOSS FUNCTION

AND

Cpk

The loss function can be used to evaluate process performance. It provides an emphasis on both reducing variability and centering the process, since those actions have a net effect of reducing the value of the loss function. Process performance is normally evaluated using Cpk. Cpk is calculated using the following equation:  upper spec limit − X C pk = minimum   3 * standard deviation 

(

)

,

X − lower spec limit   3 * standard deviation 

(

)

where X = the average of the process. Both the loss function and Cpk emphasize minimizing the variability and centering the process on the target. The relative benefits of the two can be summarized as follows: Loss function • Provides more emphasis on the target • Relates to customer costs • Can be used to prioritize the effect of different processes Cpk • Is easier to understand and use • Is based only on data from the process and specifications • Is normalized for all processes The loss function represents the type of thinking that must go into making strategic management decisions regarding the product and process for critical characteristics. Cpk is an easily used tool for monitoring actual production processes.

SL3151Ch09Frame Page 403 Thursday, September 12, 2002 6:05 PM

Design of Experiments

403

-16----20----24-

-16----20----24-

-16----20----24-

-16----20----24-

-16----20----24-

Case 1

Case 2

Case 3

Case 4

Case 5

Case 1

Case 2

Case 3

Case 5

Case 4

Average Sigma

20 1.33

18 0.67

17.2 0.4

20 2.82

20 0.67

C pk Loss (assume k = 2)

1

1

1

0.47 16

2

3.56

8.89

16

0.89

FIGURE 9.16 A comparison of Cpk and loss function.

Figure 9.16 shows Cpk and the value of the loss function for five different cases. In each of these cases, the specification is 20 ± 4 and the value of k in the loss function is $2 per unit2. Both Cpk and the loss function emphasize reducing the part-to-part variability and centering the process on target. The use of Cpk is recommended in production areas to monitor process performance because of the ease of understanding the clear relationship of Cpk and the other SPC tools. Management decisions regarding the location of distributions with small variability within a large specification tolerance should be based on a loss function approach. (See cases 2 and 5 in Figure 9.16.) The loss function approach should be used to determine the target value and to evaluate the relative merits of two or more courses of action because of the emphasis on cost and on including customer satisfaction as a factor in making basic product and process decisions. These questions also lend themselves to the use of design of experiments. The relationship of the loss function to the signal-to-noise DOE calculations used by Dr. Taguchi will now be discussed.

SIGNAL-TO-NOISE (S/N) Signal-to-Noise is a calculated value that Dr. Taguchi recommends to analyze DOE results. It incorporates both the average response and the variability of the data. S/N is a measure of the signal strength to the strength of the noise (variability). The goal is always to maximize the S/N. S/N ratios are so constructed that if the average response is far from the target, re-centering the response has a greater effect on the S/N than reducing the variability. When the average response is close to the target, reducing the variability has a greater effect. There are three basic formulas used for calculating S/N, as shown in Table 9.16. S/N for a particular testing condition is calculated by considering all the data that were run at that particular condition across all noise factors. Actual analysis techniques will be covered later.

SL3151Ch09Frame Page 404 Thursday, September 12, 2002 6:05 PM

404

Six Sigma and Beyond

TABLE 9.16 Formulas for Calculating S/N Signal-to-Noise (S/N) Smaller the better (STB)

[ ∑x ] −10 log [1 / n∑1 / x ] −10 log [1 / n( S − V ) / V ]

−10 log10 1 / n

2

2

Larger the better (LTB)

10

Nominal the best (NTB)

where

Loss Function (L)

10

m

o

o

[ ∑x ] L = k [1 / n∑1 / x ] L = k 1/ n

2

2

(

L = k σ 2 + offset 2

)

(∑ x ) / n V = (∑ x − S ) / ( n − 1) 2

Sm =

2

o

m

The relationships between S/N and loss function are obvious for STB and LTB. The expressions contained in brackets are the same. When S/N is maximized, the loss function will be minimized. For the NTB situation, the total analysis procedure of looking at both the raw data for location effects and S/N data for dispersion effects parallels the loss function approach. Examples of these analysis techniques are given in the next section. S/N is used in DOE rather than the loss function because it is more understandable from an engineering standpoint and because it is not necessary to compute the value of k when comparing two alternate courses of action. S/N calculations are also used in DOE to search for “robust” factor values. These are values around which production variability has the least effect on the response.

MISCELLANEOUS THOUGHTS Many statisticians disagree with the use of the previously defined S/N ratios to analyze DOE data. They do not recognize the need to analyze both location effects and dispersion (variance) effects but use other measures. Dr. George Box’s 1987 report is recommended to the reader who wishes to learn more about this disagreement and some of the other methods that are available. In brief, Dr. Box disagrees with the STB and LTB S/N calculations and finds the NTB S/N to be inefficient. The approach that he supports is to calculate the log (or ln) of the standard deviation of the data, log(s), at each inner array setup in place of the S/N ratio. The log is used because the standard deviation tends to be lognormally distributed. The raw data should be analyzed (with appropriate transformations) to determine which factors control the average of the response, and the log(s) should be analyzed to determine which factors control the variance of the response. From these two analyses, the experimenter can choose the combination of factors that gives the response that best fills the requirements. The data in Table 9.17 illustrate some of the concerns with the NTB S/N ratio. The first three tests (A through C) have the same standard deviation but very different S/N,

SL3151Ch09Frame Page 405 Thursday, September 12, 2002 6:05 PM

Design of Experiments

405

TABLE 9.17 Concerns with NTB S/N Ratio Raw Data (4 Reps.)

Test A B C D E

1 15 18 24 42.55

2 11 21 24 42.8

4 12 19 28.12 50

5 14 22 28.12 50

Standard Deviation

NTB S/N

1.83 1.83 1.83 2.38 4.23

3.89 17.03 20.78 20.78 20.78

while the last three tests (C through E) have the same S/N but very different standard deviations. The NTB S/N ratio places emphasis on getting a higher response value. This approach might lead to difficulties in tuning the response to a specific target. It should be noted that Taguchi does discuss other S/N measures in some of his works that have not been widely available in English. An alternate NTB S/N ratio is available in the computer program ANOVA-TM, which is distributed by Advanced Systems and Designs, Inc. (ASD) of Farmington Hills, Michigan and is based on Taguchi’s approach. This S/N ratio is: NTBπ S / N = −10 log( s 2 ) = −20 log( s )

Maximizing this S/N is equivalent to minimizing log(s). Examples using this S/N ratio will be developed later.

ANALYSIS The purpose of this section is to: 1. Introduce graphical and numerical analysis of experimental data 2. Present a method for estimating a response value and assigning a confidence interval for it 3. Discuss the use and interpretation of signal-to-noise (S/N) ratio calculations

GRAPHICAL ANALYSIS In the example in Section 2, Timothy and Christine calculated and plotted the average response at each factor level. Since the experimental design they used (an L8) is orthogonal, the average at each level of a factor is equally impacted by the effect of the levels of the other factors. This allows the graphical approach to have direct usage. This example from section 2 is shown in Table 9.18. The factor level plots are shown in Figure 9.17. Factors C1, C2 and C11 clearly have a different response for each of their two levels. The difference between levels is much smaller for the other factors. If the

SL3151Ch09Frame Page 406 Thursday, September 12, 2002 6:05 PM

406

Six Sigma and Beyond

TABLE 9.18 L8 with Test Results Levels for Each Suspected Factor for Each of 8 Tests

Test Number

C1

C2

C7

C11

C13

C15

C16

Test Result

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

10 13 15 17 14 16 19 21

1 2 3 4 5 6 7 8

Note: The C numbers (e.g., C11, C13) are factor names.

C2

–

–

1 2 LEVEL

C16

1 2 LEVEL

–

–

– – – – – – –

–

–

C15

– – – – – – 1 2 LEVEL

1 2 LEVEL

–

– – – – – –

–

–

–

C13 18 17 16 15 14 13

–

R E S P O N S E

– – – – – –

– – – – – – 1 2 LEVEL

1 2 LEVEL

C11

C7

– – – – – –

– – – – – –

–

C1 18 17 16 15 14 13

–

R E S P O N S E

1 2 LEVEL

FIGURE 9.17 Plots of averages (higher responses are better).

goal of the experiment was to identify situations that minimize or maximize the response, C1, C2 and C11 are important while the others are not. Graphical analysis is a valid, powerful technique that is especially useful in the following situations: 1. When computer analysis programs are not available 2. When a quick picture of the experimental results is desired 3. As a visual aid in conjunction with computer analysis

SL3151Ch09Frame Page 407 Thursday, September 12, 2002 6:05 PM

Design of Experiments

407

TABLE 9.19 ANOVA Table Column

Source

1 2 3 4 5 6 7 Error (pooled error) Total

C1 C2 C7 C11 C13 C15 C16

df

SS

MS

F Ratio

S’

%

1 1 1* 1 1* 1* 1*

28.125 45.125 0.125 10.125 0.125 0.125 0.125

28.125 45.125 0.125 10.125 0.125 0.125 0.125

225 361

28.000 45.000

33.38 53.65

81

10.000

11.92

4 7

0.500 83.875

0.125 11.982

0.875 83.875

1.04

Note: df = degrees-of-freedom; MS = mean square; SS = sum of squares.

Once the experiment has been set up correctly, the graphical analysis can be easily used and can point the way to improvements.

ANALYSIS

OF

VARIANCE (ANOVA)

As was mentioned earlier, mathematical calculations and detailed discussions will not be included in this chapter. The interested reader should consult Volume V of this series or references listed in the Bibliography for rigorous mathematical discussions. The approach given here will focus on the interpretation of the ANOVA analysis. ANOVA is a matrix analysis procedure that partitions the total variation measured in a set of data. These partitions are the portions that are due to the difference in response between the levels of each factor. The number of degrees of freedom (df) associated with an experimental setup is also the maximum number of partitions that can be made. Consider the L8 experiment from section 2 that was illustrated previously in the graphical analysis section. Table 9.19, which is an ANOVA table, summarizes the analysis. The column number shows to what column of the orthogonal array the source (factor) was assigned. Normally, the column number is not shown in an ANOVA table. The df column shows the df(s) associated with the factor in the source column. The SS column contains the sums of squares. The SS is a measure of the spread of the data due to that factor. The total SS is the sum of the SS due to all of the sources. The MS or mean square column shows the SS/df for each source. The MS is also known as the variance. The row with “error” in the source column is left blank in this experiment. If one of the columns had not been assigned or if the experiment had been replicated, then the unassigned dfs would have been used to estimate error. Error is the nonrepeatability of the experiment with everything held as constant as possible. The ANOVA technique compares the variability contribution of each factor to the variability

SL3151Ch09Frame Page 408 Thursday, September 12, 2002 6:05 PM

408

Six Sigma and Beyond

due to error. Factors that do not demonstrate much difference in response over the levels tested have a variability that is not much different from the error estimate. The df and SS from these factors are pooled into the error term. Pooling is done by adding the df and SS into the error df and SS. Pooling the insignificant factors into the error can provide a better estimate of the error. Initially, no estimate of error was made in the L8 example because no unassigned columns or repetitions were present. Because of this, a true estimate of the error could not be made. However, the purpose of the experiment was to identify the factors that have a usable difference in response between the levels. In this experiment, the factors with relatively small MS were pooled and called “error.” Pooling requires that the experimenter judge which differences are significant from an operational standpoint. This judgment is based on the prior knowledge of the system being studied. In the example, factors C7, C13, C15, and C16 have much lower MS than do the other factors and are pooled to construct an error estimate. The * next to a df indicates that the df and SS for that factor were pooled into the error term. The F ratio column contains the ratio of the MS for a source to the MS for the pooled error. This ratio is used to statistically test whether the variance due to that factor is significantly different from the error variance. As a quick rule of thumb, if the F ratio is greater than three, the experimenter should suspect that there is a significant difference. Dr. Taguchi does not emphasize the use of the F ratio statistical test in his approach to DOE. A detailed description of the use of the F test can be found in Box, Hunter, and Hunter (1978), and a practical explanation is included in Volume V of this series. In the determination of the SS of a factor, the non-repeatability of the experiment is still present. The number in the “S” column is an attempt to totally remove the SS due to error and leave the “pure” SS that is due only to the source factor. The error MS times the df is subtracted from the SS to leave the pure SS or S′ value for a factor. The amount that is subtracted from each non-pooled factor is then added to the pooled error SS and the total is entered as the error S′. In this way the total SS remains the same. The % column contains the S′ value divided by the total SS times 100%. This gives the percent contribution by that factor to the total variation of the data. This information can be used directly in prioritizing the factors. In the experiment that has been discussed, C2 makes the greatest contribution, C1 contributes less, and C11 contributes still less. It can be argued that the graphical analysis can display those conclusions quite well. In more complicated experiments with many factors and factors with a large number of levels, however, the ANOVA table can display the analysis in a more concise form and quickly lead the experimenter to the most important factors.

ESTIMATION

AT THE

OPTIMUM LEVEL

The ANOVA table is used to identify important factors. The experimenter refers to the average response at each level of the important factors to choose the best combination of factor levels. All of the best levels can be combined to estimate the responses at the best factor combination. Consider the case where the second level

SL3151Ch09Frame Page 409 Thursday, September 12, 2002 6:05 PM

Design of Experiments

409

of factor A (A2), the third level of factor B (B3), the first level of factor C (C1), and the interaction of C1 and D1 are determined to be the best combination of factors. An estimate of the response at these conditions can be made using the equation:

(

) [(

) (

) (

) (

µˆ opt = T + ( A2 − T ) + B3 − T + C1 − T + C1D1 − T − C1 − T − D1 − T

)]

where T = the average response of all the data; A2 = the average of the data run at A2; B3 = the average of the data run at B3; C1 = the average of the data run at C1; and D1 = the average of the data run at D1. Each factor that is a significant contributor appears in a manner similar to A2, B3 and C1 above. The term in brackets [ ] addresses the optimum level of the CD interaction and is an example of the way in which interactions are handled.

CONFIDENCE INTERVAL

AROUND THE

ESTIMATION

A 90% confidence interval can be calculated for confirmatory tests using the equation:

(

)

(

µˆ opt ± F1,dfe,.05 * MSe * 1 / ne + 1 / nr

)

where F1,dfe,.05 is a value from an F statistical table. The F values are based on two different degrees of freedom and the desired confidence. In this case, the first degree of freedom is always 1 and the second is the degree of freedom of the pooled error (dfe). The desired confidence is .05 since .05 in each direction (±) sums to a 10% confidence. MSe is the mean square of the pooled error term; nr is the number of confirmatory tests to be run; and ne is the effective number of replications and is calculated as follows: ne =

Total Number of Experiments Sum of the dfs of all the factors and interactions that are significant and appear in the equation plus 1 df for the mean.

For the µˆ opt that was just considered, ne is calculated as follows: Source

df

A B C CD Mean Total

1 2 1 1 1 6

SL3151Ch09Frame Page 410 Thursday, September 12, 2002 6:05 PM

410

Six Sigma and Beyond

Quadratic Only Response

Response

Response

Linear Only

1 2 3 Factor Level

1 2 3 Factor Level

Both Linear and Quadratic

1 2 3 Factor Level

Response

FIGURE 9.18 ANOVA decomposition of multi-level factors.

1

2

3

Supplier FIGURE 9.19 Factors not linear.

Consider that an L36 was run with no repetitions. ne = 36/6 = 6.0

INTERPRETATION

AND

USE

The confidence interval about the estimated value is used as a check when verification runs are made. If the average of the verification runs does not fall within the interval, there is strong reason to believe that a very important factor may have been left out of the experiment.

ANOVA DECOMPOSITION

OF

MULTI-LEVEL FACTORS

When a factor is tested at two levels, an estimate of the linear change in response between the two levels can be made. When a factor is tested at more than two levels, more complex relationships must be investigated. With a three-level factor, both the linear and quadratic relationships can be investigated. These relationships are demonstrated in Figure 9.18. This relationship is important to consider even when the factor levels are not continuous (e.g., different machines or suppliers). Consider the situation in Figure 9.19. The dotted line is the linear response and indicates no significant difference. However, Supplier 2 is different from Suppliers 1 and 3. This difference can be found only if the quadratic relationship is considered. The number of higher order relationships that can be investigated is determined by the degrees of freedom of the source — see Table 9.20.

SL3151Ch09Frame Page 411 Thursday, September 12, 2002 6:05 PM

Design of Experiments

411

TABLE 9.20 Higher Order Relationships Levels of a Factor

df

Relationships

2 3 4 5 etc.

1 2 3 4

Linear Linear, quadratic Linear, quadratic, cubic Linear, quadratic, cubic, quartic

TABLE 9.21 Inner OA (L8) with Outer OA (L4) and Test Results L8 Test No.

A 1

B 2

C 3

D 4

E 5

F 6

G 7

1 2 3 4 5 6 7 8

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

L4 (on side)

Test Results

Z Y X

1 1 1

2 2 1

2 1 2

1 2 2

25 25 18 26 15 18 20 19

27 27 21 23 11 15 17 20

30 21 19 27 12 17 21 20

26 19 22 28 14 18 18 17

In the ANOVA table, the number of relationships that should be investigated is the same as the df. The total SS for factor is decomposed into parts with unit dfs. These parts are the linear, quadratic, cubic, etc. parts of the relationship. Each part can then be treated separately, and the parts with small MS are pooled into the error term. The type of relationship that remains as significant can guide the experimenter in investigating the level averages.

S/N CALCULATIONS

AND INTERPRETATIONS

Control factors and noise factors were introduced in Section 3. Control factors appear in an orthogonal array called an inner array. Noise factors that represent the uncontrolled or uncontrollable environment are entered into a separate array called an outer array. The following example of an L8 linear (control) array with an L4 outer (noise) array was first presented in Section 3. Actual responses and factor names are added here — see Table 9.21 — in the development of the example. This type of experimental setup and analysis evaluates each of the control factor choices (L8 array factors) over the expected range of the uncontrollable environment

SL3151Ch09Frame Page 412 Thursday, September 12, 2002 6:05 PM

412

Six Sigma and Beyond

TABLE 9.22 The STB ANOVA Table Source

df

SS

MS

F Ratio

S’

%

A B C D E F G Error (pooled error) Total

1 1 1 1 1* 1* 1

18.487 0.864 4.232 1.295 0.223 0.213 4.362

18.487 0.864 4.232 1.295 0.223 0.213 4.362

84.803 3.963 19.413 5.940

18.269 0.646 4.014 1.077

61.53 2.18 13.53 3.63

20.009

4.144

13.96

2 7

0.436 29.676

0.218 4.239

1.526

5.14

(L4 array factors). This assures that the optimal factor levels from the L8 array will be robust. An S/N can be calculated for each test situation. These S/N ratios are then used in an ANOVA to identify the situation that maximizes the S/N. Smaller-the-Better (STB) The following S/N ratios are calculated for the STB situation using the equations given in Section 4 and assuming that the optimum value is zero and that the responses shown represent deviations from that target: Test Number

STB S/N

1 2 3 4 5 6 7 8

28.65 27.32 26.05 28.32 22.34 24.63 25.61 25.59

The S/N ratios for testing situations are then analyzed using an ANOVA table. The STB ANOVA table for the example is shown in Table 9.22. The ANOVA table indicates that factors A, G, and C are the most significant contributors. Inspection of the level averages shows that the highest S/N values (least negative), in order of contribution, occur at A2, G2, C2, D1, B1. Estimation of the S/N at the optimal levels can be made from the S/N level averages using the technique discussed earlier in this section. Likewise, estimation of the raw data average response at the optimal level can be made from the response level averages at the optimal S/N factor levels.

SL3151Ch09Frame Page 413 Thursday, September 12, 2002 6:05 PM

Design of Experiments

413

TABLE 9.23 The LTB ANOVA Table Source

df

SS

MS

F Ratio

S’

%

A B C D E F G Error (pooled error) Total

1 1 1 1 1* 1* 1

18.292 1.121 4.160 1.271 0.396 0.264 4.947

18.292 1.121 4.160 1.271 0.396 0.264 4.947

55.442 3.397 12.605 3.852

17.966 0.791 3.830 0.941

58.99 2.60 12.58 3.09

14.991

4.617

15.16

2 7

0.660 30.454

0.330 4.351

2.310

7.59

Larger-the-Better (LTB) The same data will be used to demonstrate the LTB notation. In this case, the optimum value is infinity. Examples of this include strength or fuel economy. The following S/N ratios are calculated using the LTB equation given in Section 4.

Test Number

LTB S/N

1 2 3 4 5 6 7 8

28.57 26.98 25.94 28.23 22.08 24.54 25.48 25.52

The S/N ratios for testing situations are then analyzed using an ANOVA table. The LTB ANOVA table for the example is shown in Table 9.23. Inspection of the ANOVA table and the level averages shows that the highest S/N values occur at A1, G1, C1, D2, B2. Interpretation of the LTB analysis is similar to that of the STB analysis. Nominal the Best (NTB) Analysis of the NTB experiment is a two-part process. Again, the same data will be used to illustrate this approach. The target value will be assumed to be 24 in this case.

SL3151Ch09Frame Page 414 Thursday, September 12, 2002 6:05 PM

414

Six Sigma and Beyond

TABLE 9.24 The NTB ANOVA Table Source

df

SS

MS

F Ratio

A B C D E F G Error (pooled error) Total

1* 1 1* 1* 1 1 1

0.193 9.618 0.006 0.333 17.816 2.477 10.424

0.193 9.618 0.006 0.333 17.816 2.477 10.424

3 7

0.532 40.867

0.177 5.838

S’

%

54.339

9.441

23.10

100.655 13.994 58.893

17.639 2.300 10.247

43.16 5.63 25.07

1.240

3.03

The S/N values are analyzed. The following S/N are calculated: Test Number

STB S/N

1 2 3 4 5 6 7 8

21.93 15.96 20.78 21.60 17.03 21.59 20.33 22.56

The S/N ratios for testing situations are then analyzed using an ANOVA table. The NTB ANOVA table for the example is shown in Table 9.24. The ANOVA table and the level averages indicate that E1, G1, B2, F1 are the optimal choices from an S/N standpoint. These are the factor choices that should result in the minimum variance of the response. The ANOVA analysis and level averages of the raw data are then investigated to determine if there are other factors that have significantly different responses at their different levels but are not significant in the S/N analysis. These factors can be used to tune the average response to the desired value but do not appreciably affect the variability of the response. The ANOVA table of the raw data is shown in Table 9.25. From this ANOVA table, it can be seen that the significant contributors to the observed variability of the data averages are the factors A, G, C, D, and F. This can be combined with the S/N analysis and interpreted as follows: a. Factors that influence variability only — B, E b. Factors that influence both variability and average response — G c. Factors that influence the average only — A, C d. Factors that have little or no influence on either variability or average response – D, F

SL3151Ch09Frame Page 415 Thursday, September 12, 2002 6:05 PM

Design of Experiments

415

TABLE 9.25 Raw Data ANOVA Table Source A B C D E F G X Y Z Error (pooled error) Total

df 1 1* 1 1 1* 1 1 1* 1* 1* 21 26 31

SS

MS

F Ratio

S’

%

392.000 8.000 72.000 18.000 2.000 18.000 98.000 0.125 3.125 0.000 106.750 120.000 718.000

392.000 8.000 72.000 18.000 2.000 18.000 98.000 0.125 3.125 0.000 5.083 4.615 23.161

84.940

387.385

53.95

15.601 3.900

67.385 13.385

9.39 1.86

3.900 21.235

13.385 93.385

1.86 13.01

143.075

19.93

The results from this experiment indicate that factors B, E, and G should be set to the levels with the highest S/N. Factor G should be set to the level with the highest S/N rather than using it to tune the average since its relative contribution to S/N variability is greater than its contribution to the variability of raw data. This decision might change based on cost implications and the ability to use factors A and C to tune the average response. Factors A and C should be investigated to determine if they can be set to levels that will allow the target value of 24 to be attained. This may be possible with factors that have continuous values. Factors with discrete choices such as supplier or machine number cannot be interpolated. Factors D and F should be set to the levels that are least expensive. A series of confirmation runs should be made when the optimum levels have been determined. The average response and S/N should be compared to the predicted values.

COMBINATION DESIGN Combination design was mentioned in Section 3 as a way of assigning two twolevel factors to a single three-level column. This is done by assigning three of the four combinations of the two two-level factors to the three-level factor and not testing the fourth combination. As an example, two two-level factors are assigned to a threelevel column as in Table 9.26. Note that the combination A1B2 is not tested. In this approach, information about the A.B interaction is not available, and many ANOVA computer programs are not able to break apart the effect of A and B. The sum of squares (SS) in the ANOVA table that is due to factor A.B contains both the SS due to factor A and the SS due to factor B. These two SSs are not additive since the factors A and B are not orthogonal. This means:

SL3151Ch09Frame Page 416 Thursday, September 12, 2002 6:05 PM

416

Six Sigma and Beyond

TABLE 9.26 Combination Design Factor A

Factor B

Three Level Column Combined Factor (A.B)

1 2 2

1 1 2

1 2 3

SSAB ≠ SSA + SSB

The SS of A and B can be calculated separately as follows:

( = (T

SSA = TAB1 − TAB2 SSB

AB2

− TAB3

) / (2 * r ) ) / (2 * r ) 2

2

where TAB1 = the sum of all responses run at the first level of AB; TAB2 = the sum of all responses run at the second level of AB; TAB3 = the sum of all responses run at the third level of AB; and r = the number of data points run at each level of AB. The MS of A and B then can be separately compared to the error MS to determine if either or both factors are significant. The df for both A and B is 1. If one of the factors is significant and the other is not, the ANOVA should be rerun with the significant factor shown with a dummy treatment and the other factor excluded from the analysis. EXAMPLE The following factors will be evaluated using an L9 orthogonal array: Factor

Number of Levels

A B C D E

2 2 3 3 3

A and B will be combined into a single three-level column. The test array and results are shown in Table 9.27. The sum of the data at each level of AB is: for AB = 1, the sum is 17 + 9 + 8 = 34; for AB = 2, the sum is 40 + 28 + 17 = 85; for AB = 3, the sum is 28 + 22 + 27 = 77.

SL3151Ch09Frame Page 417 Thursday, September 12, 2002 6:05 PM

Design of Experiments

417

TABLE 9.27 L9 OA with Test Results A

B

A.B

C

D

E

Test Results

1 1 1 2 2 2 2 2 2

1 1 1 1 1 1 2 2 2

1 1 1 2 2 2 3 3 3

1 2 3 1 2 3 1 2 3

1 2 3 2 3 1 3 1 2

1 2 3 3 1 2 2 3 1

7 3 5 22 13 9 12 12 15

10 6 3 18 15 8 16 10 12

Sum of the Test Results 17 9 8 40 28 17 28 22 27

TABLE 9.28 ANOVA Table Source A.B (A) (B) C D E Error (pooled error) Total

df 2 1 1 2 2 2 9 9 17

SS

MS

F Ratio

S’

%

250.778 (216.750) (5.333) 100.778 33.778 32.444 36.000 36.000 453.778

125.389 216.750 5.333 50.389 16.889 16.222 4.000 4.000 26.693

31.347 54.188 1.333 12.597 4.222 4.056

242.778

53.50

92.778 25.778 24.444

20.45 5.68 5.39

68.000

14.99

(

) ( ) 2

SSA = 24 − 85 / 2 * 6 SSA = 216.75

(

) ( ) 2

SSB = 85 − 77 / 2 * 6 SSB = 5.33

The ANOVA table is for the data shown — see Table 9.28. The decomposed SS for A and B are shown in parentheses and are not added into the total SS.

SL3151Ch09Frame Page 418 Thursday, September 12, 2002 6:05 PM

418

Six Sigma and Beyond

TABLE 9.29 Second Run of ANOVA Source

df

SS

MS

F Ratio

S’

%

A C D E Error (pooled error) Total

1 2 2 2 10 10 17

245.444 100.778 33.778 32.444 41.334 41.334 453.778

245.444 50.389 16.889 16.222 4.133 4.133 26.693

59.386 12.192 4.086 3.925

241.311 92.512 25.512 24.178

53.18 20.39 5.62 5.33

70.265

15.48

The F ratio for factor B indicates that the effect of the change in factor B on the response is insignificant. Factor B is excluded from the analysis and factor A is analyzed with a dummy treatment. The ANOVA table for this analysis is shown in Table 9.29. The analysis continues using the techniques described in this section.

MISCELLANEOUS THOUGHTS The purpose of most DOEs is to predict what the response will be at the optimum condition. Confirmatory tests should be run to assure the experimenter that the projected results are valid. Sometimes, the results of the confirmatory tests are significantly different from the projected results. This can be due to one or more of the following: • There was an error in the basic assumptions made in setting up the experiment. • Not all of the important factors were controlled in the experiment. • The factors interacted in a manner that was not accounted for. • The response that was measured was not the proper response or was only a symptom of something more basic (see Section 2). • An important noise factor was not included in the experiment (e.g., the experimental tests were run on sunny days while the confirmatory tests were run on a rainy day). • The experimental test equipment is not capable of providing consistent, repeatable test results. • A mistake was made in setting up one or more of the experimental tests. The experimenter who is faced with data that does not support the prediction is forced to ask which of these problems affected the results. It is important that all of these problems be considered and investigated, if appropriate. If two or more of these problems coexisted, correcting only one problem may not improve the experimental results. Even though it may seem that the experiment was a failure, that is not necessarily true. Experimentation should be considered an organized approach to uncovering a

SL3151Ch09Frame Page 419 Thursday, September 12, 2002 6:05 PM

Design of Experiments

419

TABLE 9.30 L8 with Test Results and S/N Values L8 Test No.

A 1

B 2

C 3

D 4

E 5

F 6

G 7

1 2 3 4 5 6 7 8

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

Z Y X

1 1 1

2 2 1

2 1 2

1 2 2

s

–20 log(s)

Test Results

25 25 18 26 15 18 20 19

27 27 21 23 11 15 17 20

30 21 19 27 12 17 21 20

26 19 22 28 14 18 18 17

2.16 3.65 1.83 2.16 1.83 1.41 1.83 1.41

–6.690 –11.249 –5.229 –6.690 –5.229 –3.010 –5.229 –3.010

working knowledge about a situation. The “failed” experiment does provide new knowledge about the situation that should be used in setting up the next iteration of experimental testing. The prior statement may sound too idealistic for the “real” world where deadlines are very important. A failed experiment may cause some people to doubt the usefulness of the DOE approach and extol the virtues of traditional one-factor-at-a-time testing. However, all of the problems listed above that could cause a DOE to fail will also cause a one-factor-at-a-time experiment to fail. In DOE, the problem will be found fairly early since relatively few tests are run. In one-factor-at-a-time testing, the problem may not surface until many tests have been run, or the problem may not even be identified in the testing program. In this case, the problem may not show up until production or field use. The importance of meeting real-world deadlines makes the planning stage of the experiment critical. Proper planning, including consideration of the experience and knowledge of experts, will enable the experimenter to avoid many of the possible problems. Deadlines are never a good excuse for not taking the time to adequately plan an experiment. AN EXAMPLE The data used to demonstrate the S/N calculations in this section will be analyzed here using the approach, NTBII S/N = –10 log (s2) = –20 log (s). This approach was discussed earlier in this chapter. The data set is repeated in Table 9.30. The S/N ratios for the testing situations are then analyzed using an ANOVA table. The NTBII ANOVA table for the example is shown in Table 9.31. To help interpret the ANOVA table, the level standard deviation averages and the level S/N averages are shown for the significant factors in Table 9.32. To give a visual impact of the spread of the data and what the above table really means, it would be wise to plot the data for each factor level. The plots of the average standard deviation by factor level are shown in Figure 9.20.

SL3151Ch09Frame Page 420 Thursday, September 12, 2002 6:05 PM

420

Six Sigma and Beyond

TABLE 9.31 ANOVA Table for Data from Table 9.30 Source

df

SS

MS

F Ratio

S’

%

A B C D E F G Error (pooled error) Total

1 1 1 1* 1 1* 1*

22.379 4.531 4.531 0.313 13.670 1.200 1.200

22.379 4.531 4.531 0.313 13.670 1.200 1.200

24.746 5.010 5.010

21.474 3.627 3.627

44.90 7.58 7.58

15.117

12.766

26.69

3 7

2.713 47.823

0.904 6.832

6.330

13.24

TABLE 9.32 Significant Figures from Table 9.31

Factor

Level

Average Standard Deviation

A

1 2 1 2 1 2 1 2

2.36 1.61 2.12 1.79 2.12 1.79 1.67 2.26

B C E

NTBII S/N –7.465 –4.120 –6.545 –5.039 –6.545 –5.039 –4.485 –7.099

The ANOVA table and the level average standard deviations indicate that A2B2C2E1 are the optimal choices from an NTBII S/N standpoint. The analysis of the raw data remains the same as shown in the chapter. The average level of the response should be targeted using the results of the raw data analysis. This is true regardless of whether the goal is as small as possible, as large as possible, or to meet a specific value. The variance should be minimized by maximizing the NTBII S/N. The experimenter must make the trade-off between the choice of factor levels that adjust the response average and the choice of factor levels that minimize the variance of the response. A comparison of the results of the two methods shows clear differences. As an example, for the situation where a specific value is targeted (NTB), the factor level choices are: NTB — B2E1G1 to minimize variability, A and C set to achieve target; NTBII — B2E1 to minimize variability, G set to achieve target. If the target is attainable using factor G, use A2C2 to minimize variability, otherwise, set C and/or A to achieve target.

SL3151Ch09Frame Page 421 Thursday, September 12, 2002 6:05 PM

Design of Experiments

421

Standard Deviation

2.5

2

1.5 1

2

Factor A

1

2

Factor B

1

2

Factor C

1

2

Factor E

FIGURE 9.20 Plots of the average standard deviation by factor level.

There is no complete agreement among statisticians and DOE practitioners as to which approach gives better results. As a general rule, the reader is encouraged to: 1. Plot the data including raw and/or transformed values, level averages and standard deviations, and any other information that seems appropriate. One picture is worth a thousand words. 2. Analyze the data using the appropriate analysis techniques. 3. Compare the results to the data plots in order to determine which set of results makes the most sense. Perform this comparison fairly and resist the temptation to choose the results solely on whether they support convenient conclusions. 4. Run confirmation tests. DOE is a powerful tool that can help the experimenter get the most out of scarce testing resources. However, as with any powerful tool, care must be taken to understand how to use the tool and how to interpret the results.

ANALYSIS OF CLASSIFIED DATA The purpose of this section is to: 1. Discuss the classified attribute analysis and classified variable analysis approaches to analyzing classified responses. 2. Present examples of how these techniques are used.

SL3151Ch09Frame Page 422 Thursday, September 12, 2002 6:05 PM

422

Six Sigma and Beyond

TABLE 9.33 Observed Versus Cumulative Frequency Observed Frequency

Cumulative Frequency

2 1 1

2 3 4

Class I Class II Class III

CLASSIFIED RESPONSES Some experimental responses cannot be measured on a continuous scale although they can be divided into sequential classes. Examples include appearance and performance ratings. In these situations, three to five rating classes are generally the optimum number because this number allows major differences in the responses to be identified and yet does not require the rater to identify differences that are too subtle. Two related techniques are used to analyze classified responses: 1. Classified attribute analysis is used when the total number of items rated is the same for every test matrix setup. 2. Classified variable analysis is used when the total number of items rated is not the same for every test matrix setup. Three to five responses at each experimental setup are recommended to give a good evaluation of the class distribution of responses at that setup. As with continuous measurements, more responses at each setup allow smaller differences to be identified.

CLASSIFIED ATTRIBUTE ANALYSIS This technique converts the observed frequency in each class into a cumulative frequency for the classes. As an example, if there are three classes, the observed and cumulative frequencies might be as shown in Table 9.33. It is assumed that the user will use a computer program to analyze the classified data. The specific input format will depend on the computer program used. The mathematical derivations and philosophies of this approach will not be presented here. For more information see Volume V of this series as well as Taguchi (1987) and Wu and Moore (1985). EXAMPLE Three grades are used to evaluate paint appearance of a product. They are “Bad,” “OK,” and “Good.” Seven factors (A through G), each at two levels, are evaluated to determine the combination of factor levels that optimizes paint appearance. Five products are evaluated at each testing situation in an L8 orthogonal array. Test results are shown in Table 9.34.

SL3151Ch09Frame Page 423 Thursday, September 12, 2002 6:05 PM

Design of Experiments

423

TABLE 9.34 Attribute Test Setup and Results Frequency in Each Grade A

B

C

D

E

F

G

Bad

OK

Good

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

2 3 4 0 0 1 0 0

3 2 1 2 4 3 3 1

0 0 0 3 1 1 2 4

TABLE 9.35 ANOVA Table (for Cumulative Frequency) Source

df

A B C D E F G Error (pooled error) Total

2 2 2* 2* 2* 2 2* 64 72 78

SS

MS

11.668 6.678 0.125 3.668 2.259 7.935 2.259 45.409 53.720 80.000

5.834 3.39 0.063 1.834 1.130 3.986 1.130 0.710 0.746 1.026

F Ratio

S’

%

7.820 4.476

10.179 5.186

12.72 6.48

5.319

6.443

8.05

58.196

72.75

The ANOVA analysis for this set of data is shown in Table 9.35. Note that the degrees of freedom are calculated differently from the non-classified situation. The df of each source is: (the number of levels of that factor – 1) * (the number of classes – 1) In this example, the number of levels of each factor is two and the number of classes is three. For each factor,

( )( )

df = 2 − 1 * 3 − 1 = 2 The total df = (the total number of rated items – 1) * (the number of classes – 1). Thus, the total df for this example is:

SL3151Ch09Frame Page 424 Thursday, September 12, 2002 6:05 PM

424

Six Sigma and Beyond

TABLE 9.36 The Effect of the Significant Factors Observed Frequency

A1 A2 B1 B2 F1 F2 Total

% Rate of Occurrence (R.O.)

Cumulative Frequency

Cumulative % R.O.

Bad

OK

Good

Bad

OK

Good

Bad

OK

Good

Bad

OK

Good

9 1 6 4 2 8 10

8 11 12 7 10 9 19

3 8 2 9 8 3 11

45 5 30 20 10 40

40 55 60 35 50 45

15 40 10 45 40 15

9 1 6 4 2 8

17 12 18 11 12 17

20 20 20 20 20 20

45 5 30 20 10 40 25

85 60 90 55 60 85 73

100 100 100 100 100 100 100

Cumulative Rate of Occurrence - %

Factor Effects 100 90 80 70 60 50 40 30 20 10 0 A-1

A-2

B-1

B-2

F-1

F-2

Factor - Level Bad

OK

Good

FIGURE 9.21 Factor effects.

(

)( )

df = 40 − 1 * 3 − 1 = 78 The error df is the total df minus the df of each of the factors. From the ANOVA table, factors A, B, and F are identified as significant. The effects of these factors are shown in Table 9.36 and Figure 9.21.

Although interpretation and use of the ANOVA table in classified attribute analysis is the same as for the non-classified situation, a significant difference does

SL3151Ch09Frame Page 425 Thursday, September 12, 2002 6:05 PM

Design of Experiments

425

Factor Effects Cumulative Rate of Occurrence - %

100 90 80 70 60 50 40 30 20 10 0 A-1

A-2

A-3

B-1

B-2

B-3

C-1

C-2

C-3

Factor - Level Class 1

Class 2

Class 3

FIGURE 9.22 Factor effects.

exist in estimating the cumulative rate of occurrence for each class under the optimum condition. Percentages near 0% or 100% are not additive. The cumulative of occurrence can be transformed using the omega method to obtain values that are additive. In the omega method, the cumulative percentage (p) is transformed to a new value (Ω) as follows:

(

)

Ω = −10 log10 l / p − 1 [the units of Ω are decibels (db).] Using this transformation, the estimated cumulative rate of occurrence for each class at the optimum condition (A2B2F1) is calculated as follows:

(

) (

) (

db of µˆ = db of T + db of A2 − db of T + db of B2 − db of T + db of F1 − db of T

)

The estimated cumulative rate of occurrence for each class for the optimum condition is: Class 1

(

) (

)

db of µˆ = db of .25 + db of .05 − db of .25 + db of .20 − db of .25

(

)

+ db of .10 − db of .25

(

) (

) (

)

= −4.77 + −12.79 + 4.77 + −6.02 + 4.77 + −9.54 + 4.77 = −18.81 µˆ = 1%

SL3151Ch09Frame Page 426 Thursday, September 12, 2002 6:05 PM

426

Six Sigma and Beyond

TABLE 9.37 Rate of Occurrence at the Optimum Settings Class

Cumulative Rate of Occurrence

Rate of Occurrence

Bad OK Good

1% 27% 100%

1% 26% 73%

Class 2

(

) (

)

db of µˆ = db of .73 + db of .60 − db of .73 + db of .55 − db of .73

(

)

+ db of .60 − db of .73 = −4.25 µˆ = 27%

These results are summarized in Table 9.37.

CLASSIFIED VARIABLE ANALYSIS Classified variable analysis is used when the number of items evaluated is not the same for all test matrix setups. As with classified attribute analysis, the computer analyzes the cumulative frequencies. EXAMPLE Four factors (A, B, C and D) are suspected of influencing door closing efforts for a particular car model. An experiment was set up that evaluated each of these factors at three levels. An L9 orthogonal array was used to evaluate the factor levels. Door closing effort ratings were made by a group of typical customers. Each customer was asked to evaluate the doors on a scale of one to three as follows: Class 1 2 3

Description of Effort Unacceptable Barely acceptable Very good feel

The experimental setup and test results are shown in Table 9.38 and Figure 9.22. The ANOVA analysis for this set of data is shown in Table 9.39. From the ANOVA table, factors A, B and C are identified as significant. The effects of these factors are shown in Table 9.40.

SL3151Ch09Frame Page 427 Thursday, September 12, 2002 6:05 PM

Design of Experiments

427

TABLE 9.38 Door Closing Effort: Test Setup and Results

A

B

C

D

Number of Ratings

1 1 1 2 2 2 3 3 3

1 2 3 1 2 3 1 2 3

1 2 3 2 3 1 3 1 2

1 2 3 3 1 2 2 3 1

5 4 5 4 4 4 5 5 4

Class% Rate of Occurrence

Ratings by Class

Class Cumulative Frequency (%)

1

2

3

1

2

3

1

2

3

1 2 2 0 0 0 3 4 3

3 1 3 0 1 1 2 1 1

1 1 0 4 3 3 0 0 0

20 50 40 0 0 0 60 80 75

60 25 60 0 25 25 40 20 25

20 25 0 100 75 75 0 0 0

20 50 40 0 0 0 60 80 75

80 75 100 0 25 25 100 100 100

100 100 100 100 100 100 100 100 100

TABLE 9.39 ANOVA Table for Door Closing Effort Source A B C D Error (pooled error) Total

df

SS

MS

F Ratio

S’

%

4 4 4 4* 1782 1786 1798

871.296 34.404 25.125 4.827 864.291 869.118 1800.000

217.824 8.601 6.296 1.207 0.485 0.487 1.001

447.277 17.661 12.928

869.348 32.456 23.234

48.30 1.80 1.29

874.962

48.61

The choice of the optimum levels is clear for factors A and B. A2 and B1 are the best choices. Two different choices are possible for factor C, depending on the overall goal of the design. If the goal is to minimize the occurrence of unacceptable efforts, C1 is the best choice. If the goal is to maximize the number of customer ratings of “very good,” then C2 is the best choice. For this example, C1 will be chosen as the preferred factor setting. The estimated rate of occurrence for each class for the optimum setting, A2B1C1, can be calculated using the omega method. The estimated rates are shown in Table 9.41. The df for the factors are calculated in the same way as with the Classified Attribute Analysis, i.e., df = (the number of levels of that factor – 1) * (the number of classes – 1). In Classified Variable Analysis, the total number of items evaluated at each condition is not equal. To “normalize” these sample sizes, percentages are analyzed and the “sample size” for each test setup becomes 100 (for 100%). The total df is (the number of “sample sizes” – 1) * (the number of classes – 1). For this example, the total df is:

SL3151Ch09Frame Page 428 Thursday, September 12, 2002 6:05 PM

428

Six Sigma and Beyond

TABLE 9.40 The Effects of the Door Closing Effort Factor & Level A1 A2 A3 B1 B2 B3 C1 C2 C3 Total

% Rate of Occurrence

Cumulative% Rate of Occurrence

Class 1

Class 2

Class 3

Class 1

36.7 0 71.7 26.7 43.3 38.3 33.3 41.7 33.3 36.1

48.3 16.7 28.3 33.3 23.3 36.7 35.0 16.7 41.7 31.1

15.0 83.3 0 40.0 33.3 25.0 31.7 41.7 25.0 32.8

36.7 0 71.7 26.7 43.3 38.3 33.3 41.7 33.3 36.1

Class 2

Class 3

85.0 16.7 100.0 60.0 66.6 75.0 68.3 58.4 75.0 67.2

100 100 100 100 100 100 100 100 100 100

TABLE 9.41 Rate of Occurrence at the Optimum Settings Cumulative Rate of Occurrence

Class 1 (unacceptable) 2 (barely acceptable) 3 (very good feel)

Rate of Occurrence

0% 13.4% 100%

(

0% 13.4% 86.6%

)( )

df = 900 − 1 * 3 − 1 = 1798

The error df is the total df minus the df of each of the factors.

DISCUSSION

OF THE

DEGREES

OF

FREEDOM

In both classified attribute analysis and classified variable analysis, the total degrees of freedom are much greater than the number of items evaluated. The interpretation of the F ratios and the calculation of a confidence interval are complicated by the large number of degrees of freedom and will not be addressed here. The analysis techniques for classified responses are not as completely developed as are the techniques for the analysis of continuous data. In Dr. Taguchi’s approach, the emphasis is on using the percent contribution to prioritize alternative choices. Although better statistical techniques may be developed to handle classified data, classified attribute and classified variable analyses can be used to identify the large contributors to variation in classified responses.

SL3151Ch09Frame Page 429 Thursday, September 12, 2002 6:05 PM

Design of Experiments

429

MISCELLANEOUS THOUGHTS As we just mentioned in the discussion of the degrees of freedom, there is no consensus among statisticians regarding the best method to use to analyze classified data. A method that is an alternate to the ones described in this section is to transform the classified data into variable data and analyze the data as described in Section 5. A drawback to this approach is that the relative difference in the transformed values should reflect the relative difference in the classifications, and this is sometimes difficult to achieve. A simple example from the medical field will illustrate this. Four different groups of patients suffering from the same disease are each given a different medicine. The purpose is to determine which medicine is best. The response classes are shown in below: Class A B C

Description of Effect Patient improves No change in patient Patient dies

If Class A is given a value of 1 and Class B is given a value of 2, what should Class C be given? Is the difference between Classes B and C the same as the difference between Classes A and B? Twice the difference? Three times? Dr. George Box is of the opinion that this difficulty can be overcome by analyzing the variable data using several different transformations from the classifications. In most instances, the choice of the best response will not be affected by the different relative values placed on the classifications and, in every case, the data will be much easier to analyze and interpret. The example given earlier dealing with classified attribute data will be worked as an example. EXAMPLE Three grades are used to evaluate paint appearance of a product. They are “Bad,” “OK,” and “Good.” The classified data are transformed into variable data as follows: Bad = 1; OK = 3; Good = 4. This puts emphasis on avoiding situations that result in “bad” responses. Seven factors (A through G), each at two levels, are evaluated to determine the combination of factor levels that optimizes paint appearance. Five products are evaluated at each testing situation in an L8 orthogonal array. Test results are shown in Table 9.42. The ANOVA analysis for the raw data is shown in Table 9.43. Plotting of the data and inspection of the level averages reveal that the best factor choices are: A2B2F1. The ANOVA analysis for the NTB S/N ratios is shown in Table 9.44. Plotting of the S/N data and inspection of the level averages reveals that the best factor choices are: A2B2E2F1. The best choices overall are: A2B2E2F1. This compares with the best choice of A2B2F1 from the accumulation analysis on page 425. Each of the methods has one further disadvantage. Using the transformation approach, it is not possible to make a projection of what the distribution of classes would look like at the optimum settings. The accumulation analysis was not able to identify the effect on the standard deviation of the ratings due to factor E. Each approach tells a different part of the story and both should be used to get the full picture.

SL3151Ch09Frame Page 430 Thursday, September 12, 2002 6:05 PM

430

Six Sigma and Beyond

TABLE 9.42 OA and Test Setup and Results Test Setup and Results Frequency in Each Grade A

B

C

D

E

F

G

Bad

OK

Good

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

2 3 4 0 0 1 0 0

3 2 1 2 4 3 3 1

0 0 0 3 1 1 2 4

Transformed Data 1 1 1 3 3 1 3 3

1 1 1 3 3 3 3 4

3 1 1 4 3 3 3 4

3 3 3 4 4 4 4 4

TABLE 9.43 ANOVA for the Raw Data Source

df

SS

MS

A B C D E F G Error (pooled error) Total

1 1 1* 1* 1* 1 1* 32 36 39

11.03 3.03 0.03 2.03 2.03 7.23 2.03 21.60 27.70 48.98

11.03 3.03 0.03 2.03 2.03 7.23 2.03 0.68 0.77 1.26

F Ratio

S’

%

14.33 3.93

10.26 2.26

20.94 4.61

9.39

6.46

13.18

30.01

61.27

DYNAMIC SITUATIONS This section discusses: 1. What dynamic test situations are 2. How a test plan should be set up in a dynamic situation 3. The analysis of test data

DEFINITION In many instances, the experimenter knows that the optimum response for a system changes with levels of an input signal. Using the signal-to-noise techniques described

SL3151Ch09Frame Page 431 Thursday, September 12, 2002 6:05 PM

Design of Experiments

431

TABLE 9.44 ANOVA Table for the NTB S/N Ratios Source

df

SS

MS

A B C D E F G Error (pooled error) Total

1 1 1* 1* 1 1 1* 3

23.81 23.81 0.39 0.39 13.21 23.81 3.49

23.81 23.81 0.39 0.39 13.21 23.81 3.49

4.26 88.91

1.42 12.70

7

F Ratio

S’

%

16.76 16.76

22.39 22.39

25.18 25.18

9.29 16.76

11.79 22.39

13.26 25.18

9.95

11.19

in the previous sections would yield incorrect results. These techniques emphasize repeatability across all levels of the noise factors. In a dynamic situation, the experimenter wants different responses depending upon the level of an input signal factor. Two examples are: 1. If two or more length measurement devices are compared, the standard lengths to be measured become signal factor levels for comparison. The experimenter would want a measurement device that gives a reading that is relative to the different standard lengths measured and is repeatable at each standard measured. 2. Several factors are to be included in an experiment to determine the combination that optimizes vehicle braking distance. The tests are run at two different vehicle speeds. The vehicle speed would be treated as a signal factor. The experimenter would want the braking distance to be repeatable at each vehicle speed and reflect the customers’ needs and desires for braking distance at each vehicle speed. These needs and desires would not be the same for all vehicle speeds. (Note: It might seem that the goal should be to minimize the braking distance at each vehicle speed; however, if the braking were too abrupt, the driver might lose control of the vehicle.)

DISCUSSION The analysis of dynamic test data can be complicated. The following conditions exist in the examples that follow. If these conditions are not present in a dynamic experiment, help from a statistician should be sought before setting up the experiment. Conditions 1. Signal factors will be assigned to an outer array in the experimental setup. 2. The signal factor(s) will have either two or three levels.

SL3151Ch09Frame Page 432 Thursday, September 12, 2002 6:05 PM

432

Six Sigma and Beyond

TABLE 9.45 Typical ANOVA Table Setup Source

df

SS

V

Signal error

dfs df1 – dfs

SSs SS1 – SSs

Vs Ve

Total

df1

SS1

3. If there are three levels for a signal factor, the intervals between the adjacent levels will be equal. 4. The experimental test includes either noise factors in an outer array or repetitions so that for each inner array control factor setup, two or more runs are made for each signal factor. Analysis The general approach that is used to analyze the data is: 1. The test results for each inner array setup (test number) are separately analyzed using analysis of variance (ANOVA). The ANOVA table for these analyses will be shown in a typical format as Table 9.45. 2. A nominal-the-best signal-to-noise ratio is calculated for each inner array setup from these ANOVA tables as follows:  SS − V  s e S / N = 10 log10   V * r * s * h2   e

where r = the number of data in each level of the signal factor for this inner array setup; s = 0.5 if the signal factor has two levels or s = 2.0 if the signal factor has three levels; and h = the interval between the adjacent levels of the signal factor. 4. The calculated S/N ratio for each inner array setup is then used in a nominal-the-best (NTB) S/N analysis of variance to determine which control factor settings should be used to reduce variability and which should be used to tune the response to the desired output. The application of these steps will be developed more fully through the following examples. EXAMPLE 1 Two different types of automatic optical measurement devices are to be compared. Two orientations of the devices are possible, horizontal or vertical. These are assigned to an L4 inner array as follows:

SL3151Ch09Frame Page 433 Thursday, September 12, 2002 6:05 PM

Design of Experiments

433

TABLE 9.46 L4 OA with Test Results Test Matrix

Test Number

T

O

1 2 3 4

1 1 2 2

1 2 1 2

T

S1

×O 1 2 2 1

S2

F1

F2

F1

F2

NTB S/N

9.8 10.2 9.6 10.2

9.7 9.9 9.9 9.8

20.4 20.3 19.6 19.7

20.2 20.1 20.0 19.5

19.33 14.94 12.05 12.65

Factor

Column

Type (T) Orientation (O) T × O Interaction

1 2 3

Items with two different surface finishes will be measured by the devices. Surface finish (F) will be a noise factor. Two standard lengths of 10 and 20 mm will be evaluated. These will be the two levels of the signal factor (S). The test matrix and test results for the experiment are shown in Table 9.46. For test number 1, the S/N ratio is calculated from the ANOVA table for just the runs in test number 1. S

F

Test Result

1 1 2 2

1 2 1 2

9.8 9.7 20.4 20.2

The ANOVA table for these data is shown in Table 9.47. The S/N ratio is calculated as:  SS − V  s e S / N = 10 log10  2  Ve * r * s * h   111.303 − 0.013  S / N = 10 log10    0.013 * 2 * 0.5 * 102  S / N = 19.33

SL3151Ch09Frame Page 434 Thursday, September 12, 2002 6:05 PM

434

Six Sigma and Beyond

TABLE 9.47 ANOVA Table — Raw Data Source

df

SS

V

Signal error Total

1 2 3

111.303 0.026 111.328

111.303 0.013

TABLE 9.48 ANOVA Table (S/N Ratio Used as Raw Data) Source

df

SS

MS

T O T×O Error (pooled error) Total

1 1* 1* — 2 3

20.794 4.494 5.153 9.647

20.794 4.494 5.153 4.824

30.647

10.147

F Ratio 4.311

S’

%

15.970 14.471

52.46 47.54

An S/N ratio for each of the other test setups is calculated in a similar manner. These S/N ratios are then analyzed using the S/N ratio as a single response for each test setup — see Table 9.48. The level averages for the data are: Level Averages NTB S/N Data T1 T2 Overall

O1

O2

Overall

19.33 12.05 15.69

14.98 12.65 13.82

17.16 12.35 14.76

Inspection of the data shows that the setting of T that gives the highest S/N is level 1. Although there are not enough test setups to allow the statistical identification of level 1 of factor O as the optimum, the data suggests that orientation 1 might work the best with device 1 and should be further investigated. The level averages of the raw data are shown in Table 9.49. The predicted averages are calculated using the techniques given in Section 5 using the interaction of T1 and O1 (assumed) as the optimum setting.

SL3151Ch09Frame Page 435 Thursday, September 12, 2002 6:05 PM

Design of Experiments

435

TABLE 9.49 Level Averages — Raw Data

T1 T2 O1 O2 Average of S Overall Average at T1O1

S1

S2

Overall

9.90 9.88 9.75 10.03 9.89

20.25 19.70 20.05 19.90 19.98

15.08 14.79 14.90 14.97 14.93 15.03

(

) µˆ = 14.93 + [(15.03 − 14.93) − (15.08 − 14.93) − (14.90 − 14.93)] + (9.98 − 14.93)

for S = 1 10 mm

µˆ = 9.87 mm

(

) µˆ = 14.93 + [(15.03 − 14.93) − (15.08 − 14.93) − (14.90 − 14.93)] + (9.98 − 14.93)

for S = 2 20 mm

µˆ = 19.96 mm Note that the readings at the optimum do not average out to the standard exactly. This assumes that the output reading can be calibrated to reflect the standard measured. The emphasis in the approach is to provide readings with low variability at each standard level output. This example was very simple and it may seem that the ANOVA was not really necessary. In many cases, the inner array will be more complicated than an L4, and the technique shown in this example will help the experimenter make an informed choice.

EXAMPLE 2 The effect of several factors on vehicle braking distance is to be investigated. The control factors to be investigated are assigned to an L8 orthogonal inner array as follows: Column 1 2 3 4 5 6 7

A — Content of material “Z” in the brake pads B — Content of material “Y” in the brake rotors A × B interaction C — Hydraulic fluid type Unassigned Unassigned D — Brake pad design

SL3151Ch09Frame Page 436 Thursday, September 12, 2002 6:05 PM

436

Six Sigma and Beyond

TABLE 9.50 OA Setup and Test Results for Example 2 S1

A X

S2

T1

T2

T1

T2

A

B

B

C

D

P1

P2

P1

P2

P1

P2

P1

P2

NTB S/N

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 2 1 2 1 1 2

39.9 42.5 45.0 40.7 39.4 37.5 36.0 39.6

40.6 42.8 41.3 39.7 40.1 37.3 38.4 40.4

40.4 42.4 41.0 40.5 39.7 37.6 36.9 40.5

40.6 42.7 44.8 40.7 38.1 37.3 37.9 40.3

140.9 143.0 141.2 141.3 139.9 137.6 135.1 139.7

141.2 142.4 143.1 140.7 139.7 138.0 139.3 140.1

140.7 141.1 143.4 140.9 141.1 137.1 138.4 142.0

139.6 142.8 142.7 139.7 139.7 137.2 136.1 138.5

15.73 14.62 5.90 15.25 12.74 20.64 6.58 9.89

Noise and signal factors are assigned to an L8 outer array as follows: Column 1 2 3 4 5 6 7

S — Vehicle speed (30 mi/h or 60 mi/h) T — Tire size Unassigned P — Pavement type (asphalt or concrete) Unassigned Unassigned Unassigned

In this example, vehicle speed is a signal factor. It is not possible that the braking distance would be the same when starting from 30 mi/h vs. 60 mi/h and therefore, different responses are expected. The experimenter has determined through market research that for this type of vehicle, the customer would prefer that the braking distance be 35 feet from 30 mi/h and 130 feet from 60 mi/h. The test setup and results are shown in Table 9.50. The unassigned columns are not shown to conserve space and to make the table more presentable. The outer array is also shown somewhat differently, with column 1 (factor S) at the top, column 2 (factor T) in the middle, and column 4 (factor P) at the bottom. Although this arrangement can be used to present the data, the unassigned columns should be added back to the arrays to aid the experimenter’s understanding of the analysis and the application of the inner and outer L8 orthogonal arrays. For the first test setup, the S/N ratio is calculated from the ANOVA table for the data in the first row. The ANOVA table for these data is:

ANOVA Table (Raw Data) Source

df

SS

V

Signal Error Total

1 6 7

20090.101 1.786 20091.888

20090.101 0.298

SL3151Ch09Frame Page 437 Thursday, September 12, 2002 6:05 PM

Design of Experiments

437

TABLE 9.51 ANOVA Table (S/N Ratio Used as Raw Data) Source

df

A B A×B C Unassigned Unassigned D Error (pooled error) Total

1* 1 1 1 1* 1* 1 — 3 7

SS

MS

0.340 85.217 7.431 47.288 1.103 4.307 28.313

0.340 85.217 7.431 47.288 1.103 4.307 28.313

5.750 173.998

1.917 24.857

F Ratio

S’

%

44.453 3.876 24.668

83.300 5.514 45.371

47.87 3.17 26.08

14.769

26.396

15.17

13.418

7.71

The S/N ratio is calculated as follows:

 SS − V  s e S / N = 10 log10   V * r * s * h2   e  20090.101 − 0.298  S / N = 10 log10    0.298 * 4 * 0.5 * 302  S / N = 15.73 An S/N ratio for each of the other test setups is calculated in a similar manner. These S/N ratios are then analyzed using the S/N ratio as a single response for each test setup — see Table 9.51. The ANOVA table indicates that factors B, C, D, and the interaction of factors A and B are significant. The level averages for these factors are:

Level Averages B1 A1 15.18

B2 A2

A1

A2

C1

C2

D1

D2

16.69

10.58

8.24

10.24

15.10

14.55

10.79

The ANOVA table and the level averages indicate that B1, C2 and D1 are the optimal choices from an S/N standpoint. These are the factor choices that should result in the minimum variance of the response. An analysis of the raw data would identify the signal factor as the most significant contributor to the variation of the data. However, this information is not useful. To increase the ability of the analysis to clearly show the significant control factors, the target braking distance for each of the signal factor levels is subtracted from all of the data collected at that signal factor level. This reduces the percent level of

SL3151Ch09Frame Page 438 Thursday, September 12, 2002 6:05 PM

438

Six Sigma and Beyond

TABLE 9.52 Transformed Data S1

A

S2

T1

T2

T1

T2

A

B

X B

C

D

P1

P2

P1

P2

P1

P2

P1

P2

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 2 1 2 1 1 2

4.9 7.5 10.0 5.7 4.4 2.5 1.1 4.6

5.6 7.8 6.3 4.7 5.0 2.3 3.4 5.4

5.4 7.4 6.0 5.5 4.7 2.6 1.9 5.5

5.6 7.7 9.8 5.7 3.1 2.3 2.9 5.3

10.9 13.0 11.2 11.3 9.9 7.6 5.1 9.7

11.2 12.4 13.1 10.7 9.7 8.0 9.3 10.1

10.7 11.1 13.4 10.9 11.1 7.1 8.4 12.0

9.6 12.8 12.7 9.7 9.7 7.2 6.1 8.5

TABLE 9.53 ANOVA Table for the Transformed Data Source A B A×B C Unassigned Unassigned D S T Unassigned P Unassigned Unassigned Unassigned A×S D×S Error (pooled error) Total

df

SS

MS

F Ratio

S’

%

1 1* 1* 1* 1* 1* 1 1 1* 1* 1* 1* 1* 1* 1* 1* 47* 30 63

150.369 1.56E-4 0.473 0.833 2.600 2.213 101.758 382.691 0.083 0.170 0.375 1.658 0.508 2.520 0.083 0.098 16.441 58.055 692.874

150.369 1.56E-4 0.473 0.833 2.600 2.213 101.758 382.691 0.083 0.170 0.375 1.658 0.508 2.520 0.083 0.098 0.988 0.968 10.998

155.340

149.401

21.56

105.122 395.342

100.790 381.723

14.55 55.09

60.959

8.80

contribution of the signal factor and increases the percent level of contribution of the control factors while maintaining their relative order of contribution. This transformation makes the effects of the control factors more visible but does not affect their significance. The transformed data are shown in Table 9.52. The ANOVA table for these data is shown in Table 9.53.

SL3151Ch09Frame Page 439 Thursday, September 12, 2002 6:05 PM

Design of Experiments

439

The interactions between all the columns of the inner array and all the columns of the outer array are available for investigation. For this example, only the A × S and D × S interactions are investigated to give an indication of whether factors A and D “behave” consistently at the two levels of the signal factor. The analysis indicates that control factors A and D are significant contributors to the variation of the data. The difference in responses between the two levels of these factors is independent of the signal level. The analysis also identified the signal factor, S, as an important contributor to the data variation. This, of course, was already known. The level averages are:

A1 A2 D1 D2 Average of S Overall Average

S1

S2

Overall

6.58 3.59 3.86 6.30 5.08

11.54 8.41 8.68 11.28 9.98

9.06 6.00 6.27 8.79 7.53

The predicted averages are calculated using the techniques given in Section 5 using A2 and D1 as the optimum settings and adding the values that were subtracted prior to the ANOVA.

(

for S = 1 30 mph

[(

)

) (

) (

)] (

)

) (

) (

)] (

)

µˆ = 7.53 + 6.00 − 7.53 − 6.27 − 7.53 − 14.90 − 14.93 + 5.08 − 7.53 + 35 µˆ = 37.29 feet

(

for S = 2 60 mph

[(

)

µˆ = 7.53 + 6.00 − 7.53 − 6.27 − 7.53 − 14.90 − 14.93 + 5.08 − 7.53 + 130 µˆ = 137.19 feet

Factor B should be set to level 1 and factor C should be set to level 2 to maximize the S/N ratio. Since the target values were not obtained at the optimum settings, the experimenter must either continue to investigate other ways of reducing the stopping distance or accept the consequences of failing to fully satisfy the customer’s requirements.

MISCELLANEOUS THOUGHTS Let us close this section with a discussion of the two examples dealing with the NTBII S/N approach. The NTB S/N ratio for a dynamic situation is:  SS − V  s e NTB S / N = 10 log10  2  Ve * r * s * h 

SL3151Ch09Frame Page 440 Thursday, September 12, 2002 6:05 PM

440

Six Sigma and Beyond

This equation was explained earlier. Using the same terminology, the NTBII S/N = NTBII S/N = –10 log (Ve) which equals –20 log (error standard deviation). For Example 1 The calculations for the NTB S/Ns were discussed earlier. The same steps are followed for the NTB approach until the final S/N calculation. The two sets of S/N ratios are contrasted below: Test Number

NTB S/N

NTBII S/N

1 2 3 4

19.33 14.94 12.05 12.65

19.03 14.88 12.04 13.01

When the NTBII S/N ratios were analyzed, the ANOVA table and the interpretation of the level averages were essentially the same as those for the NTB S/N. For Example 2 The calculations for the NTB S/N were discussed earlier. The NTBII analysis had suggested that the standard deviation of the data might be related to the average of the data. In other words, the spread of the stopping distances might be greater at standard one (30 mi/h) than at standard two (60 mi/h). Using the procedure given in pages 395–397, the averages and standard deviations were compared as follows: 1. For each vehicle speed, the average stopping distance and the standard deviation were calculated (16 averages and 16 standard deviations total). 2. The log (standard deviation) was plotted versus the log (average). 3. The slope was estimated to be in the range of 0.2 to 0.3 with large scatter in the data. By comparing this value to Item 4 on page 396, it was determined that there was not a strong need to transform the data. The NTBII S/N ratios were calculated for the untransformed data. The two sets of S/N ratios are compared below: Test Number

NTB S/N

NTBII S/N

1 2 3 4 5 6 7 8

15.73 14.62 5.90 15.25 12.74 20.64 6.58 9.89

5.26 4.19 –4.51 4.62 2.21 10.18 –3.87 –0.56

SL3151Ch09Frame Page 441 Thursday, September 12, 2002 6:05 PM

Design of Experiments

441

When the NTBII S/N ratios were analyzed, the ANOVA table and the interpretation of the level averages were essentially the same as those for the NTB S/N. The reader is encouraged to run the analysis to confirm this. The analysis of the raw data did not change. The conclusions also remained the same as before. For these two examples, the NTB and NTBII methods give equivalent results. However, this does not prove equivalency of the methods. On other sets of data, differences in the results obtained have been demonstrated. Of the two methods, the NTBII approach is easier to understand since maximizing the NTBII S/N is the same as minimizing the error variance for the chosen combination of factor levels. (The experimenter should always analyze the data completely, plot the data, compare the results to the data plots, and run confirmation tests.)

PARAMETER DESIGN This section provides an example of how the DOE technique is used to determine design factor target levels. This approach is an upstream attempt to develop a robust product that will avoid problems later in production. The emphasis at this stage is on using wide tolerance levels to provide a product that is easy to manufacture and still meets all requirements.

DISCUSSION After the basic design of a product is determined, the next step is to determine to what levels the components of that product should be set to ensure that the target will be met. The experience of the designer or design team is useful in establishing the starting values for the investigation. This investigation begins by determining what the component target values should be using wide component tolerances. This is called parameter design. If the resultant variability around the product response target is too great, the next step is to determine which tolerances should be tightened. This approach, Tolerance Design, will be discussed in Section 9. Example A particular product has been designed with five components (A through E). The target response for the product is 59.0 units. Field experience has indicated that when the response differs from the target by five units, the average customer would complain and the repair cost would be $150. From this information, the k value in the loss function can be calculated. k = $150 / 52 = $6.00 per unit 2 A brainstorming group that consisted of the designer and other experts in this area determined that the response is linear over the component range of interest and that the components should be evaluated at the levels shown in Table 9.54.

SL3151Ch09Frame Page 442 Thursday, September 12, 2002 6:05 PM

442

Six Sigma and Beyond

TABLE 9.54 Components and Their Levels Levels (Units are those appropriate to each component)

Component (Factor) A B C D E

Low

High

1000 400 50 1300 1200

1500 700 70 2200 1600

Note: Factor A is more expensive to manufacture at the higher level than at the lower level. These factors will be assigned to an L8 inner array. The two unassigned columns will be used for an estimate of the experimental error. An L8 outer array contains the low cost tolerances as follows: Low Cost Tolerance (Factor)

Low

High

A B C D E

–50 –15 –5 –200 –100

+50 +15 +5 +200 +100

The tolerance amounts are added to/subtracted from the control level as indicated by the outer array. The brainstorming group suspects that two other noise factors are significant, namely, the temperature (T) and humidity (H) of the assembly environment. The noise and tolerance factors are combined into an L8 outer array. The testing setup and test results are shown in Table 9.55. An understanding of the way the testing matrix is interpreted can be reached by considering the factor A. When the inner array column associated with factor A has a value of 1, the value of A is 1000. When there is a 2 in that column, the value of A is 1500. The actual test values of A are also determined by the tolerance value of A in the outer array. If the outer array value of A is 1, then 50 is subtracted from the value of A determined in the inner array. If the outer array value is 2, then 50 is added to it. This can be summarized as follows: Actual Test Values of A Inner Array Value of A 1 2

Outer Array Tolerance Value of A 1

2

950 1450

1050 1550

SL3151Ch09Frame Page 443 Thursday, September 12, 2002 6:05 PM

Design of Experiments

443

TABLE 9.55 L8 Inner OA with L8 Outer OA and Test Results Outer Array

L8 Inner Array A

B

C

D

E

X

Y

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

T H E D C

1 1 1 1 1

2 2 2 2 1

2 2 1 1 2

1 1 2 2 2

2 1 2 1 2

1 2 1 2 2

1 2 2 1 1

2 1 1 2 1

B A

1 1

1 1

2 1

2 1

1 2

1 2

2 2

2 2

NTB S/N

51.4 58.6 52.1 62.6 47.2 50.1 40.1 46.6

49.5 56.7 59.1 60.0 45.3 48.8 38.6 43.2

48.9 56.8 58.9 61.6 45.2 48.4 37.7 42.6

56.6 60.3 62.7 67.2 51.3 54.6 44.6 49.4

52.7 58.2 61.5 62.5 47.3 51.0 40.7 46.6

50.8 56.3 51.1 59.4 45.4 49.1 38.7 39.2

46.0 52.8 47.0 56.0 41.3 44.5 34.8 39.2

51.9 56.7 50.9 62.7 47.4 50.4 39.7 45.4

24.34 28.36 19.58 25.59 24.28 24.86 22.99 23.12

Note: X and Y are the unassigned columns that will be used to estimate error.

The ANOVA table and level averages for the most significant factors for the S/N and raw data are shown in Table 9.56. From the S/N level averages, D2B1E2 is clearly the best setting for S/N. The estimated S/N at that setting is:

(

) (

) (

S / N = 24.14 + 25.48 − 24.14 + 25.46 − 24.14 + 25.30 − 24.14

)

= 27.96 Since A1 is preferred from a cost standpoint and D2 is preferred from the S/N analysis, the next step is to determine if the value of C can be adjusted to attain the target of 59. The average response at A1D2 is: Average Response = 50.6 + (56.23 – 50.6) + (53.18 – 50.6) = 58.8 To reach a target of 59, the value of C that is included in the average response calculation must have a level average of 50.8 since: Target = Average Response at A1D2 + Effect due to C 59.0 = 58.8 + (50.8 – 50.6)

SL3151Ch09Frame Page 444 Thursday, September 12, 2002 6:05 PM

444

Six Sigma and Beyond

TABLE 9.56 ANOVA Table (NTB) and Level Averages for the Most Significant Factors Source

df

SS

A 1* 0.860 B 1 13.965 C 1 2.533 D 1 14.455 E 1 10.845 X 1* 2.477 Y 1* 10.424 Error — (pooled error) 3 1.906 Total 7 43.704 S/N Level Averages Factor Level 1 Level 2 D B E

22.80 25.46 22.98

MS 0.860 13.965 2.533 14.455 10.845 2.477 10.424

F Ratio

S’

%

21.992 3.989 22.764 17.079

13.330 1.898 13.820 10.210

30.50 4.34 31.62 23.36

4.446

10.17

0.635 6.243

25.48 22.82 25.30

Average of all data = 24.14 Source

df

Raw Data ANOVA Table SS MS F Ratio

A 1 2032.883 9.533 B 1 435.244 C 1 427.973 D 1 13.231 E 1 3.658 X 1* 3.563 Y 1* 88.125 A-Tol. 1 1.995 B-Tol. 1* 113.156 C-Tol. 1 49.879 D-Tol. 1 3.754 239.089 E-Tol. 1* 30.754 H 1 140.969 T 1* 161.297 Error 49* (pooled error) 54 Total 63 3570.410 Level Averages Factor

Level 1

A C D Average of

Level 2

56.23 44.96 47.99 53.21 48.01 53.18 all data = 50.60

2032.883 9.533 435.244 427.973 13.231 3.658 3.563 88.125 1.995 113.156 49.879 3.754 239.089 30.754 140.969 161.297 56.673

S’

%

680.577 3.192 145.713 143.279 4.430

2029.896 6.546 432.257 424.986 10.244

56.85 0.18 12.11 11.90 0.29

29.503

85.138

2.38

37.883 16.699

110.169 46.892

3.09 1.31

80.043

236.102

6.61

188.180

5.27

SL3151Ch09Frame Page 445 Thursday, September 12, 2002 6:05 PM

Design of Experiments

445

The target value for C can be interpreted from the tested levels and the level averages as follows: Target = 50 +

50.8 − 47.99 70 − 50 53.21 − 47.99

(

)

= 60.77 Note: Value of C

Response

50 Target 70

47.99 50.8 53.21

This value, 60.77, is at the same percentile between 50 and 70 as 50.8 is at between 47.99 and 53.21. In summary, the recommended target values are: Factor A B C D E

Target Value 1000 400 60.77 2200 1600

The estimated average is 59.0 and the estimated S/N is 27.96. The 90% confidence on the average is:

(

)

59.0 ± 4.02 * 2.987 * 1 / 16 + 1 / 8 or , 59.0 ± 1.50

A set of verification runs is not made using the recommended factor target values given previously. The tolerance levels from the outer array are used to define an L8 verification run experiment as shown in Table 9.57. The average response is 59.5 and the S/N is 27.3. Since the average response and the S/N are close to the predicted values, the verification runs confirm the prediction. If the average response and S/N did not confirm the predictions, the data could be analyzed to determine which factors have response characteristics different from those predicted. The information from the verification runs cannot be used directly in the loss function, since the observed variability may be affected by testing only at the tolerance limits. The center portions of the factor distributions are not represented in these tests. For the situation where the change in response is assumed to be a linear increase or decrease across the tolerance levels, the loss function can be easily calculated as follows:

SL3151Ch09Frame Page 446 Thursday, September 12, 2002 6:05 PM

446

Six Sigma and Beyond

TABLE 9.57 Variation Runs Using Recommended Factor Target Values A-Tol

B-Tol

C-Tol

D-Tol

E-Tol

H

T

Test Result

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

1 2 2 1 2 1 1 2

60.2 57.9 59.5 64.8 59.4 58.6 55.7 59.9

1. If it can be assumed that the Cpk in production will be 1.0 or greater for all specified tolerances, then the difference between the tolerance limits will be equal to or greater than six times the production standard deviation for each component parameter. 2. The difference between the response level averages for the two tolerance limits will equal six times the production response standard deviation since the product response is linearly related to the component parameter level. 3. The response variance due to each tolerance is additive since the response effect of each component tolerance is additive. (Variance = Std. Dev.2) 4. The effect of noise factors can be treated in a similar manner. In this example, the levels of humidity were set at the average humidity ± 2 times the humidity standard deviation. The change in response is assumed to be linear across the change in humidity. The difference in response between the two levels represents four times the response standard deviation. The response variance can be calculated as shown in Table 9.58. The response variance will be 3.9970. The loss function can be calculated from the equation:

(

Loss = k σ2 + offset 2

) (

)

2  Loss = $6.00 3.9970 + 59.5 − 59.0   

= $25.482 per piece For a production run of 50,000 pieces, the total loss would be $1,274,100. In the situation where the change in response is non-linear across the tolerance levels or noise factor levels, a computer simulation can be used to determine the distribution of product response for each component taken singly and for the total

SL3151Ch09Frame Page 447 Thursday, September 12, 2002 6:05 PM

Design of Experiments

447

TABLE 9.58 Calculated Response Variance Tolerance for Factor

Response Difference Between Tol Limits

Response Production Std. Dev.

A B C D E

2.2 1.0 2.2 1.6 0.1

0.37 0.16 0.37 0.27 0.02

0.1334 0.0278 0.1344 0.0711 0.0003 0.3680

3.2

0.80

0.6400 2.9890 3.9970

Humidity (H) Error Variance

Variance

assembly. This situation occurs when the highest (lowest) response occurs at the component nominal and the response decreases (increases) as the distance from the component nominal increases. The purpose of these calculations is to estimate the response variance for the total assembly population so that the loss function can be calculated. Once the value of the loss function has been calculated, it can be compared to the cost of tightening the tolerances so as to determine the optimal tolerance limits. This technique will be discussed in the following section.

TOLERANCE DESIGN This section illustrates: 1. How tolerance limits can be set so that the product will meet customer requirements repeatedly with the widest possible tolerances. The goal is to choose the most cost efficient tolerance levels. 2. How prior knowledge about the response characteristics of the component levels of a product can be efficiently used.

DISCUSSION After the target level for each component has been determined using parameter design, the loss function value of the product design is compared to design guidelines and to the cost of improving the production processes to meet tightened tolerances. If it costs less to tighten the tolerance than the resulting reduction in the loss function, then for the long run it is better to tighten the tolerance. The evaluation of the tolerance limits and the selective tightening of the limits is called tolerance design. As in parameter design, an example will illustrate this approach. The reader may need to review Volume V of this series.

SL3151Ch09Frame Page 448 Thursday, September 12, 2002 6:05 PM

448

Six Sigma and Beyond

TABLE 9.59 Cost of Reducing Tolerances Low Cost Tolerance

High Cost Tolerance

Component

Low

High

Low

High

% Reduction

Cost to Change the Tolerance for 50,000 Pieces (dollars)

A C D

–50 –5 –200

+50 +5 +200

–40 –4 –150

+40 +4 +150

20 20 25

5,000 15,000 9,000

Example Continuing the example from the previous section, the loss function with low cost tolerance was calculated to be $25.482 per piece or $1,274,100 for the production run of 50,000 pieces. This calculation was based on the assumptions that: 1. The tolerance spread is equal to six times the production standard deviation (Cpk = 1). 2. The response changes linearly across the tolerance limits. 3. The sum of the variance contributions for the components is the total assembly response variance. 4. Humidity was set at levels that are the average humidity ±2 times the standard deviation, and the response changes linearly across these levels. If any of these assumptions cannot be made, a computer simulation using the appropriate assumptions can be used to determine the total assembly response variance. The calculation of the response variance was shown in Table 9.58, and we are going to repeat it again. From the response variance contribution table and the calculations, it can be seen that the tolerances for factors A, C, and, to a lesser degree, D are the largest component tolerance contributors to the total response variance. The cost of reducing those tolerances is shown in Table 9.59. Since the response is linearly related to the component levels, a reduction of 20% in the tolerance spread will result in a reduction of 20% in the response spread. In the situation where the response is not linear, it would be necessary to run a computer simulation, as was mentioned previously. The impact of tightening the tolerance of each of the three components is summarized in Table 9.60. The variance % reduction per $1000 cost indicates that the reduction of the tolerance of component A should be investigated first since the % reduction per $1000 cost is the greatest. The situation for a reduction of 20% in the tolerance limits of component A is summarized in Table 9.61. The response variance will be 3.9484. If the same 0.5 offset is assumed, the loss function can be calculated as:

SL3151Ch09Frame Page 449 Thursday, September 12, 2002 6:05 PM

Design of Experiments

449

TABLE 9.60 The Impact of Tightening the Tolerance Response Difference Response Between Tightened Production Component Tolerance Limits Std. Dev.

Tightened Response Variance

Cost of Tightened Tolerance (dollars)

A C D

1.76 1.76 1.20

0.293 0.293 0.200

0.0858 0.0858 0.0400

5000 15,000 9000

Component

Original Variance

Tightened Variance

Variance % Reduction

Variance % Reduction per $1000 Cost

A C D

0.1344 0.1344 0.0711

0.0858 0.0858 0.0400

36.16 36.16 43.74

7.23 2.41 4.86

TABLE 9.61 Reduction of 20% in the Tolerance Limits of Component A Tolerance for Component A B C D E

Humidity (H) Error variance

Response Difference Between Tol. Limits

Response Production Std. Dev.

Response Variance

1.76 1.00 2.20 1.60 0.10

0.293 0.160 0.670 0.270 0.020

0.0858 0.0278 0.1344 0.0711 0.0003 0.3194

3.2

0.80

0.6400 2.9890 3.9484

(

Loss = k σ2 + offset 2

) (

)

2  Loss = $6.00 3.9484 + 59.5 − 59.0   

= $25.1904 per piece

SL3151Ch09Frame Page 450 Thursday, September 12, 2002 6:05 PM

450

Six Sigma and Beyond

TABLE 9.62 Reduction of Tolerance Limits for Component D Tolerance for Component A B C D E

Humidity (H) Error variance

Response Difference Between Tol. Limits

Response Production Std. Dev.

Response Variance

1.76 1.00 2.20 1.60 0.10

0.293 0.160 0.370 0.200 0.020

0.0858 0.0278 0.1344 0.0400 0.0003 0.2883

3.2

0.80

0.6400 0.2883 3.9173

For a production run of 50,000 pieces, the total loss would be $1,259,520. This is a $14,580 decrease in the loss function from the low cost tolerance situation. Since the decrease in the loss function is more than the $5000 cost of tightening the tolerance on A, it is advantageous in the long run to tighten that tolerance. The 0.50 offset is assumed to be a constant to provide a basis to compare improvement in only the variance part of the equation. In some situations, the actions taken to reduce the response variance may also result in a better-centered response distribution. The next step is to evaluate the loss function with the tolerance limits reduced for component D — see Table 9.62. The response variance will be 3.9173. If the same 0.5 offset is assumed, the loss function can be calculated as:

(

Loss = k σ2 + offset 2

) (

)

2  Loss = $6.00 3.9173 + 59.5 − 59.0   

= $25.0038 per piece For a production run of 50,000 pieces, the total loss would be $1,250,190. This is a $9330 decrease in the loss function from the situation with only the tolerance of A tightened. Since the decrease in the loss function is more than the $9000 cost of tightening the tolerance on D, it is advantageous in the long run to tighten that tolerance. We can do the same for component C — see Table 9.63. The response variance will be 3.8687. If the same 0.5 offset is assumed, the loss function can be calculated as:

SL3151Ch09Frame Page 451 Thursday, September 12, 2002 6:05 PM

Design of Experiments

451

TABLE 9.63 Reduction of Tolerance Limits for Component C Tolerance for Component

Response Difference Between Tol. Limits

Response Production Std. Dev.

Respons Variance

1.76 1.00 1.76 1.20 0.10

0.293 0.160 0.293 0.200 0.020

0.0858 0.0278 0.0858 0.0400 0.0003 0.2397

3.2

0.80

0.6400 2.9890 3.8687

A B C D E

Humidity (H) Error Variance

(

Loss = k σ2 + offset 2

) (

)

2  Loss = $6.00 3.8687 + 59.5 − 59.0   

= $25.1904 per piece For a production run of 50,000 pieces, the total loss would be $1,235,610. This is a $14,580 decrease in the loss function from the situation with only the tolerances of A and D tightened. Since the cost of tightening the tolerance on component C is $15,000, it would not be advantageous to tighten that tolerance. So far, the tolerance design has been entirely a paper exercise based on the tests run during the parameter design and the assumptions about the relationships between the component levels and the response. A set of confirmation runs should be made with the tolerance limits for components A and D tightened. An L8 orthogonal array is used for the confirmation runs with the levels set, test setup, ANOVA table, and level averages as shown in Table 9.64. The test setup and results: A-Tol

B-Tol

C-Tol

D-Tol

E-Tol

H

1 1 1 1 2 2 2 2

1 1 2 2 1 1 2 2

1 1 2 2 2 2 1 1

1 2 1 2 1 2 1 2

1 2 1 2 2 1 2 1

1 2 2 1 1 2 2 1

Test Result 1 2 2 1 2 1 1 2

59.7 57.5 59.5 63.6 59.3 58.6 55.8 59.3

SL3151Ch09Frame Page 452 Thursday, September 12, 2002 6:05 PM

452

Six Sigma and Beyond

TABLE 9.64 L8 OA used for the Confirmation Runs with the Levels Set, Test Setup, ANOVA Table, and Level Averages Tolerance Levels Low Level Column (1) A-Tol. B-Tol. C-Tol. D-Tol. E-Tol. H (Humidity) Unassigned

1 2 3 4 5 6 7

High Level (2)

Nominal

+45 +15 +5 +150 +100 High

1000 400 60.77 2200 1600 —

–45 –15 –5 –150 –100 Low

The average response is 59.2 and the S/N is 28.5. The ANOVA table and level averages for all of the factors from the verification runs are: ANOVA Table Source

df

SS

MS

F Ratio

S’

%

A-Tol. B-Tol. C-Tol. D-Tol. E-Tol. H T Error (pooled error)

1 1 1 1 1* 1 1* — 2

6.661 1.201 9.461 2.761 0.101 13.781 0.551

6.661 1.201 9.461 2.761 0.101 13.781 0.551

20.433 3.684 29.022 8.469

6.335 0.875 9.135 2.435

18.35 2.53 26.46 7.05

42.273

13.455

38.98

0.652 4.931

6.61

7

0.652 34.519

2.282

Total

Level Averages Factor

Level 1

Level 2

A-Tol B-Tol C-Tol D-Tol E-Tol H

60.1 58.8 58.1 58.6 59.3 60.5

58.3 59.6 60.3 59.8 59.1 57.9

As mentioned in the last section, this information cannot be used directly in the loss function since the observed variability may be affected by testing only at the

SL3151Ch09Frame Page 453 Thursday, September 12, 2002 6:05 PM

Design of Experiments

453

TABLE 9.65 Response Variance Tolerance for Factor A B C D E

Humidity (H) Error variance

Response Difference Between Tol. Limits

Response Production Std. Dev.

1.8 0.8 2.2 1.2 0.2

0.30 0.13 0.37 0.20 0.03

0.0900 0.0178 0.1344 0.0400 0.001 0.2833

2.6

0.65

0.4225 0.3260 1.0318

Variance

tolerance limits. The center portions of the distributions are not represented in these tests. Since the change in response is assumed to be a linear increase or decrease across tolerance levels, the loss functions can be easily calculated. The Cpk in production is assumed to be 1.0 or greater for all specified tolerances. The difference between the tolerance limits will be equal to six times the production standard deviation for each component parameter for a Cpk of 1. Since the product response is linearly related to the component parameter level, the difference between the response level averages for the two tolerance limits will equal six times the production response standard deviation. Since the response effect of each component tolerance is additive, the response variance due to each tolerance is additive. In a similar manner, the difference in response between the two humidity levels represents four times the response standard deviation. For this example, the response variance can be calculated as shown in Table 9.65. The response variance will be 1.0318. The loss function can be calculated from the equation:

(

Loss = k σ2 + offset 2

) (

)

2  Loss = $6.001.0318 + 59.5 − 59.0   

= $7.6908 per piece For a production run of 50,000 pieces, the total loss is estimated to be $384,540 compared to $1,274,100 before the tolerance design. This is a $889,560 reduction from the original estimate of the value of the loss function.

SL3151Ch09Frame Page 454 Thursday, September 12, 2002 6:05 PM

454

Six Sigma and Beyond

The reduction is due to three difference elements: 1. The mean was relocated from 59.5 to 59.2. 2. The error variance was reduced from 2.989 to 0.326. 3. The tolerances for components A and D were tightened. Humidity Note that humidity was identified as an important contributor throughout this example. The experimenter should investigate the possibility of controlling humidity to further reduce the loss function. If either the effect of humidity on the design can be minimized or the humidity can be controlled, the loss function could be greatly reduced. Testing Eighty tests were used in the example for the last and present sections. These tests were used as follows: • Determine the target levels — 64 tests. • Confirm the choice of targets — 8 tests. • Determine the tolerances to tighten — 0 tests (based on prior knowledge and simulation). • Confirm the performance with tightened tolerances — 8 tests.

DOE CHECKLIST Action Describe in measurable terms how the present situation deviates from what is desired. Identify the proper people to be involved in the investigation and the leader of the investigation. Obtain agreement from those involved on: Scope of the investigation Other constraints, such as time or resources Obtain agreement on the goal of the investigation. Determine if DOE is appropriate or if other research should be done first. Use brainstorming to determine what factors may be important and which of them could interact. Choose a response and measurement technique that: Relates to the underlying cause and is not a symptom Is measurable Is repeatable Determine the test procedure to be used. Determine which of the factors are controllable and which are noise. Determine the levels to be tested for each factor. Choose the appropriate experimental design for the control and noise factors.

Complete

SL3151Ch09Frame Page 455 Thursday, September 12, 2002 6:05 PM

Design of Experiments

Action Obtain final agreement from all involved parties on the: Goal Test procedure Approach Timing of the work plan Allocation of roles Arrange to obtain appropriate parts, machines and testing facilities. Monitor the testing to assure proper procedures are followed. Use the appropriate techniques to analyze the data. Run confirmatory experiments. Prepare a summary report of the experiment with conclusions and recommendations.

455

Complete

SELECTED BIBLIOGRAPHY Bowker, A.H. and Lieberman, G.J., Engineering Statistics, Prentice-Hall, Inc., Englewood Cliffs, NJ, 1972. Box, G.E.P., Report No. 26, Studies in Quality Improvement: Signal-to-Noise Ratios, Performance Criteria and Transformation, The College of Engineering, University of Wisconsin — Madison, 1987. Box, G.E.P. and Draper, N.R., Empirical Model Building and Response Surfaces, John Wiley & Sons, New York, 1987. Box, G.E.P., Hunter, W.G., and Hunter, J.S., Statistics for Experimenters, John Wiley & Sons, New York, 1978. Brown, R.M. and Burke, M.I., Framing of Design of Experiments (DOE), Proceedings from the American Society for Quality Control 42nd Annual Quality Congress, May 1988. Fleiss, J.L., Statistical Methods for Rates and Proportions, John Wiley & Sons, New York, 1981. Hicks, C.R., Fundamental Concepts in the Design of Experiments, Holt, Rinehart and Winston, New York, 1982. Ishikawa, K., Guide to Quality Control, Asian Productivity Organization, Tokyo, 1983. Kapur, K.C. and Lamberson, L.R., Reliability in Engineering Design, John Wiley & Sons, New York, 1977. Taguchi, G., System of Experimental Design, Volumes 1 and 2, UNIPUB: Kraus International, White Plains, NY, and American Supplier Institute, Dearborn, MI, 1987. Taguchi, G. and Konishi, S., Orthogonal Arrays and Linear Graphs: Tools for Quality Engineering, American Supplier Institute, Dearborn, MI, 1987. Taguchi, G. and Wu, Y., Introduction to Off-Line Quality Control, Central Japan Quality Control Association, Nagaya, Japan, 1979. Wu, Y. and Moore, W.H., Quality Engineering Product and Process Design Optimization, American Supplier Institute, Dearborn, MI, 1985. Japanese Industrial Standard, General Tolerancing Rules for Plastics Dimensions – JIS K 7109–1986, Japanese Standards Association, Tokyo, 1986.

SL3151Ch09Frame Page 456 Thursday, September 12, 2002 6:05 PM

SL3151Ch10Frame Page 457 Thursday, September 12, 2002 6:03 PM

10

Miscellaneous Topics — Methodologies THEORY OF CONSTRAINTS (TOC)

Every organization that wishes to achieve significant improvement with modest capital investment must address five critical questions: 1. What are the key areas within the organization for competitive improvement? 2. What are the key technologies and techniques that will improve these key competitive areas at least cost to the organization? 3. How do these improvement and investment opportunities relate (i.e., how can they be applied in an integrated, supportive and logical manner)? 4. In what sequence should these opportunities be addressed? 5. What are the real financial benefits going to be? Certainly, many other questions must be dealt with, but these are the five issues that frequently cause the most difficulty in industrial and business planning today. So, how can the theory of constraints (TOC) help? Before we answer this, let us examine the fundamental concepts of TOC.

THE GOAL The fundamental goal of any for-profit organization is to make money. In fact, practical experience tells us that the owners/shareholders of such organizations demand this end result performance. However, is this definition of the goal complete? The notion of continuous or ongoing improvement has proven to be extremely powerful in all aspects of life. Therefore, does the application of this notion to our definition of the goal have any important impact? Most organizations strive to improve their money-making performance year after year. So, the goal is really to make more money now and in the future. In this manner, it is impossible to make short-term decisions that bolster short-term profitability while compromising longerterm profitability without violating our goal definition. If the owners are responsible for determining the goal of the organization, what is the role of management? Clearly, management must develop strategies and tactics that are appropriate for achieving the goal. Unlike the goal, these strategies and tactics must be flexible and responsive to changing conditions. The goal of for-profit organizations has been the same for over 1000 years and shows all indications that it will continue in good health.

457

SL3151Ch10Frame Page 458 Thursday, September 12, 2002 6:03 PM

458

Six Sigma and Beyond

Now, before an organization can develop its own customized strategies and tactics, it must first address at least one prerequisite. How would the management team of an organization know whether a particular strategy or tactic was effective (i.e., contributed to making more money)? Some set of measures would have to be used to gauge the degree of success. As a matter of fact, the implementation of a few strategic and/or tactical candidates may not be measurable. These candidates would likely not be selected for actual implementation. So, what should be the high-level measures that lead us to understand the impact of our strategic and tactical efforts on our goal?

STRATEGIC MEASURES For TOC, three measures have become the pillars of the methodology. They are: 1. Throughput (T) — The net rate at which the organization generates and contributes new money, primarily through sales. 2. Investment (I) — The money the organization spends on “stuff,” shortand long-term assets, which can ultimately be converted into T. 3. Operating Expense (OE) — The money the organization spends converting I into T. What do we really mean by these strategic measures, starting with T? T implies that no one within the organization can get a “gold star” or rest easy until the product has been sold. Simply designing or producing the product is not enough. Instead, for anyone’s efforts to count toward the generation of new money, products must not be just designed and produced but sold as well. Therefore, we are not going to allow elements of the organization to play output performance games. No longer will we allow elements of the organization to hide behind inventory profits. With regard to I, TOC combines all the materials (e.g., short-term asset investments in raw materials, work-in-process materials, and unsold finished goods materials) captured within the organization, together with the traditional capital assets (e.g., plant, equipment, land) of the organization. It is easy to visualize how materials are converted into T, but how can these traditional capital assets of the organization be converted into T? The 1980s were known as the decade of acquisitions and takeovers. During this period, many owners sold portions of companies (e.g., plants, divisions) to someone else in return for money. In their eyes, they converted specific capital assets into T. OE seems to possess a similar composition. In OE, we have traditionally thrown direct labor expenses, indirect labor expenses, overhead expenses, sales expenses, and general and administrative expenses into one big pot. Why? We have noticed that over the past 20 or 30 years, the direct correlation between the level of total operating expense of a typical organization versus the level of business it enjoys has gradually eroded. In fact, today it is not uncommon to find organizations whose level of business can fluctuate greatly upwards or downwards, while their true “outof-pocket” operating expense spending hardly budges. Thus, many expenses that we traditionally view as variable (i.e., proportional to level of business activity) are no longer proportional to level of business activity.

SL3151Ch10Frame Page 459 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

NET PROFIT, RETURN

ON INVESTMENT, AND

459

PRODUCTIVITY

At this point, it is not unreasonable to ask the question: “How do T, I, and OE provide clear indications of the impact of our actions on the goal?” Before we can directly answer that question, we should examine T a little more closely. What did we mean in our definition of T, “new money generated and contributed primarily through sales?” Is T simply equal to sales revenue generated for each time period? If you were the proprietor of a small business, say, a dry cleaner, you would not pay yourself a monthly salary. Rather, you would pay yourself in accordance with what money was left at the end of the month after all your expenses were paid (e.g., labor, insurance, raw materials suppliers). As a small business proprietor, you may encounter periods of time where business is so bad that you not only do not make a profit (i.e., no money left over at the end of the month to pay yourself), but you do not have enough money to pay all your expenses. What do you do? Well, you probably carefully ration the money you did generate that month. Which expenses will you pay first? Normally, you are not excited about the prospect of asking your suppliers to wait for their payments. If presented with this situation, they may elect to no longer service you. So, you will probably pay them first. It is interesting to notice that the types of expenses that you would pay first are directly proportional to the level of your business. In other words, the quantity of raw materials and component parts you purchase from your suppliers, together with the level of subcontracting services you buy from your subcontractors, go up when business levels go up and vice versa. Expenses that clearly demonstrate variability proportional to level of business activity can be thought of as the true variable expenses. Once you have paid these true expenses from the money your business generated through sales, you can use the money left to cover your own internal recurring and fixed operating expenses. Therefore, the money left after all these expenses are covered is contributed back to the company to cover its OE. Anything left over at this point is pre-tax net profit. In summary: T = sales – true variable expenses (TVE) Net profit (NP) = T – OE The net profit generated per period in relation to the total investment base of the company employed is an important relative measure. This relative measure is frequently thought of as return on investment and can be summarized as follows: Return on Investment (ROI) = NP/I = (T – OE)/I Finally, how could we strategically measure the productivity of the organization? The traditional definition of productivity compares the value of the output generated to the money spent in the generation process. From the T, I, and OE perspective, this would lead us to the following conclusion:

SL3151Ch10Frame Page 460 Thursday, September 12, 2002 6:03 PM

460

Six Sigma and Beyond

Productivity = T/OE In other words, the productivity of the organization can be described as simply the T dollars generated for each OE dollar spent.

MEASUREMENT FOCUS It has often been said that if you focus on everything, you will end up focusing on nothing. So in this context, which of the three strategic measures (T, I, or OE) should an organization focus its improvement efforts on? Before we answer this question, it may be helpful to determine which measure typical organizations traditionally have focused on. It is our experience that traditional focus is on OE. Why? Here are some popular reasons: • • • • •

OE is perceived to be the easiest to control. It is thought that OE changes can be made quickly. OE is where all of our “people” costs reside. Productivity improvement justifications usually focus on OE reduction. Traditional product costs are partially derived by allocating OE to the per product level.

Now that we have a perspective as to why conventional improvement efforts are commonly focused on OE, is this the proper focal point for the modern, lean, quickresponse enterprise? In order to answer this question, we should first examine the desired improvement trends for T, I, and OE from the perspective of long-term continual improvement: From this perspective, reducing I and OE while increasing T would have positive impact on NP and ROI. However, how far can we go in our efforts to reduce I and OE? From a pure academic perspective, our I and OE improvement efforts are bound by zero, and in reality, we cannot go below some non-zero threshold without driving the enterprise out of business. Also, each step we take closer to our I and OE practical reduction limits, the more energy, effort, resources, and time it usually takes to generate real financial benefits. In other words, long-term I- and OE-focused reduction efforts are prone to the laws of diminishing returns. On the other hand, T appears to have no obstacle in the path of its long-term continual improvement. Thus, it would appear that over the long term, our improvement efforts should be focused primarily on T. Now, does this imply that I and OE are unimportant? Of course not. First, the net profit and ROI relationships rely heavily on I and OE. Therefore, we cannot implement any T improvement opportunity without first understanding its impact on I, OE, and therefore, net profit and ROI. Second, obvious opportunities for real and meaningful I and OE reduction should always be pursued. However, this also implies that our efforts to identify improvement opportunities will not be preoccupied by I and OE. Rather, we will marshal and focus our efforts on identifying and implementing T improvement opportunities. Now, what is the implication of shifting continual improvement focus from the traditional OE perspective to the more modern perspective of T? Another analogy

SL3151Ch10Frame Page 461 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

461

may aid our investigation. Assume you can go to your local optician and purchase a pair of “OE” glasses. When you get back to your company and put these glasses on, you are able to clearly, easily, and rapidly identify and prioritize opportunities for OE reduction, and as a result, you amass a list of OE reduction projects (A, B, C, D, and E). A few days later, you return to that same optician and purchase a pair of “T” prescription glasses. Upon returning to your company, you put on your new T glasses and are able to clearly, easily, and rapidly identify and prioritize opportunities for increasing T. Using your instincts, do you think the list of T projects would be the same or different from OE reduction project list? Over the past five years and thousands of respondents, the answer has unanimously been “different.” In other words, our instincts tell us that perspective is critical in guiding continual improvement efforts. Therefore, viewing the organization “through the eyes of T” may be the most effective method of driving continual improvement within the modern, lean, quick-response enterprise.

THROUGHPUT

VERSUS

COST WORLD

A great deal of attention has been paid to the comparison between cost world and throughput world decision making. Let us take time to simply examine this relationship from the perspective of strategic or organization-wide productivity (T/OE). Generally, the traditional focus on OE reduction generates a consistent degradation in T. However, the rate of OE reduction is greater than the resulting rate of T decrease, thus leading to a misleading improvement in productivity. Obviously, this false sense of security and improvement cannot be sustained for a prolonged period of time. In addition, most organizations are capable of maintaining productivity levels even in the midst of significant T losses. These are classic symptoms of what has become known as the death spiral. On the other hand, in the situation where throughput world T, OE, and T/OE condition in which T will typically increase steadily while OE increases at a slower rate or is held constant, the result is increased organization-wide productivity performance. The attention is clearly being paid to increasing T, even if OE must be increased as well. The best financial arbitrator of such improvement options is incremental ROI.

OBSTACLES

TO

MOVING

INTO THE

THROUGHPUT WORLD

When information is passed from the bottom of the organization to the top, it rarely gets there without significant interpretation and summarization along the way. Some may even say that information rarely gets to the top of the organization without distortion. Likewise, as policies flow down the organization from the top to each lower level, rarely are these policies not interpreted prior to their implementation. This distortion, some may say, also occurs to policies as they cascade down the organization. Why does this distortion occur? Usually, individuals and departments at each level of the organization’s pyramid will interpret information flowing up or policies cascading down to their own best advantage. So, what perspective or framework do they use in determining how to perform this interpretive process? There is

SL3151Ch10Frame Page 462 Thursday, September 12, 2002 6:03 PM

462

Six Sigma and Beyond

an old saying that goes, “Tell me how you will measure me, and I will tell you how I will behave.” Therefore, the manner in which each group within the organization is measured directly impacts that group’s actions. To see how this phenomenon occurs, let us examine a hypothetical organization with sales, finance, manufacturing, product development (PD), and quality departments within that organization: Typical measurable characteristics are: Finance • Cash flow management • Return on assets • Return on investment Sales • Volumes Product development • Development budgets • Development schedules Quality • Defect rates • Returns • Scrap rates • Warrantee claims Manufacturing • On-time delivery What do you notice about these individual department or local measures? First, many of them exist. Second, they are all different. But there is something else. Let us say that I manage manufacturing and my on-time delivery is getting worse. Such local measures as on-time delivery are important to me; they are like my professional report card. So I conclude that in order for me to improve my local measure, I need to add production capacity by purchasing and installing a new piece of equipment. However, when I take my recommendation to the finance people, they veto the proposal not because they are inherently nasty people, but because the proposal will make their return on assets local measures worse. So I rethink my strategy and conclude that I could improve on-time delivery by simply speeding up my production equipment. However, when I do that, the quality manager reacts negatively, to say the least, because the defect rate local measure gets worse. It appears that frequently my efforts to improve those measures for which I am held accountable hurt other local measures for which others in the organization are held accountable. In other words, our traditional local measures are frequently in conflict with one another. In the quick response, lean, modern organization, can such a fundamental conflict be allowed to continue? Most believe it cannot. Can we overcome this situation? The simplest solution would be to measure all departments and functions with the same measures. What measures come to mind first? How about T, I, and OE? Of these fundamental measures, which should take precedence within the quickresponse, lean, modern enterprise? We concluded earlier that T provided the best

SL3151Ch10Frame Page 463 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

463

perspective regarding continual improvement. Therefore, we are not suggesting that we treat each department within the enterprise as an enterprise unto itself focusing on its own T. Instead, we are suggesting that “each local area be measured on its effort and how those efforts contribute directly to improving the global T of the entire enterprise.”

THE FOUNDATION ELEMENTS

OF

TOC

When we view organizational opportunities for improvement through the eyes of T, we are really asking ourselves a simple question, “What limits our ability to improve the T of our organization?” This simple question leads quickly to the conclusion that the typical organization is nothing more than a system composed of tightly interlinked or interdependent subsystem components (e.g., departments and functions). A common analogy describing a system is that of a chain. Each link in the chain represents individual organization departments and/or functions. Each department is dependent upon the succeeding and preceding departments or links. The overall performance of the chain is usually described in terms of tensile strength. Therefore, when the question, “What limits our ability to improve the overall performance of the chain?” is asked, the answer becomes obvious: its weakest link. The concept of an organization as a collection of interdependent subsystem components whose improved performance is based upon its single weakest link is fundamental to systems thinking and is critical to our effort to view the organization through the eyes of T. TOC calls the organization’s weakest link its constraint. The organization’s constraint is that element of the organization that limits its ability to improve performance relative to T. Therefore, the two fundamental concepts of TOC are: 1. View the organization through the eyes of T. 2. Develop a common performance measurement system derived from T. These two elements are related to design for the six sigma (DFSS) methodology because: 1. Design engineering knows the critical features of the products that are processed through the specific manufacturing process A. 2. Manufacturing engineering knows the critical manufacturing process steps performed by the specific manufacturing process A. 3. Integrating the above two insights (the essence of DFSS) frequently allows the joint design/manufacturing engineering team to offload a few noncritical process steps now performed by process A to some other nonconstraint manufacturing process such as process B.

THE THEORY

OF

NON-CONSTRAINTS

Most managers in their everyday life quantify their daily projects in the following prioritization scheme:

SL3151Ch10Frame Page 464 Thursday, September 12, 2002 6:03 PM

464

Six Sigma and Beyond

• Tier I — most important projects to a particular department • Tier II — moderately important projects taking priority after Tier I projects • Tier III — low-priority projects that are scheduled but rarely addressed First, let us examine how these lists would be constructed from the traditional management perspective. What type of projects would make it into Tier I for manufacturing process (MP) #1 department? Obviously, process improvement projects focused primarily, if not exclusively, at MP #1 department. How about MP #3 department? The same. And MP #5, #6, and so on? The same. By looking at the organization’s various lists of Tier I process improvement projects, can we determine where the leverage point of the organization is located? Can we determine what the organization’s key improvement thrust is? No! In fact many, many different issues are key to the organization. It appears that the organization is focusing on everything. Now, from the T perspective, what type of projects would make it into Tier I for MP #1 department? Obviously, process improvement projects focused primarily, if not exclusively, at MP #4 department. How about MP #3 department? The emphasis is also focused primarily on process improvement projects at MP #4 department. And for MP #5, #6, and so on? The same. By looking at the organization’s various lists of Tier I process improvement projects, can we determine where the leverage point of the organization is located? Can we determine what the organization’s key improvement thrust is? Yes! In fact, it appears that the organization has been able to synchronize the process improvement efforts of all of its non-constraint resources from the T perspective. If any logistical system’s T performance is limited by its constraint resource and there can only be one constraint resource within a system at a single point in time, then the rest of the organization’s resources must be non-constraints. Therefore, in terms of sheer numbers, non-constraints dominate the organization. From this perspective coupled with the Tier I process improvement example above, we have discovered that we may have made a mistake naming this methodology the theory of constraints. In reality TOC is not primarily about the poor overworked people in the constraint department. TOC is not primarily about whether the constraint people can invent a 25th hour in the day or an eighth day in the week. Rather, one of the most powerful elements of TOC is the synchronization of the organization’s nonconstraints so as to improve the T performance of the entire system.

THE FIVE-STEP FRAMEWORK

OF

TOC

A great deal of ground has been covered in this section. Let us summarize the lessons derived in this section by listing the five-step implementation process known as The Five-Step Framework of the Theory of Constraints (TOC): 1. Identify the organization’s constraint. 2. Develop plans to exploit the organization’s constraint (e.g., squeeze out as much T performance improvement as possible from the existing constraint resource).

SL3151Ch10Frame Page 465 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

465

3. Subordinate the actions of non-constraint resources to the implementation of step #2 (e.g., ensure that all non-constraint functions support and are synchronized in the implementation efforts of #2). 4. Elevate the organization’s constraint (e.g., augment the constraint). 5. Once the constraint has been broken, go back to Step #1, but beware of organizational inertia (e.g., be sure to clearly communicate to the entire organization that it is time to search for the next constraint). This common-sense process of improvement is neither complex to understand nor difficult to implement. In fact, practical implementation experience tells us that TOC logically integrates many of the traditional tools in the improvement toolbox in an effort to improve the T performance of the entire organization. Such tools as 8D, design of experiments, and work team now become more effective in their application when used through the eyes of T. Finally, the power of TOC in the DFSS process is the understanding of constraint and the action that must be taken to remove the constraint. In essence, TOC is a viable methodology to eliminate the hidden factory from a design perspective.

SELECTED BIBLIOGRAPHY Goldratt, E., Theory of Constraints, North River Press, Inc., Great Barrington, MA, 1990. Goldratt, E. and Cox, J., The Goal, 2nd ed., North River Press, Inc., Great Barrington, MA, 1992. Goldratt, E., Satellite Program: Facilitators Handbook. North River Press, Inc., Great Barrington, MA, 1999. Goldratt, E. Late Night Discussions on the Theory of Constraints, North River Press, Inc., Great Barrington, MA., 1998.

DESIGN REVIEW A typical design review is a process, and it must be (a) multi-phased and (b) involved in the different design phases. In fact, the reviews should also extend to the operations and support phases. It is of paramount importance in any given design review to consider the feedback of customer information because quite often it reveals factors of concern that may have been forgotten or considered too lightly. If design reviews are not taken seriously in the sequential design phases, warranty costs can well exceed any early budgetary considerations. A very typical sequential design review is given in the SAE R&M Guideline (1999, p. 16) shown in Table 10.1. Even though Table 10.1 identifies the core objectives of design review, there is more to it than just a cursory outline of requirements (Stamatis, 2002). For example: System Requirements Review — This is the first review with the customer where the customer specifies the level of cost-effectiveness that the manufacturer is expected to meet. It is at this meeting(s) that customer and manufacturer come to some agreement not only on the reliability and

SL3151Ch10Frame Page 466 Thursday, September 12, 2002 6:03 PM

466

Six Sigma and Beyond

TABLE 10.1 Design Review Objectives Design Phase

Review Objectives

1. Concept

Concept review: Focus on feasibility of proposed design approach Preliminary design review: Validate the capability of the evolving design to meet all technical requirements Build: Address issues resulting from machine build and runoff testing In-plant installation: Conduct failure investigation of problem areas for continuous improvement

2. Development and design

3. Build and install

maintainability (R&M) characteristics but also the adjunct attributes of availability, dependability, and capability. System Initial Design Review — This meeting(s) provides the final definition of the system functional requirements, firming up what was discussed earlier. The allocation of R&M values accompanied by the attributes that support these is usually accomplished at this review. Preliminary Design Review — This is where the configuration items are reviewed and the complexities and technologies are discussed. The objects here are to (1) establish design adequacy and determine risks involved with the proposed design methods and techniques, (2) harmonize the proposed design with the specifications, and (3) ensure the compatibility of the physical and functional characteristics with each other and with the operating and maintenance environments. Resolution of the entire system could be accomplished at this review where there is some finalization on at least “first intentions.” This would include initial sketches and drawings, mockups, simulation models, and prototypes. Interim Design Review — Formality of meetings continue by the manufacturer reviewing progress with the customer. Data feedback on all testing and analysis starts to get more intense at this stage. The customer and supplier review the status of the milestones to ensure being on target. Critical Design Review — By this stage the design should be firm enough to lock in the design parameters and make preparations for the qualification testing. The customer by all means should be present at this design review to assure that all particulars of the contract are agreed upon. Formal Qualification Review — This is the final review to precede full-scale production. The qualification models were fabricated and assembled with production tooling, but at this stage customer and manufacturer should be ready for the production line. At this meeting the review team confirms that the total package is as agreed upon and that it meets all the terms of the contract.

SL3151Ch10Frame Page 467 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

467

In-house Reviews — In the event problems occur after production starts, additional reviews may be necessary. Most R&M texts do not discuss this type of review because theoretically the design is “frozen” and the only problems, if there are problems, are production types. But in the real world anomalies do occur, and they have to be addressed. The design review concept should continue until the customer is totally convinced that that the product is of high quality. Many questions arise in the course of design review meetings. Some organizations have a checklist they use to ensure that “nothing falls through the cracks” and that they cover all the potential problems that may occur. A generic checklist, which can be used at design review meetings or by the designer to ensure the integrity of the design, is shown as Table 10.2. For obvious reasons, this list is not an exhaustive list to cover all situations. Rather, it is a list that may be modified and act as a catalyst for further discussion. Here, we must give the reader a very strong caution. Concurrent or simultaneous engineering is not a design review function, but it is closely related to it. Concurrent engineering is the process by which all disciplines that design, manufacture, inspect, sell, use, and maintain the product work together to develop and produce it. Milestones are established where the various disciplines accomplish specific tasks simultaneously before proceeding on to the next task. Traditionally, each step in the design process occurred one step at a time and extended over a relatively long duration. In concurrent engineering, several tasks such as manufacturing engineering work with the quality engineer and design to set up their responsibilities while the designer is still working with the design. A comparison between traditional and concurrent engineering is shown in Table 10.3.

FAILURE MODE

AND

EFFECT ANALYSIS (FMEA)

Even though the FMEA is discussed in Chapter 6, it is important to discuss here the relationship between FMEA and design review. FMEA, as we already have said, is a methodology that helps in identifying potential and known failure modes and then arriving at a probability of occurrence and detection. In addition, a good FMEA should recognize interfacing failures in components, subsystems, and systems themselves. Typical interfacing problems may occur in the form of proximity, information, material, and energy transfer. In all cases, the ultimate goal of any FMEA (Concept/System, Design, Process, or Machinery) is to reduce as many failure modes as possible, or to reduce the probability of the failure modes as much as possible. The process of doing an FMEA is a down-up approach. For a very detailed explanation, see Stamatis (1995). An FMEA in design review should not be confused with a Fault Tree Analysis (FTA). Whereas the FMEA is a down-up approach, the FTA is a top-down approach to failure analysis starting with the undesirable “top event” and progressively determining all the ways the failure may happen. The analysis starts early in the concept phase and is continually monitored through the subsequent stages. A comparison between FMEA and FTA is given below:

SL3151Ch10Frame Page 468 Thursday, September 12, 2002 6:03 PM

468

Six Sigma and Beyond

TABLE 10.2 Design Review Checklist • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Do specified components meet their reliability requirements? Can off-the-shelf items be used for particular functions? Does the design meet functional requirements? Were standard components and assemblies considered and used where possible? Were all environmental impacts considered? Did all components pass the environmental testing? Were corrections made where necessary? Have critical characteristics been considered? Have failure histories been investigated? Did each component and material meet its requirements under environmental extremes of the specification? Were there enough data for reliability calculations? Was the complete unit tested? Were the weak links in the design corrected? Does demonstrated reliability meet required specification or is redesign indicated? Are predicted and allocated reliabilities compatible? Are trade-offs necessary? Were manufacturing and quality assurance (QA) considered in the design? Is redundancy necessary to meet reliability requirements? Can the environment be changed or be protected? Is heating, cooling, shock mounting, shielding, or better insulation required? Were all failure modes corrected to prevent recurrence? Has storage capability been studied? Should specifications be written to ensure 100% test and inspection? Are there suitable manufacturing and QA procedures to ensure good quality? Is the item designed as simply as possible? Have all human factors been considered? Have sharp corners been considered? Have all potential stress risers been eliminated? Can other cognizant disciplines help in writing specifications or offer improvements in the design? Can suppliers provide reliability values for their components? If so, are the values compatible with the overall system? Has enough testing been performed to validate the required reliability for designated components? Will early testing or screening help eliminate infant mortality type failures? Have maintainability requirements been considered?

TABLE 10.3 Comparison Between Traditional and Concurrent Engineering Function

Traditional Engineering

Concurrent Engineering

Organization

Engineering is unique and separate from other departments Each department waits for output from previous department

Engineering is part of multifunctional team with team objectives Tasks progress simultaneously working with engineering

Timing of outputs and inputs

SL3151Ch10Frame Page 469 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

An FMEA starts with the origin of a failure mode and then proceeds to find the causes, probabilities, and corrective actions. An FMEA considers all potential failure modes that can be produced through mental generation. An FMEA uses an inductive approach. Less engineering is needed for an FMEA. There is limited safety assessment in an FMEA. The failure paths are not delineated in an FMEA. An FMEA looks at each failure mode separately.

469

An FTA starts with an accident or undesirable failure event, determines the causes, then the origin of the causes, then what can be accomplished to avoid the failure. An FTA studies only negative outcomes that warrant further analysis. The FTA uses a deductive approach. Skilled personnel are required for an FTA. The FTA manages risk assessment and safety concerns. The FTA provides a good assessment of the failure paths, and the control points are well enhanced. An FTA demonstrates a more selective method of showing the relationship among events that interact with one another.

The steps of the FMEA are: 1. Define the team. 2. Define the system, design, process, or machinery (block diagram, process flow diagram). 3. Construct the P-diagram. 4. Define the function. 5. Identify the failure(s). 6. Complete the tabular form. 7. Analyze (evaluate) the completed FMEA. 8. Recommend any corrections for design changes. 9. Document the analysis and its results. To help in the generation of FTA, the Rome Air Development Center (RADC) has formulated a seven-step approach. These steps obviously are very generic, and each organization that is using them must modify them to fit their purpose. The seven steps are: 1. Define the system, ground rules, and any assumptions to be used in the analysis. 2. Develop a simple block diagram of the system showing inputs, outputs, and interfaces. 3. Define the top event (ultimate failure effect or undesirable event) of interest. 4. Construct the fault tree for the top event using the rules of formal logic. 5. Analyze the completed fault tree. 6. Recommend any corrections for design changes. 7. Document the analysis and its results. FTA is one of the few tools that can depict the interaction of many factors and manage to consider the event that would trigger the failure or undesirable event.

SL3151Ch10Frame Page 470 Thursday, September 12, 2002 6:03 PM

470

Six Sigma and Beyond

Designers should consider the safety aspects of their designs early in the concept stage so that necessary changes can be accomplished before they have a chance of occurring.

REFERENCES Anon. Reliability and maintainability guideline for manufacturing machinery and equipment. M-110.2. 2nd ed. Society of Automotive Engineers, Warrendale, PA. and National Center for Manufacturing Sciences, Ann Arbor, MI, 1999. Stamatis, D.H. Guidelines for Six Sigma design review — Part 1. Quality Digest. pp. 27–31. Part 2. pp. 26–30, April and May 2002. Stamatis, D.H. Failure Mode and Effect Analysis: From Theory to Execution. Quality Press, Milwaukee, WI, 1995.

SELECTED BIBLIOGRAPHY Hu, J. et al., Role of Failure Mechanism Identification in Accelerated Testing, Journal of the IES. July/Aug. 1993, 39–45. Keceioglu, D., Reliability and Life Testing Handbook, vols. 1 and 2, Prentice-Hall, Upper Saddle River, NJ, 1993. King, J., New Directions for Reliability, QualityEngineering-1, 79–89, 1988. O’Connor, P., Practical Reliability Engineering, John Wiley & Sons, New York, 1991. Phadke, M.S., Quality Engineering Using Robust Design, Prentice-Hall, Upper Saddle River, NJ, 1989.

TRADE-OFF STUDIES Trade-off studies are designed for balancing both business and technical issues and optimizing the product for the customer, whether internal or external to the organization. Trade-off studies: • Are a structured, analytical method for objectively identifying, defining, and evaluating alternatives • Are designed for analytically presenting, evaluating, and weighting decision information based upon program targets, objectives, goals, and technical requirements • Ensure that the selected alternative is the best at meeting the program objectives, goals and technical requirements We conduct them to: • Promote an objective evaluation and minimize subjective selection • Force requirements to drive the evaluation of the alternative • Ensure that sufficient information for making a decision is provided

SL3151Ch10Frame Page 471 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

471

• Demonstrate that the alternatives satisfy the requirements (as they are understood at the time of the evaluation) • Document the evaluation Of course, the question quite often is, “How do I know if I need to conduct a trade-off study?” The answer depends on the following questions: • Does the decision require input and concurrence from several organizations? • Does the decision require balancing inputs that may conflict and/or are inversely related? • Is there a choice between several viable/acceptable alternatives? • Is a quick, comprehensive, and defensible decision needed? If the answer is yes to one or more of these questions, a trade-off study may be the best approach for selecting the optimum solution. To conduct the trade-off study appropriately, there are some preliminary requirements. They are: • Prior to declaring or encountering a critical design freeze • When balancing major systems or their components’ functional performance • When considering several design alternatives at any level (e.g., systems, subsystems, and components and so on). • When conflict exists among targets, objectives, and requirements (e.g., maximizing one has negative effects on the others) • When establishing dominant attributes or prioritizing customer requirements As with any other methodology and tool, with a trade-off study we expect to have some kind of deliverables at the end of the analysis. Typical deliverables are: • An alternative selected largely on the basis of fact, which is acceptable to and defensible by all the stakeholders • Complete documentation (Evidence Book) outlining how and why the decision was made • A risk list identifying areas of concern for all the alternatives investigated • Sensitivity analysis showing the stability of the selected alternative

HOW

TO

CONDUCT

A

TRADE-OFF STUDY: THE PROCESS

Several steps are required when conducting a complete trade-off study. Each step aids in ensuring that the end decision best meets the stated customer requirements. Each step is discussed in detail below. A checklist is also provided to assist you in conducting your trade-off study. The steps are: 1. Construct the Preliminary Matrix The trade-off study matrix consists of two major components — the alternative list and the category list:

SL3151Ch10Frame Page 472 Thursday, September 12, 2002 6:03 PM

472

Six Sigma and Beyond

The alternative list: This list is simply a listing of each alternative being considered. The alternatives are listed across the top of the trade-off study matrix, with one alternative per column. The category list: The category list consists of musts and wants arranged by assessment items. Each assessment item is broken down into various measurable discriminators. An example: Attribute category: safety Assessment item: frontal impact Discriminator: 5 mph bumpers; bumper material; crumple space The first step in the trade-off study process is creating a preliminary matrix. You must identify both the alternatives being examined and the list of assessment items and discriminators. Draw assessment items and discriminators from program requirements, corporate data, QFD studies, CAE analysis, etc. The preliminary matrix acts as a discussion catalyst at the first team meeting. After assembling the assessment list, sort it into those that are imperatives (or musts) and those that are desirables (or wants). 2. Select and Assemble the Cross-Functional Team The goal of this step is to ensure that affected parties are adequately represented. It is better for a group to decline participation than to be overlooked in the team assembly process. The team is composed of representatives from each group impacted by the decision being made (it must be cross-functional and multidisciplinary). Team size varies depending on the subject and scope of the project (initial meetings should include no more than nine to twelve people). Team membership is based on contribution potential not approval needs. Approval takes place during the presentation of results at the end of the process. 3. Assign Team Members’ Roles and Responsibilities Although all team members will play a role in the trade-off study process, three key positions must be filled to ensure process success, as follows: Team champion Is usually a program manager, project manager, or someone empowered by those individuals to carry out the selection of an alternative Is the individual who must design, build, or approve the selected alternative Supports and participates in the process, and accepts (and backs) the team’s consensus decision Provides the resources to accomplish the task at hand Lead facilitator Guards against duplicating efforts, provides information to individuals between meetings, generally coordinates the entire process

SL3151Ch10Frame Page 473 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

473

Is an empowered member of the team whose role is to coordinate the work of the ranking process Resolves any overlaps in evaluation that occur during the ranking process Coordinates the open issues identified by the team Compiles the ranking process data/results into the evidence book Process coordinator Ensures that all affected parties are represented and that each team member understands the process Is responsible for aiding the team leader in assembling the team Schedules and runs all core team meetings Provides and explains the trade-off study methodology and supporting tools 4. Assign Ranking Teams To Evaluate the Alternatives Ranking teams are designed to evaluate each alternative within a particular category or assessment item. Individuals are assigned to these teams based on their particular specialty. For example, a transmission design engineer would be assigned to the task of assessing (or ranking) an alternative’s ability to handle a particular level of input torque; an individual from marketing would determine an alternative’s potential volumes; a financial expert may be assigned to develop marginal costs, and so on. After the core team modifies and approves the preliminary matrix, determine the personnel necessary for conducting the ranking within each category. Assign team members to ranking teams based on their specialty or to the category that affects their area/product. Due to the critical aspect of the ranking teams, team members require specific direction on: 1. Ranking/evaluation methodologies 2. Documentation format and content 3. Reporting their findings, conclusions, and issues Identification of Ranking Methods The ranking team’s primary function is ranking each alternative’s ability to achieve the discriminator (for example, the alternative providing the best crash test gives the highest rank). Measurable discriminators are a must, and the ranking team must devise a method for determining each alternative’s ability to meet that discriminator (for example, is mileage, bumper material, or crumple space the discriminator with the best measurability and effectiveness for our expected result). This ranking is accomplished as follows: As each ranking team delves into the ranking process, the team members may expand or contract the discriminator list to better evaluate the alternative’s performance for a given assessment item. Each team selects the ranking method it feels is appropriate and is defensible to the core team. Whenever possible, the rankings should be based on actual test results, CAE analysis, or numerical analysis (i.e., facts or directly observable data).

SL3151Ch10Frame Page 474 Thursday, September 12, 2002 6:03 PM

474

Six Sigma and Beyond

Expert opinion and subjective rankings should be used as a last resort when time, cost, knowledge, or value limit the reliance on tests, rigorous analysis, or computer simulation. An alternative that fails in achieving any assessment item defined as a project must is eliminated from the evaluation. The ranking team informs the lead facilitator who stops further evaluation of that alternative by all other ranking teams (this is done to limit resource expenditure). When ranking the alternatives, the one (or more) alternatives best satisfying a particular discriminator get the highest numerical rank for that discriminator. This is usually done by counting the number of alternatives and using that value as the highest rank (i.e. with four alternatives, the best would receive a rank of four, the next a three, and so on). Alternatives do not have to be forced ranked, nor does one have to receive the top score. If all or some of the alternatives have equal ability to satisfy the discriminator, they would receive equal rank. If that ability is high, they would all receive a top rank; if that ability is poor, they all may receive a low rank Development of Standardized Documentation A secondary result of a trade-off study is the evidence book documenting the entire decision-making process. Although secondary, this activity needs to be taken very seriously in order to defend the core team’s decision to those both inside and outside the group. Documentation being produced by the ranking teams must be consistent in format and content to ensure ease in assembling the evidence book. To ensure this: Provide each ranking team a standard format for reporting their findings. Make sure that the format includes, at a minimum: The ranking results The ranking method Advantages and disadvantages of each alternative (as defined by the ranking teams) Any risks associated with the selection of each alternative Any issues identified during the ranking process Include all supporting documentation generated during the ranking process in an appendix, attachment, or separately defined section. Timing for Report out of Selection Process Once the members of a ranking team complete the ranking process, they forward their completed documentation to the lead facilitator to place in the evidence book. Each team is then responsible for preparing a presentation of its findings to the entire core team. The presentation includes a summary of the rankings, methods, issues, and risks associated with the selected alternative. This presentation is then made to the entire core team at the final core meeting. 5. Weight the Various Categories While the ranking teams proceed with the ranking process, the process coordinator and team leader pull together the necessary or key personnel to assign weightings to the various assessment items.

SL3151Ch10Frame Page 475 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

475

The weighting rule: Assign weightings according to the assessment item’s importance or impact on satisfying the customer and company needs/requirements and ensuring the optimum decision (for this point in the program). With key personnel developing the weightings in parallel to the ranking process, we are able to: Assign the weightings at a higher corporate level assuring better alignment with corporate vision. Work more quickly and efficiently toward balancing the weightings. 6. Compile the Evidence Book Now that the weightings have been assigned, the next step is for the lead facilitator to compile the evidence book. There are several steps to this process: Organize the ranking teams’ documentation in category sequence as it appears on the trade-off study matrix. Calculate each alternative’s score within the assessment item by determining an assessment item average. Simply sum an alternative’s rank for the various discriminators and divide by the number of discriminators. Continue calculating alternatives’ scores as additional ranking teams report out. Once each ranking team has reported out, the lead facilitator develops a summary of the evidence book for distribution during the final presentation. This summary contains: The completed trade-off study matrix, including tallied alternative scores A section outlining the identified advantages and disadvantages of each alternative Identified risks associated with each alternative 7. Present the Results When each ranking team has reported its results to the lead facilitator, the process coordinator reassembles the core team for a presentation of the results. Copies of the summary document are distributed to each core team member three to five days prior to the meeting. Each ranking team then presents its findings, methods, and issues to the entire team. Sensitivity Analysis: The purpose of the sensitivity analysis is to determine the robustness of the selected alternative. The process allows the group to ask various “what if ” questions regarding a particular ranking or weighting and receive an immediate answer, such as: “What if ” the ranking of that assessment category were inverted, would the alternative still be chosen? “What if ” the weighing of that assessment item were lowered, would it change the selection?

SL3151Ch10Frame Page 476 Thursday, September 12, 2002 6:03 PM

476

Six Sigma and Beyond

It is recommended that a laptop computer, loaded with the trade-off study matrix, be brought to the presentation by the lead facilitator. Modifications can be made to the rankings within an assessment item or weightings on a category to see how that would affect the overall decision. This will identify how sensitive the decision is to certain changes and give the group an immediate feel for the selected alternative’s robustness. Here is a typical trade-off study process checklist: 1. Constructing the preliminary matrix • Consideration has been given to all attribute categories. • All discriminators are measurable, now. • Assessment items considered “musts” are truly “musts.” • Assessment items considered “wants” are only “wants.” 2. Selection and assembly of the cross-functional team • All affected activities have been invited to participate. • All participants are empowered by their management. 3. Assigning team members’ roles and responsibilitie. • Team champion will design, build, or approve the selection. • Lead facilitator is in a position to coordinate the ranking teams. • Process coordinator is willing to schedule and run meetings. 4. Assigning team members to ranking teams • Ranking teams have the right specialists to accomplish their task. • Standardized report format has been established and agreed upon. • Acceptable ranking methodologies have been agreed upon. • Freedom in expanding/contracting the discriminator list has been conveyed. • Timing and report out procedures are understood by each team. 5. Weightings of the various categories • Identification of key personnel is complete. • Weightings are being conducted in parallel to ranking teams’ evaluation. • Assigned weightings align with customer and corporate wants. 6. Compilation of evidence book • Ranking team documentation organized to trade-off study matrix • Alternative scores calculated (weight × rank = score) • Trade-off study summary completed, for final presentation 7. Presentation of results • Entire core team reassembled • Laptop with trade-off study matrix available for sensitivity analysis • Consensus decision reached as to which alternative to pursue

GLOSSARY

OF

TERMS

Alternative rank — Defines how well an alternative compares to other alternatives in achieving a particular assessment item or discriminator. Assessment item — A particular attribute of a given category.

SL3151Ch10Frame Page 477 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

477

Categories — Major classes of assessment items Discriminator — A specific portion of an attribute. Evidence book — The book containing the complete set of documentation created by the core team, ranking teams, and weighting teams during their evaluation of the alternatives. Musts and wants — Each assessment item is grouped according to whether it is a must or a want. Musts are defined as those items that an alternative has to meet in order for it to garner further consideration. When an alternative does not meet a must, it is dropped from the study, unless it can be brought in line with the must or the must is modified. Wants are those items that are needed to reach maximum customer and corporate satisfaction, but an alternative would not be discarded for failing to meet them. These wants are weighted and determine which alternative gets selected from those that meet all of the musts. Ranking team — The sub-core group/team that is assembled to evaluate (and rank) the various alternatives within a particular assessment item. This team also identifies risks, advantages, disadvantages, and issues encountered during the alternative evaluation. Trade-off study matrix — A tabular chart used to list alternatives being evaluated (across the top) and assessment items being used to differentiate the alternatives (down the left-hand side). Alternative rankings, assessment item weightings and alternatives score are also tallied on the matrix. (A matrix can easily be created using any spreadsheet program.) Weightings — A value given to an assessment item, categorized as a want, to show its relative importance to the other assessment items. Techniques used to develop weightings include pair-wise comparison, 100% weightings (sum of weightings add to 100), and many others.

SELECTED BIBLIOGRAPHY Bain, L., Statistical Analysis of Reliability and Life Testing Models: Theory and Methods, Marcel Dekker, New York, 1991. Hubka, V., Principles of Engineering Design, Butterworth Scientific, London, 1982. Kapur, K.C. and Lamberson, L.R., Reliability in Engineering Design, John Wiley & Sons, New York, 1977.

COST OF QUALITY The purpose of costs in quality is to establish “the method” for collecting, maintaining, and using quality cost data so that they become the conscience (the driving force) of the organization for continual improvement. Once this conscience has been realized, then a real effort is put in place in the area of quality improvement opportunity (QIO) for quality audit, product procedures process, and the overall system.

SL3151Ch10Frame Page 478 Thursday, September 12, 2002 6:03 PM

478

Six Sigma and Beyond

Scrap or rework Suppliers Reject Receiving CQ

Inputs

Production Process

Final product QC

Output Accept

Process QC

Customer reports Technology costs

Quality management

Contracts Standards Drawings

FIGURE 10.1 Quality cost: The quality control system.

This is based on the notion that “quality” is defined as satisfying the customer’s needs. How does the cost of quality/quality improvement opportunities satisfy the customer’s needs through the manufacturing organization? See Figures 10.1 and 10.2.

COST MONITORING SYSTEM An organization is expected to use methods that accurately monitor all cost elements. The cost of doing business (labor, materials, overhead) cannot be effectively controlled without a systematic method that effectively monitors how costs are incurred. Specifically, the organization should develop costs relative to quality that will serve as a guide for measuring plant efficiency. Standard Cost The organization should develop a method that will allow for efficiency in labor (direct and indirect); identification of material content (parts and components); appropriate measures of overhead; and documentation/development, review, and revision of these standards. Actual Costs A supplier should maintain an accurate system to record, monitor, and control labor, material and overhead costs. For example: • Compute labor efficiency reports for a period. • Establish a tolerance limit for efficiency. • Ensure that reports are received by management.

SL3151Ch10Frame Page 479 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

479

COST

Continual Improvement EXPECTATIONS

Cost Monitoring System

Cost Reduction Efforts

Standard Cost

Continual Cost Improvement Efforts

Actual Cost

Competitive Product Development

Variance

Cost Estimates

FIGURE 10.2 Costs.

• Generate monthly summary reports. • Ensure that raw materials are ordered in economical quantities to reduce and/or control the cost of material. • Develop a system that will track materials on hand. • Budget for overtime. • Charge premiums to the applicable departments as overhead. • Charge straight time portions as direct labor.

SL3151Ch10Frame Page 480 Thursday, September 12, 2002 6:03 PM

480

Six Sigma and Beyond

• Perform surveys to accurately analyze and distribute indirect inventory overhead and service department costs. • Provide adequate records for support. Variance A supplier should be able to identify and control cost variances. For example: Cost comparison report: Reviewed by the responsible departments in predefined cycles weekly, monthly, etc. Tolerance limit: A system to address the variance and provide for appropriate action. Plant report: A regular reporting of costs and variance published and reviewed by management. Level of comparison: A specific level appropriate for the commodity being produced (part number, department, cost center, etc.). Cost Reduction Efforts The organization and the supplier should cooperate fully with each other in an effort to reduce costs. Continuous efforts to reduce costs and therefore selling price are essential for the organization and the supplier, if both are to remain competitive in the market place. For a supplier to reduce costs, the efforts should be directed in the following areas: 1. Continual improvement program 2. Competitive product development and target cost achievement 3. Cost estimates

CONCEPTS

OF

QUALITY COSTS

All the gurus of quality have identified the costs as an essential part of overall quality improvement. In a summary format let us see some of them: J. Juran Among other concerns, Juran emphasizes that 1. Quality is an issue of cost. 2. To control cost, management must be equipped with experience and training. 3. Quality cost must become a part of the strategic business plan of the organization. W.E. Deming Perhaps one of the most prolific gurus in quality issues of the 20th century, Deming spent a lifetime explaining the issue of cost as one of the driving forces in a dynamic

SL3151Ch10Frame Page 481 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

481

organization by always explaining the need to end the practice of awarding business on the basis of price tag, eliminate numerical goals, and eliminate work standards. P. Crosby Crosby was by far the best salesperson of quality. He was the first to associate the bottom line with the effect of costs. He made a point of differentiating quality of conformance and nonconformance, to quantify the waste of poor quality as a per cent of sales, push for the concept of zero defects and the attributes of prevention quality as opposed to appraisal. G. Taguchi Taguchi’s contribution to the cost of quality is with the tolerance design cost accountability on specifications setting and the loss function.

DEFINITION

OF

QUALITY COMPONENTS

Depending on whom you listen to, approximately 6 to 15% of all quality problems are related to special causes (labor). The other 85 to 94% arise from faults in the company’s system. This larger percentage will continue until management changes the system. Both special and local issues are contributors to the cost of quality (CQ). Two questions follow: What is really meant by “quality cost”? What are the steps of quantifying the costs? 1. There are two major categories of quality costs. a. Inputs b. Outputs The inputs are made up of the appraisal costs, which are the costs incurred (first time through), to discover the condition of the product. These include • Incoming material inspection • Inspection and test • Maintaining accuracy of test equipment • Materials and services consumed • Evaluation of stocks • Product quality audits Another component of the inputs to quality costs is the prevention cost. These are costs incurred to keep output and appraisal costs to a minimum. They include: • Quality planning • New products review • Training • Process control • Quality data requisition and analysis

SL3151Ch10Frame Page 482 Thursday, September 12, 2002 6:03 PM

482

Six Sigma and Beyond

• Quality reporting • Improvement projects • Other prevention costs (general office expenses) The outputs are made up of the internal failure costs. These result when quality issues are discovered outside an organization by the customer. They include • Scrap • Rework • Retest • Downtime • Yield losses • Disposition • Failure analysis • Fault of suppliers Another component of the outputs is the external failure costs, which are • Complaint adjustment • Returned material • Warranty changes • Allowances • Repair • Errors • Liability 2. The use of cost of quality can be quantified by giving attention, prioritizing, justifying, recognizing, and driving decision making deeper into the organization. Screening the costs throughout the organization occurs through: a. Analyzing the ingredients of established accounts b. Resorting to basic accounting documents c. Creating records for documentation d. Estimating costs using statistical tools The analysis of these four categories can be performed by various means, i.e., through descriptive statistics, graphical techniques, or advanced statistical analyses. Here we must emphasize that cost of quality is not a system that encourages, fortifies, or perpetrates the adversary position of one department against another or one company against another. Rather, it is a system that allows management to look at a specific situation compared against itself over time. The ultimate in this thinking is planning for growth. The relationship among CQ, planning, growth, and quality is heavily dependent upon management’s attitude and the employer involvement improvement opportunity of a given organization. The underlying assumption of this concept is that as one controls quality, one reduces cost and this increases profit. This assumption of CQ is important because it becomes the catalyst that causes management to address the issue of quality. Profit is the universal language of all management. The question becomes: “What does CQ provide to management that serves as a significant indicator to them?” It provides:

SL3151Ch10Frame Page 483 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

483

• A systematic method of assessing the overall effectiveness of the quality program • A means of establishing programs to meet overall needs • A method of determining problem areas and action profiles • A technique to determine the optimum amount of effort for each of the various quality activities

METHODS

OF

MEASURING QUALITY

The operating quality costs of prevention and appraisal are considered to be controllable quality costs, while the internal and external failure costs are uncontrollable. Juran has demonstrated the relationship between the controllable and uncontrollable QC curves and the direct quality cost curve over time. As the controllable costs of prevention and appraisal increase, the uncontrollable costs of internal and external failure decrease. The point where the cost of preventing and appraising exceeds the cost of correcting the product failure is the optimum operating quality cost. Mathematically, the optimization is: Let f(q) = total (internal and external) failure costs p(q) = total (appraised and prevention) prevention costs T(q) = total quality costs = f(q) + p(q) q = quality level (0 to 100% good product) T(q) = dT/dq = 0 or dp/dq = df/dq, which is the minimum This means that an additional dollar invested in prevention will produce exactly one dollar’s worth of reduced failure costs. Below the optimum it provides more than one dollar and above the optimum the opposite is true. Therefore: 1. Optimum quality depends on incremental not total elementary costs. 2. There is nothing that demands the optimum be at q = 100%. There might be a minimum rather than an optimum, and it could very well be at q = 100%. The optimum (minimum) quality cost could lie at zero defects, q = 100%, if the incremental cost of approaching zero defects is less than the incremental return from the resulting improvement. Juran asserts that prevention costs rise asymptotically, becoming infinite at 100% conformance. This implies that the incremental cost is also infinite. Since the incremental return is not, it follows from his assertion and the above mathematics that the optimum lies below 100%. The question is: Does it really take infinite investment to reach zero defects? For Crosby, on the other hand, the cost of quality bases are: 1. 2. 3. 4. 5.

Total contract sales Total cost Manpower Manpower by skill Budgeted costs

SL3151Ch10Frame Page 484 Thursday, September 12, 2002 6:03 PM

484

Six Sigma and Beyond

6. 7. 8. 9. 10.

Income after taxes Operating profit Equity earnings Strategic managed cost Constant dollars

COMPLAINT INDICES The user costs associated with failures can be grouped into five categories. R = repair cost E = effectiveness loss (idle labor) C = extra capacity required because of product downtime D = damage caused by failure L = lost income (profit) If these costs are measured each year over the life of the product, then the failure cost (Cf) is n

Cf =

∑ (L + i) j ( R + E + C + D + L ) L

j

j

j

j

j

j =1

where n = life of the product and i = the yearly interest rate.

PROCESSING 1. 2. 3. 4. 5.

AND

RESOLUTION

OF

CUSTOMER COMPLAINTS

Satisfying the complaint Preventing a recurrence of isolated complaints Pareto analysis An in-depth analysis of the vital few Further statistical analysis

TECHNIQUES

FOR

ANALYZING DATA

1. The seven tools of total quality costs a. Pareto chart b. Course and effect diagram c. Stratification chart d. Check sheet e. Histogram f. Scatter plot g. Graphs and control charts

SL3151Ch10Frame Page 485 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

2. 3. 4. 5. 6. 7.

485

Defect matrices Cost analysis Spare parts use growth curves Probability paper Simulation studies Statistical modeling a. Abnormality control chart. The abnormality chart is a chart that addresses the following questions/concerns: i. How did it happen? Date: Place: Lot number: Item name: Found By: Description: ii. How it was found? iii. Emergency measure taken iv. Investigation of the causes v. Cause(s) vi. Measures taken to prevent recurrence vii. How will these measures affect similar processes? viii. How to proceed next?

FORMAT

FOR

PRESENTATION

OF

COSTS

Type of Standard

Standard is Based on

Engineered

Studies made by engineers, e.g., material usage, labor hours

Historical

Statistical computation of past performance Market studies to discover performance of competitors Broad program of final results needed and allocation to subprograms, e.g., reliability goals

Market Planned

Managers Use the Report to Answer the Question Are we attaining the results that the engineering studies showed were obtainable? Are we getting better or worse? Were do we stand compared to our competitors? Are we going to be able to attain the overall planned goal?

This format must serve as a catalyst to management to provide attention, prioritization, justifications, recognition, and corrective action.

LAWS 1. 2. 3. 4. 5.

OF

COST

OF

QUALITY

We cannot reduce cost without affecting quality. We can improve quality without increasing cost. We can reduce cost by improving quality. Cost of quality drives the system. If quality costs money, do not do it.

SL3151Ch10Frame Page 486 Thursday, September 12, 2002 6:03 PM

486

Six Sigma and Beyond

TABLE 10.4 Typical Monthly Quality Cost Report (Values in Thousands of Dollars) October Actual

Variance

18.3 4.6 2.6 7.3 2.4 35.2 7.7%

3.2 06 0.9 2.1 3.4 10.2

9.6 32.5 14.1 1.4 4.1 61.7 13.5%

1.8 15.4 27.4 1.1 1.6 9.7

14 6 197.2 25.2 6.8 14.1 0.8 258.7 56.4% 8.6 41.8 25.6 21.9 4.9 0.0 102.8 22.4% 458.4 6.5 8.8 16.7

9.6 124.3 8.1 2.3 6.6 0.2 129.9

1.6 1.2 0.3 27.0 4.0 0.0 30.3 79.7

Year to Date Category A. Prevention Cost 1. Quality engineering 2. Design and development 3. Quality planning by others 4. Quality training 5. Other Total prevention cost % of total quality cost B. Appraisal Cost 1. Inspect and test incoming materials 2. Inspection and test 3. Product quality audits 4. Materials and services consumed 5. Equipment calibration and maintenance Total appraisal cost % of total quality cost C. Internal Failure Cost 1. Scrap 2. Rework 3. Failure analysis 4. Reinspection 5. Fault of supplier 6. Downgrading Total internal cost % of total quality cost D. External Failure Cost 1. Complaints 2. Rejected and returned 3. Repair 4. Warranty Charges 5. Errors 6. Liability Total external cost % of total quality cost Total operating cost Measurement Bases 1. Direct labor ($/man-hour) 2. Sales (%) 3. Manufacturing Costs (%)

Actual

Variance

190.1 61 8 20 7 46.8 312 350 6 9.4%

10.1 7.5 73 20.3 25.0 55 2

87.3 323.0 140.9 16.5 23.4 591.1 15.9%

7.1 105.0 269.7 8.8 00 166.4

50.0 1305.6 185.1 88.0 152.1 8.1 1788.9 48.1 75.3 403.6 256.5 226.6 28.5 0.0 990.5 26.6% 3721.1 5.3 9.0 16.3

8.0 557 6 0.4 3.0 77.2 19 621.5

5.3 26.4 3.5 263.4 10.2 0.0 291.2 108.7

SL3151Ch10Frame Page 487 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

487

DATA SOURCES Typical sources for cost of quality (CQ) are: Types of Data

Sources

Field performance data Trend of sales by model customer, etc. Trends of competitor activities, dealer reactions, and other “field intelligence” Extent of field replacements due to failures in service Competitive quality ratings Independent quality ratings Results of research on quality research Cost summary

Customer service department Internal sales analysis Reports of field sales force Sales of spare parts Customers who buy from multiple sources Independent laboratories Government departments; institutions Monthly quality cost report — see Table 10.4

INSPECTION DECISIONS What to inspect

When to inspect

How much to inspect Type of measurements Who inspects

Where to inspect

Raw materials Processes Products Prior to supplier shipment Upon receipt from suppliers Before start of processes During processes Prior to costly processes Prior to irreversible processes Prior to covering processes (painting) After processes Before shipping to customers 100% inspection Sampling inspection Variable measurement (continuous) Attribute measurement (discrete) External suppliers Workers themselves Quality inspectors Work stations Inspection stations Laboratories

PREVENTION COSTS (SEE TABLE 10.5) APPRAISAL COSTS (SEE TABLE 10.6) INTERNAL FAILURE COSTS (SEE TABLE 10.7) EXTERNAL FAILURE COSTS (SEE TABLE 10.8)

SL3151Ch10Frame Page 488 Thursday, September 12, 2002 6:03 PM

488

Six Sigma and Beyond

TABLE 10.5 Prevention Costs Cost Element

Description/Definition

Where to Obtain/ How to Calculate Cost

1. Quality planning By quality department Quality engineers SQA engineers Reliability engineers Statisticians Other planning

All costs (salary and administrative) related to the planning of an effective quality system that translates customer requirements into the manufacturing process; test and inspection planning costs are reported separately (see #2)

Salary budget reports Expense budget reports Estimates Department budget reports (allocated) Time sheets Purchase orders Estimates

By other departments Manufacturing engineering Controller’s office Systems Administrative Purchasing/other

Allocated costs for time spent in quality planning by personnel not reporting to the quality department

2. Test and inspection planning

Costs of planning and procuring developing test and inspection equipment (excluding actual equipment costs, which are part of appraisal costs) Development costs for test and inspection processes

3. Qualification of new products/ processes/equipment

Costs for qualifying new products, processes, and equipment (including of test and inspection) to meet customer requirements

Department budget reports (allocated) Purchase orders Estimates

Department budgets (allocated) Launch budget (allocated) Purchase orders Estimates

4. Quality training

All costs for developing, implementing, operating, and maintaining formal quality training (including statistical training)

Training budget Purchase orders Estimates

5. Other prevention expenses

All other costs associated with planning, implementing, and maintaining a quality system not specifically included elsewhere

Estimates Adjustments (including negative costs)

SL3151Ch10Frame Page 489 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

489

TABLE 10.6 Appraisal Costs Cost Element

Description/Definition

1. Incoming and receiving inspection and test 2. In-process inspection and test

All costs of inspectors, supervision, lab, and clerical personnel working on incoming material; includes costs to visit or station personnel at supplier locations Salaries and associated costs of all staff performing in-process inspection and testing either 100% or sampling; includes materials consumed during tests Costs of tests, inspection, and lab equipment; equipment maintenance and purchased services also included

3. Test and inspection equipment 4. Product quality reviews 5. Field performance evaluations 6. Other appraisal costs

Personnel expenses for performing quality reviews on in-process or finished products Costs incurred in field testing for product acceptance at a customer’s site, prior to releasing the product All other appraisal costs not specifically covered elsewhere

Where to Obtain/ How to Calculate Cost Department budgets (allocated) Process sheet standards Inspection sheets standards Estimates Same as # 1

Department budgets (allocated) Purchase orders/maintenance contracts Estimates Department budgets (allocated) Estimates Field inspection reports Department budgets (allocated) Estimates Estimates Adjustments (including negative costs)

DIAGNOSTIC GUIDELINES TO IDENTIFY MANUFACTURING PROCESS IMPROVEMENT OPPORTUNITIES 1. Identify the process to be evaluated. 2. Become acquainted with the process by reviewing process sheets and through discussion with line supervision. 3. Visit each operation to review for type of cost incurred, appraisal, internal failure, etc. 4. Talk to individual operators to define further what goes wrong at each operation; mis-assembly, wrong tools, poor setup, etc.; note machine numbers, part numbers, and shift. 5. Identify and quantify failures at each operation; scrap, damage, rework, etc., by shift. 6. Use the existing financial system to assign the cost of direct/indirect labor, benefits, material, etc., to each operation within the process. 7. For each operation calculate the cost of scrap, rework, testing, inspection, production checks, sorting, and audits. Also calculate the costs associated with return sales, warranty, and customer loyalty.

SL3151Ch10Frame Page 490 Thursday, September 12, 2002 6:03 PM

490

Six Sigma and Beyond

TABLE 10.7 Internal Failure Costs Cost Element

Description/Definition

1. Rework and repair internal fault internal fault supplier fault

Costs of reworking defective product; includes costs associated with the renew and dispositioning of non-conforming purchased products

2. Scrap

3. Troubleshooting and failure analysis

All scrap losses incurred resulting defective purchased materials/products and incorrectly performing manufacturing operations; costs charged to suppliers are not included; scrap value, less handling charges, may be included as an offset Costs incurred in analyzing non-conforming product to determine causes

4. Reinspect and retest

Costs to reinspect or retest products that previously failed

5. Excess inventory

Inventory costs resulting from producing defective products; includes storage of defective product and added inventory of good product to cover production shortfalls Costs to revise a product or process due to production of defective product All other costs related to the production of defective product nor specifically included elsewhere

6. Design and process changes 7. Other internal failure costs/offsets

Where to Obtain/How to Calculate Cost Cost accounting reports Detective material reports Department budgets (allocated) Estimates Salvage reports Defective material reports Estimates

Department budgets (allocated) Problem reports Department budgets (allocated) Estimates Cost accounting reports Department budgets (allocated) Estimates Estimates Estimates Adjustments

8. Sum these costs to obtain the total cost of quality within the process. 9. State this cost as a fraction of the total cost of the process or as a dollar amount that represents the opportunity for improvement in the process. 10. Ensure continuous improvement through ongoing process analysis (plan, do, check, act).

DIAGNOSTIC GUIDELINES TO IDENTIFY ADMINISTRATIVE PROCESS IMPROVEMENT OPPORTUNITIES 1. Identify the process or procedure to be evaluated. 2. Become acquainted with the process or procedure by reviewing instruction sheets and procedure manuals and by generating a unique process flow diagram; discuss with local supervision.

SL3151Ch10Frame Page 491 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

491

TABLE 10.8 External Failure Costs Cost Element

Description/Definition

1. Warranty

All warranty costs that can be allocated to a manufacturing location due to the production of defective product or incoming material; includes internal processing and investigation of warranty All costs associated with manufacturing location fault for recall campaigns or liability claims Costs of handling and accounting for defective product returned or rejected by the consuming plant or customer Costs to reinspect or retest defective product at the customer’s site Salary and administrative costs to handle meetings, visits, etc. with customer personnel resulting from the receipt of defective product Costs to revise the product or process to satisfy the customer who received defective product Extraordinary costs that result from attempting to satisfy a customer whose expectations were not met with previously received defective product All other costs related to defective product reaching the customer not specifically covered elsewhere

2. Recalls and product liability claims 3. Products returned or rejected

4. Reinspection and retest 5. Customer and field contacts

6. Design and process changes 7. Customer goodwill

8. Other external failure costs

Where to Obtain/ How to Calculate Cost Warrant reports (allocated) Department budgets (allocated) Estimates

Recall reports Corporate liability settlement reports Department budgets (allocated) Department budgets (allocated) Returned material reports Sales and service reports Department budgets (allocated) Estimates Department budgets (allocated) Estimates

Department budgets (allocated) Estimates Travel and expense reports Department budgets (allocated) Estimates

Estimates Adjustments (including negative costs)

3. Review each operation for the type of cost incurred; appraisal (checks, reviews, etc.), internal failure (blueprint errors, incomplete forms, etc.). 4. Talk to individual employees to define further what goes wrong at each operation; redundant operations, misfiling, improper direction, delays, etc. 5. Identify and quantify failures at each operation and their effect on subsequent operations. 6. Use the existing financial system to assign the cost of labor and material to each operation.

SL3151Ch10Frame Page 492 Thursday, September 12, 2002 6:03 PM

492

Six Sigma and Beyond

7. Calculate the cost of losses associated with the items identified in steps 4 and 5. 8. Sum these costs to obtain the total cost of quality within the process. 9. State this cost as a fraction of the total cost of the process or as a dollar amount that represents the opportunity for improvement in the process. 10. Ensure continuous improvement through ongoing process analysis (plan, do, check, act).

STEPS

FOR

QUALITY IMPROVEMENT — USING COST

OF

QUALITY

Procedure 1. Organize the team. 2. Describe the problem. Estimate the magnitude of quality costs. Identify the key business processes that have the greatest impact on the costs. 3. Define root causes. Identify and prioritize the root causes of process problems. 4. Implement interim corrective action. Establish control of the business process. 5. Implement permanent corrective action. Improve the capability of the business process. 6. Verify effectiveness of actions. Measure effect of actions identified in (4) and (5). 7. Prevent recurrence. Modify management and operating systems, practices, procedures, and processes. 8. Congratulate team. Examples Non-manufacturing measurements, which are sometimes difficult to establish, might include the following: 1. Accounting • Percent of late reports • Computer input incorrect • Errors in specific reports as audited • Percentage of significant errors in reports; total number of reports • Percentage of late reports; total number of reports; average reduction in time spans associated with important reports • Pinpointing high-cost manufacturing elements for correction • Pinpointing jobs yielding low or no profit for correction • Providing various departments with the specific cost tools they need to manage their operations for lowest cost

SL3151Ch10Frame Page 493 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

493

2. Administrative • Success in maximizing discount opportunities through consolidated ordering • Success in eliminating security violations • Success in effecting pricing actions so as to preclude subsequent upward revisions • Success in estimating inventory requirements • Success in responses to customer inquiries so as to maximize customer satisfaction • Decimal points correctly placed • Correct calculations in bills, purchase orders, journal entries, payrolls, bills of lading, etc. • Time spent in locating filed material • Percentage of correct punches in paper used during a given period versus actual output in finished pages 3. Clerical • Accurate typing, spelling, hyphenation • Decimal points correctly placed • Correct calculations in bills, purchase orders, journal entries, payrolls, bills of lading, etc. • Time spent in locating filed material • Percentage of correct punches • Paper used during a given period versus actual output in finished pages 4. Data processing • Keypunch (KP) cards thrown out for error • Computer downtime due to error • Rerun time • Promptness in output delivery • Effectiveness of scheduling • Depth of investigations by programmers • Program debugging time • KP (data entering) efficiency 5. Engineering: design • Adequacy of systems specifications • Accuracy of system block diagrams • Thoroughness of system concepts • Simulation results compared to original design or prediction • Success in creating engineering designs that do not require change in order to make them perform as intended • Success in developing engineering cost estimates versus actual accruals • Success in meeting self-imposed schedules • Success in reducing drafting errors • Success in maximizing capture rates on RFPs for which the company was a contender • Success in meeting engineering test objectives • Number of error-free designs

SL3151Ch10Frame Page 494 Thursday, September 12, 2002 6:03 PM

494

Six Sigma and Beyond

• • • • • • • • • • • • • • • • • •

Correct readings of gages and test devices Accurate specifications and standards Proper reporting and control of time schedules Reduction of engineering design changes Changes in tests or in illustrations of reports Rework resulting from errors in computer program input Advance material list accuracy Design compliance to specifications Customer acceptance of proposals Meeting schedules Thoroughness of systems concepts Accuracy and thoroughness of reports Adequacy of design reviews Compliance to specifications Adequacy of design reviews Accuracy of computations Accuracy of drawings Reduction in number of engineering non-conformances to correct errors 6. Engineering: manufacturing • Accuracy of manufacturing processes • Timely delivery of manufacturing processes to the shop • Accuracy of time study data • Accuracy of time estimates • Timely response to bid requests • Schedule compliance • Asset utilization • Accuracy and thoroughness of test processes • Adequacy and promptness of program facilitation • Application of work simplification criteria • Minimum tool and fixture authorization • Labor utilization index • Methods improvement (in hours or dollars) • Contract cost • Lost business due to price • Process change notices due to error • Tool rework to correct design • Methods improvement 7. Engineering: plant • Effectiveness of preventive maintenance program • Accuracy of estimates (dollars and details) • Accuracy of layouts • Cost of building services • Completeness of plant engineering drawings • Adequacy of scheduling • Fixed versus variable portions of overhead

SL3151Ch10Frame Page 495 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

495

• Maintenance cost versus floor space, manpower, etc. • Lost time due to equipment failures • Janitorial service • Success in meeting or beating budgets • Instrument calibration error • Fire equipment found defective • Lost time due to equipment failures • Purchase requisition errors • Schedule compliance • Timely response to bid requests • Adherence to contract specifications • Effectiveness of customer liaison • Effectiveness of cost negotiations • Status “ship not bill” • Change orders due to errors • Drafting errors found by checkers • Late releases • Time lost due to equipment failure • Callbacks on repairs 8. Finance • Billing errors (check accounts receivable overdues) • Accounts payable deductions missed • Vouchers prepared with no defects • Clock card or payroll transcription errors • Data entering errors • Computer downtime • Timeliness of financial reports • Effectiveness of scheduling program “debugging” time • Rerun time • Accuracy of predicted budgets • Clerical errors on entries • Inventory objectives met • Payroll errors • Discount missed • Amounts payable records • Billing error 9. Forecasting • Can departments function with maximum effectiveness with budgets set for them? • Can the company buy needed capital equipment, keep inventories supplied, pay its bills? • Do projects meet time schedules? • Assistance to line organizations (scheduling, planning, and control functions) • Methods for finance and cost control • Timeliness of financial reports

SL3151Ch10Frame Page 496 Thursday, September 12, 2002 6:03 PM

496

Six Sigma and Beyond

• • • •

Assets control Minimizing capital expenditures Realistic budgets Clear and concise operating policies; timely submission of realistic cost proposals • Completeness of financial reports • Effectiveness of disposition of government property • Effectiveness of cost negotiations 10. Legal • Amount of paper used versus finished pages produced • Misdelivered mail • Misfiled documents • Delays in execution of documents • Teletype errors • Patent claims omitted • Response time on request for legal opinion 11. Management • Output of staff elements, overall defects rates, budgets and schedule controls, and other factors that reflect on managerial effectiveness (In other words, the accomplishments of a manager are the sum totals of those working under him or her) • Success in developing estimates of costs versus actual accruals • Success in meeting schedules • Performance record of employees under the manager’s supervision • Success in developing realistic estimates on a PERT or PERT/cost chart • Success in minimizing use of overtime operations • All nonproduction departments can be measured • Each department should be measured against itself, using time comparisons, and preferably by itself. • The best primary goals are those that measure cost performance, delivery performance, and quality performance of the of the department. Secondary goals can be derived from these primary goals. • There should be a base against which quality, cost or delivery performance can be measured as a percentage improvement. Examples of such a base would be direct labor, the sales dollar, the material dollars, or the budget dollar. A dollar base is more meaningful to management than a physical quantity of output. • Success in effecting pricing actions so as to preclude subsequent revisions. • Pages of data compiled with no defects • Clarity and conciseness of operating procedures • Evaluations of capital investment • Errors in applying standards on process sheets • Accuracy of estimates; actual costs versus estimated costs • Effectiveness of work measurement programs

SL3151Ch10Frame Page 497 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

12. Marketing • Success in reduction of defects through suggestion submittal • Success in capturing new business versus quotations • Responsiveness to customer inquiries • Accuracy of marketing forecasts • Response from news releases and advertisements • Effectiveness of cost and price negotiations • Success in response to customer inquiries (customer identification) • Customer liaison • Effectiveness of market intelligence • Attainment of new order targets • Operation within budgets • Effectiveness of proposals • Exercise of selectivity • Control of cost of sales • Meeting proposal submittal dates • Timely preparation of priced spare parts list • Aggressiveness • Utilization of field marketing services • Dissemination of customer information • Bookings budget met • Accuracy of predictions, planning and selections • Accurate and well-managed contracts • Exploitation of business potential • Effectiveness of proposals • Control of printing costs • Application of standard proposal material • Standardization of proposals • Reduction of reproduction expense • Contract errors • Order description error • Sales order errors 13. Material • Saving made • Late deliveries • Purchase order (PO) errors • Material received against no PO • Status of unplaced requisitions • Orders open to government agency for approval • Delays in processing material received • Damage or loss items received • Claims for products damaged after shipment from our plant • Delays in outbound shipments • Complaints or improper packing in our shipments • Errors in travel arrangements

497

SL3151Ch10Frame Page 498 Thursday, September 12, 2002 6:03 PM

498

Six Sigma and Beyond

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Accuracy of route and rate information on shipments Success in meeting schedules; material shortages in production Success in estimating inventory requirements Clock card errors by employees Damaged shipments Stock shelf life exceeded Items in surplus Purchase requisition errors Effectiveness of material order follow up Adequacy and effectiveness of planning and scheduling Application of residual inventories to current needs Inventory turnover manufacturing jobs without schedules Timeliness of incorporating ECNs Timely replacement of rejected parts Adequacy of reject control plan Effectiveness of packing operation Application of residual inventories to current needs Floor shortages Labor utilization index Data processing rerun time on material programs Bad requisitions Value of termination stores and residual inventory Manpower fluctuations around mean Percent supplier material ($) rejected and returned; total material ($) purchased Number of defective suppliers (repetitive); total number of suppliers Number of single source suppliers; total number of suppliers Percent of supplier material ($) holding up production: total material $ Number of late lots received (actually holding up production); total lots received Percent of purchased material (actual); total material bid or budgeted Percent of reductions in B/M effected through purchasing effort; total material bid or budgeted. Correct quotations or rates Customers call back as promised Installation of exact equipment requested by customer Appointments kept at the time promised customers Prompt handling of complaints Accurate meter readings Courteous treatment of customers Right packages of goods ordered shipped Number of telephone numbers correctly dialed PMI rejects Savings made Material handling budget met

SL3151Ch10Frame Page 499 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

499

• Travel expense against open shop orders • Orders to government for approval — disapproved, resubmitted, and open, not approved 14. Personnel • Success in eliminating security violations • Hiring effectiveness • Thoroughness and speed of responding to suggestions • Employee participation in company sponsored activities • Administration of insurance programs • Accident prevention record • Processing insurance claims • Provision of adequate food services • Personnel security clearance errors • External classified visit authorization errors • Speedy processing of visitors through lobbies • Records accuracy • Adequacy of training programs • Thoroughness and speed of investigating suggestions • Grievances • Employment requisitions filled • Administration of insurance program • Acceptance of organization development recommendations • Effectiveness of administration of merit increases • Overhead budget performance 15. Product assurance • Participation in design reviews • Customer liaison • Technical society participation • Accuracy of proposals and contracts • Application of program policies • Prevention of field complaints • Effectiveness of reporting and recording • Customer rejects • Rejected material on the floor • Adequacy of vendor ratings • Effectiveness of field quality control • Rejects • Screening efficiency • Inspection documentation • Quality assurance audits 16. Product control • Success in developing realistic schedules • Success in developing realistic estimates • Success in identifying defective specifications • Process sheets written with no error

SL3151Ch10Frame Page 500 Thursday, September 12, 2002 6:03 PM

500

Six Sigma and Beyond

• Transportation hours without damage to product • Parts shortages in production • Downtime due to shortages 17. Production • Success in reducing the scrap, rework, and “use-as-is” categories • Success in maintaining perfect attendance records • Success in identifying defective manufacturing specifications • Success in meeting production schedules • Success in cost reduction through suggestion submittal • Success in improving first article acceptance • Performance against standard • Success in reducing required MRB action • Utilities improperly left running at close of shift • Application of higher learning curves • Floor parts shortages • Delays due to rework, material shortage, etc. • Control of overtime (nonscheduled) • Prevention of damage to work in process • Cleanliness of assigned areas • Conformance to estimates • Suggestions submitted • Labor utilization index • Defects • Asset utilization • Scrap • Utilization of correct materials, drawings, and procedures • Prevention of damage • Safety records • Inches of weld with no defects • Log book entries with no defects • Security violations • Compliance to schedules • Accuracy of estimates 18. Program Management • Liaison with customer • Financial quality of proposals (technical approach, cost, time) • Soundness of project plans • Coordination of support activities • Satisfactory field sell off • Backlog • New business volume versus budgeted 19. Publications • Compliance to specifications • Errors corrected • Thoroughness of material • Quality of production

SL3151Ch10Frame Page 501 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

501

20. Quality control • Inspection errors • Sampling program errors • Timeliness of inspection reports • Adequacy of vendor quality ratings • Returned goods and field rework due to inspection oversight; customer rejects • Quality assurance audits • Inspection documentation • Customer liaison 21. Research and development • Can it be applied? • Can it be developed? • Can it be manufactured? • Can it be marketed? 22. Security • Personnel security clearance errors • Timely and accurate processing • External classified visit authorization errors • Accurate processing of visitor identification • Effectiveness of security program • Guards, security checks, badges, passes • Records accuracy • Fire watch 23. Services: general • Promptness in reply to requests • Quality of service rendered • Blueprint and drawing control, reproduction, distribution • Test equipment maintenance and calibration • TRW communication • Reproduction facilities 24. Purchasing • Purchase order changed due to error • Late receipt of materials • Rejections due to incomplete description 25. Supervision A supervisor’s performance is measured by the overall effectiveness of the department; in other words, the supervisor is judged by the sum total of accomplishments of the people working for him or her. The worth of individual or group achievements should be evaluated against the following criteria: • Impact of potential error (abort of mission, cost effect on schedules, etc.) • Contributions of the individual or group to the prevention of error • Difficulty of the job and level of skill required • Work schedules and load impact on error potential

SL3151Ch10Frame Page 502 Thursday, September 12, 2002 6:03 PM

502

Six Sigma and Beyond

• Ability of individual to correct his/her own errors • Attitude of the individual toward work, project, or command mission

GUIDELINE COST

OF

QUALITY ELEMENTS

BY

DISCIPLINE

Note: Nonconformance elements listed reflect those that are the responsibility of that department. Engineering: price of conformance a. Design specification reviews b. Product qualification, evaluation, characterization c. Drawing checking d. Supplier evaluation e. Preventive maintenance f. Process capability studies g. Fabrication of special test fixtures h. Verification of workmanship standards i. Review of test specifications j. Failure effects/mode analysis k. Pilot production runs l. Packaging qualification m. Customer interface n. Safety review — operator safety o. Technical manuals p. Preproduction reviews q. Defect prevention program r. Schedule reviews s. Process reviews t. Early approval of production specifications u. Computer-aided design (CAD) v. First piece approval w. Agency approval x. Supplier qualification y. Special test fixture design review z. Education aa. Prototype inspection and test ab. Testing ac. Receiving sample testing ad. In-process sample testing ae. Final sample testing af. Laboratory analysis and test ag. Fault insertion test ah. Engineering audits ai. Training for special testing aj. Personnel appraisal

SL3151Ch10Frame Page 503 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

Engineering: price of nonconformance a. Warranty expense b. Engineering travel and time on problems c. Redesign d. Premium freight costs e. Material review activities f. Failure analysis (return evaluation) g. Corrective action h. Failure reports i. Return goods analysis j. Product liability (design related) Comptroller: price of conformance a. Forecasting performance b. Training and procedures c. Ledger review of P & L and balance sheet d. Budget generation e. Long-range planning f. Job description g. Cost of quality budget h. Timecard review i. Capital expenditure reviews j. Total expenditure reviews and delegation of authority k. Order entry review l. Product cost standards m. Cost reduction n. Cost of quality reviews o. Data processing report/financial report reviews p. Ledger reviews q. Invoicing review Comptroller: price of nonconformance a. Billing errors b. Inventory out of control c. Improper A/P vendor payments d. Incorrect accounting entries e. Bad debts turnovers, overdue A/R f. Payroll errors Software: price of conformance a. Software planning b. Software reliability projection/prediction c. Systems analyst interrogating activities d. Documentation review e. Learn/understand customer requirement/business f. Preparation and review of system specifications g. Flow chart review h. Correlation analysis

503

SL3151Ch10Frame Page 504 Thursday, September 12, 2002 6:03 PM

504

i. Data entering operator training j. Tape duplication and verification k. Program testing l. Function test m. Performance test n. Code verification o. Depreciation of software (outdated) p. System test q. Inspect programs Software: price of nonconformance a. Keeping track of system failures b. Going back to customer to re-evaluate c. Customer change requirement d. Recode; debug; retest e. Documentation changes Plant administration: price of conformance a. Consultants b. Preventive maintenance program c. Modeling d. Controlled/critical storage e. Environmental control f. Labor training g. Review of labor production rates h. Security i. Surveillance j. Machine maintenance — P.M. k. Machine maintenance training l. Timely machine replacement m. Periodic equipment depreciation review n. Equipment depreciation reappraisal o. Facility planning — audits p. Facility inspection and test q. Data on labor productivity r. Pilot run to check standard s. Labor surveillance t. Time card control test u. Time card audit v. Machine maintenance test w. Machine maintenance inspection x. Equipment depreciation inventory y. Equipment depreciation audit z. Equipment depreciation tracking Plant administration: price of nonconformance a. Facility planning redesign b. Facility equipment replacement c. Missed schedule

Six Sigma and Beyond

SL3151Ch10Frame Page 505 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

d. Incorrect labor level e. Increased failure f. Incorrect time g. Machine scrap h. Machine rework i. Machine downtime j. Product liability k. Equipment depreciation — obsolete l. Equipment depreciation — premature Purchasing: price of conformance a. Supplier review and approval b. Send proper specs to vendor — make it clear what is required c. Periodic seminars d. Forecasting — cost of carrying hard-to-get materials e. Material cost resulting from multiple sourcing f. Strike build-up costs g. Evaluation of supplier’s equipment that will be used to do the job h. Review supplier incoming quality practices i. Recertification of suppliers j. Incoming inspection cost k. Information — systems cost associated with vendor rating Purchasing: price of nonconformance a. Scrap b. Sorting c. Reinspection due to rejects d. Rework e. Excess inventory due to lack of confidence in vendor delivery f. Loss incurred as a result of vendor delinquencies g. Corrective action cost h. Shipping cost on returns to vendors i. Purchase order changes resulting from error j. Incoming inspection cost k. Premium freight l. Trips to suppliers to resolve problems m. Expediting cost to ensure proper deliveries (i.e., phone bill) Marketing: price of conformance a. Procedures b. Training c. Forms design d. Sales support material e. Design specs f. P&L g. Computive data h. Market forecast i. Legal and product safety review j. User market research

505

SL3151Ch10Frame Page 506 Thursday, September 12, 2002 6:03 PM

506

k. Sales support cost l. Customer survey m. Sales dollars n. Service cost by area/advertising o. Loss leaders p. Launch U.S. availability q. Pilot and field test r. Incentive programs s. Market survey Marketing: price of nonconformance a. Labor of redos — administration b. Incorrect order entry c. A/R receivables d. Special instruction e. Field service — excessive f. Warranty g. Literature reprint h. Contingent liability i. Unit productivity j. Product recall k. Loss of market share Human resources: price of conformance a. Pre-screen applications b. Interviewing c. Personnel testing d. Reference checking e. Security clearance, if necessary f. Orientation g. Training h. Job description and work plans i. Safety program j. Quality improvement program k. Exit interviews l. Performance appraisal m. Attendance tracking n. Productivity rates o. Personnel records audits p. Tracking of injuries Human resources: price of nonconformance a. Turnover rates b. Grievance tracking c. Non-timely filling of position d. Cost of injuries Manufacturing: price of conformance a. Training: supervisor hourly

Six Sigma and Beyond

SL3151Ch10Frame Page 507 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

b. Special review c. Tool/equipment control d. Preventive maintenance e. ZD program f. Identify incorrect (zero defect) specifications/drawings g. Housekeeping h. Controlled overtime i. Checking labor j. Trend charting k. Customer source inspection l. First piece inspection m. Stock audits n. Certification Manufacturing: price of nonconformance a. Rework b. Scrap c. Repair and return expenses d. Obsolescence e. Equipment/facility damage f. Repair equipment/material g. Expense of controllable absence h. Supervision of manufacturing failure element i. Discipline costs j. Lost time accidents k. Product liability Quality control: price of conformance a. Quality training b. Test planning c. Inspection planning d. Audit planning e. Product design review f. Supplier qualification g. Productibility/quality analysis review h. Process capability studies i. Machine capability studies j. Calibration of quality equipment k. Operator certification l. Incoming inspection m. In-process inspection n. Final product inspection o. Product test p. Product audit q. Test equipment r. Checking gauges, fixtures, etc. s. Prototype inspection

507

SL3151Ch10Frame Page 508 Thursday, September 12, 2002 6:03 PM

508

Six Sigma and Beyond

t. Quality systems audits u. Customer/agency audits v. Outside lab evaluations w. Life testing x. Product audits Quality control: price of nonconformance a. Scrap analysis b. Rework analysis c. Warranty cost analysis d. Concessions analysis e. Factory returns analysis f. Material review board action Industrial engineering: price of conformance a. Operator training b. Design review c. Inventory control d. Job description e. Methods description f. Test equipment description verification g. Material utilization h. Line rebalance i. Process verification j. Product control card system k. Material usage verification Industrial engineering: price of nonconformance a. Tool repair b. Tool modification c. Corrective action costs d. Engineering change order e. Purchasing change order f. Turnover g. Obsolete job description Information systems: price of conformance a. Job descriptions (written) b. Hiring and testing c. Schools d. Program documentation and testing e. Cost benefit analysis f. Risk analysis of project g. Proper communication of user requirements between user and information systems h. Verification of input data i. Test techniques j. Pilots k. Parallel runs l. Post implementation audit

SL3151Ch10Frame Page 509 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

509

Information systems: price of nonconformance a. Systems do not meet user requirements — redo b. Corrective maintenance c. Reruns d. Input cycles edit and update e. Hardware downtime f. Scheduling failures Law department: price of conformance a. Maintenance of law library b. Seminar on prevention of product liability claims c. Label copy evaluation d. Advertisement copy review e. Safety program audit f. Equal opportunity program audit g. Compliance audit of SEC h. Contract review i. Checking paperwork for errors j. Compliance audit on environmental protection agency (EPA) k. Review of federal/state submissions (new products, patents, etc.) Law department: price of nonconformance a. Product liability matters (travel, litigation, outside firms, time) b. Warranty reviews c. Penalties for late filing d. Product complaint review (internal and with regulatory agency) e. Product recalls f. Defense of patent infringement suit g. Representing grievances h. Internal department rework (rewrite, retype, etc.) i. Seminar on defending product liability suits j. Settlements

COST

OF

QUALITY

AND

DFSS RELATIONSHIP

We made the statement earlier that perhaps one of the key contributions of COQ in DFSS is to identify and eliminate the “hidden factory” cost. To do that, let us visit some of the more demanding calculations. First, we begin with a review of the Poisson distribution:

r

Y=

( np) e r!

− np

( d )r e this of course is equivalent to Y = u r!

−d

u

where n is the total number of independent trials; p is the probability of occurrence; r is the number of occurrences; u is the number of units produced; and d is the number of non-conformities (defects). This value (d) is also known as the np.

SL3151Ch10Frame Page 510 Thursday, September 12, 2002 6:03 PM

510

Six Sigma and Beyond

Under special conditions, such as normalizing per unit, the d/u = np/u. Therefore, if we substitute terms for the normalizing case where u = 1, for the special case where r = 0 (remember 0! = 1), we are able to reduce the above formulas into Y =e

−d

u

The reader will notice that this equation really reflects the first time yield (YFT) for a specific d/u. Of course, if we know the first time yield we can solve for d/u with the following formula: d/u = –ln(YFT) where ln is the natural log. If the assumptions for the Poisson model are not reasonable then we may use the binomial model Y=

m! pr qm−r r!( m − r )!

This of course, becomes Y = (1 – p)m for the special case of r = 0 where p is the constant probability of an event and q = 1 – p. In dealing with COQ issues, as Grant and Leavenworth (1980) have pointed out, the Poisson approximation may be applied when the number of opportunities for non-conformance (n) is large and the probability (p) of an event (r) is small. In fact, as n increases and r decreases, the approximation by the Poisson model improves. Furthermore, we can use COQ issues and information to generate or doublecheck the validity of the critical to quality characteristic (CTQ) as well as the critical to process characteristic (CTP). Above all, we are capable of measuring the classical perspective of yield. The traditional formula, which is based on process capability, is Y final =

S U

where Yfinal = final yield; S = number of units that pass; and U = number of units tested. Another way to view yield is to calculate the rolled throughput, which is Yrt = e − dpu or Yrt = Ytp;1 * Ytp;2 *,… * Ytp;m or the normalized variation of Ynorm = (Yrt )

1 m

where Ynorm is the normalized yield; Yrt is the rolled throughput yield; m is the number of categories; tp is the throughput yield of any given category; and dpu is defects per unit.

SL3151Ch10Frame Page 511 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

511

REFERENCES Grant, E.L. and Leavenworth, R.S., Statistical Quality Control, McGraw-Hill, New York, 1980.

SELECTED BIBLIOGRAPHY Aaron, M.B., Measure for Measure Alternative to Goodness, Motherhood & Morality, paper presented in Automach, Australia for SME, 1986. Besterfield, D.H., Quality Control, Prentice Hall, Englewood Cliffs, NJ, 1979. Carlsen, R.D., Gerber J, and McHugh, J.F., Manual of Quality Assurance Procedures and Forms, Prentice Hall, Englewood Cliffs, NJ, 1981. Ford Motor Co., Team Orientated Problem Solving. Electrical & Electronics Div., Rawsonville, MI, 1987. Juran, J.M. and Gzyna F.M., Quality Planning and Analysis. 3rd ed., McGraw-Hill, New York, 1993. Schneiderman, A.M., Optimum Quality Costs and Zero Defects: Are They Contradictory Concepts? Quality Program, Nov. 1986, pp. 23–27.

REENGINEERING Reengineering by definition is a drastic change of the process. However, if the process changes, then the job/task regarding that process must be changed as well. This section addresses the approach and method for reengineering only from the process perspective. For more details see Stamatis (1997) and Selected Bibliography. The discussion will focus on drastic changes as well as developmental changes for the process. Evolutionary changes in process are addressed by statistical process control charting and other monitoring methods and are beyond the scope of this volume (Volume IV of this series covers this topic). Both approaches to redesign merge the viewpoints of management and labor, resulting in more job satisfaction and productivity. Drastic changes are taking place across the corporate world in the areas of communication practices, corporation cultures, and productivity. These changes are the result of increased employee awareness, an advanced level of technology, competition, mergers, greater demand of quality, and in general, increasing business costs. These changes have forced management to respond in several ways, including asking employees (union and non-union) for their help. This initiative by management has resulted in participative programs such as teamwork and as of late redesigning the actual job or process. As a result of these changes, trust and open communication are cultivated and encouraged. Information sharing, as well as moving responsibility and accountability to employees themselves, is a common occurrence. This employee participation has generated a need for both job and process redesign so that an organization may be more competitive in the world markets as well as more efficient in producing its product or service.

PROCESS REDESIGN A process may change in evolutionary form and or in a very drastic approach (Chang, 1994). When a process changes in an evolutionary form, it may be because of statistical process control monitoring or some other kind of monitoring method.

SL3151Ch10Frame Page 512 Thursday, September 12, 2002 6:03 PM

512

Six Sigma and Beyond

Under this condition, the redesign process takes the form of a problem-solving approach. A typical approach is the following: Step 1. Reason for improvement — why is there a need for change? Select appropriate and applicable measures and targets. Determine any gaps. Step 2. Define problem — Whenever possible, stratify the improvement area. Look for root cause rather than symptoms. Step 3. Analysis — Verify root cause. Step 4. Solution(s) — Determine alternatives. Select best solution. Step 5. Result(s) — Verify and evaluate the elimination of the root cause by asking: Are we better off? Are we worse off? Are we the same as before? Step 6. Implementation — Review the control plan. Change it if appropriate and applicable. Standardize the process. Replicate.

THE RESTRUCTURING APPROACH When a process changes drastically, it may be because of a newly introduced technology or a process reengineering effort. If a process changes because of newly introduced technology, then this process follows a restructuring plan. Process restructuring seeks to break down complexity into manageable pieces. Rather than focusing on a single problem, the process redesigner seeks to understand how whole sets of activities and problems are interrelated. When complexity itself is a problem, restructuring is probably the best path to pursue (Rupp and Russell, 1994). The heart of the restructuring approach consists of six steps. They are: Step 1. Need for change — Reevaluate the process and the technology under consideration. Why is the technology necessary? Are there other alternatives? Step 2. Analysis of new technology — What is the cost? What is the time for implementation? What does the value analysis indicate? Step 3. Evaluate the “new” paradigm(s) — Look at the alternatives from a wider perspective. Look at as many alternatives as possible. Stretch the alternatives for results. Step 4. Design the “new” process — Define, measure and evaluate all alternatives with the following in mind: • Flow • Structure • Tasks

SL3151Ch10Frame Page 513 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

513

Step 5. Build the “new” process — Verify the process performance and effectiveness. Step 6. Implementation — Develop “new” control plan. Standardize the “new” process Replicate If, on the other hand, a process changes because of reengineering efforts, then this process is developed in four stages. The four stages are: Stage 1. Recognition for change. One of the first things that the team has to recognize is that the status quo is about to change. Once the realization sets in, then a formal analysis of what has to change must be performed and the intentions of that change must be communicated throughout the organization to those who are or will be involved. As part of the communication effort, the context of the change and the operating principles will also be communicated. Stage 2. Change content definition (formulation). In this stage, a process map is developed, so that the process targets and objectives may be declared. (In some cases, two process maps are developed as needed. One represents the old process and the second represents the new process. This is done for comparison purposes.) This stage begins the process analysis (tasks and jobs required) and determines the process changes and the new owners. Perhaps one of the most important aspects of this stage is the formulation of the baseline production requirements such as capacity, cycle time, productivity, efficiency, and quality requirements. In some cases, this is the stage where a pilot study will be designed. Stage 3. Change implementation. This is the stage where most of the tedious work has to occur. Specifically, the controls are designed for the new process, and a systematic analysis is performed to identify potential modification points and to eliminate non–value adding steps. A formal value analysis and FMEA may be performed in this stage to identify areas of opportunity and possible restructuring. Stage 4. System maintenance. This is the final stage of the reengineering process. This is where the old system officially is declared obsolete and the new system is installed with all the new structures, modifications, waste reductions, and new targets of production. This four-stage model of the process redesign identifies the general elements of the change. To complete the discussion, however, we must also address the specific tasks that the team leader (project manager) must perform and how the team members will respond. Table 10.9 summarizes these seven steps to the implementation process.

THE CONFERENCE METHOD Another way to redesign a process is the conference method. This method is based on the notion of a cross-functional design team, which has been charted by a steering

SL3151Ch10Frame Page 514 Thursday, September 12, 2002 6:03 PM

514

Six Sigma and Beyond

TABLE 10.9 Seven-Step Process Redesign Model Prime Action of Team Leader

Support Action

Introduce the process change; full disclosure about the project to all concerned

Action kick-off by management team who will be responsible to the team and team leader for follow-up Strategy is incorporated into the business plan, and the team leader acts as both changing agent and support member Managers and team leader support the team for the change; they provide encouragement, coaching, and resources as needed Managers and team leader define the system of the “change”; they provide the appropriate support as needed.

Develop strategy for implementation

Perform appropriate and applicable training to the team members Follow up with both managers and employees to develop team level information, meetings, reports, problem resolution structure, and whatever else is necessary Follow up on the reports and measurement element of this stage; evaluation of the results is also important Help managers with problem Full integration of process redesign

Team leader conducts meetings and improves quality figures; interpretation is an important information through a systematic flow upward Team leader and managers set first set of performance resolution targets Team leader reports progress; audit(s) may be conducted in order to verify targets and or modify the process or the targets

committee to design a more effective organization (Weisbord, 1987; Wilgus, 1995). Within the charter of the team it is imperative that a four-item analysis be addressed. That analysis should cover the following: 1. External influences and how the organization must change in response to these influences 2. Customer analysis 3. Vision of what the company aspires to become 4. Principles or values that will guide behavior In this method, there are three basic ways of analysis: 1. The vision conference, which is made up of two elements: a. The past and present. Here the team acknowledges practices that should be brought into the future and those that should be dropped. b. The future. Here the team reviews the vision statement and identifies the organizational values, which include customer focus, trust, and shared responsibility. 2. The technical conference, which is made up of two elements:

SL3151Ch10Frame Page 515 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

515

a. Analysis of current state. Here the core process is studied and evaluated rather than what people do. b. Creation of the ideal state. Here the group activity is focused on generating a conducive environment for breakthrough thinking. 3. The social conference. In this stage of the conference the social system of the organization is evaluated. The elements upon which the evaluation is based are structure, skills, style, symbols, and human systems. Each element is influenced by changes in the environment and must be aligned with the organization’s vision values and technical system. The goal of this stage is to generate a design that is the “best.”

THE OOAD METHOD Yet another way of redesigning the process is through Object Oriented Analysis and Design (OOAD). Basically, this method is a framework for understanding, developing, organizing, and managing projects. OOAD is the practice of examining any collection of activities (processes, automation situation, and information flow) as a series of interacting objects. As such, this method is applicable across many industries. Historically, OOAD is the result of the interaction of project management tools (PERT, CPM), system design tools (CASE), and analysis tools. As such OOAD has a valued use throughout project and process life cycle activities. OOAD does not require special computing, CASE tools, or computer programs. The four basic tools or templates basic to doing enterprise and plant OOAD are the opportunity framework, the activity object diagram, the integrated object template, and the system requirements template. From a reengineering perspective of a particular process, the OOAD has the following minimum requirements: 1. 2. 3. 4. 5. 6. 7. 8.

Motivate people and keep them involved. Get people to recognize that quality is their responsibility. Recognize other initiatives and coordinate with their activities. Get all interacting groups involved in the project (operations, automation, safety, logistics, finance, and so on). Get the information to the place and people where it is needed in a timely manner. Translate the engineering, operations, and support views into the overall organization’s view. Include the effects of “real world” practices in the design. Ensure that the technology is acquired by the client.

Once the minimum requirements are established, then the reengineering team is ready to implement the method for change. The success of the OOAD will depend upon the implementation of the following ten steps: 1. Establish the program scope and focus. Define and publish OOAD. Agree on primary focus — the core process to be reengineered.

SL3151Ch10Frame Page 516 Thursday, September 12, 2002 6:03 PM

516

Six Sigma and Beyond

2. Set high level goals and objectives. Derive from current business strategies. Identify critical processes, activities, and metrics. 3. Establish the appropriate time horizon for planning. Incorporate technology forecasts. Balance tactical with strategic intent. 4. Adopt the specific methodology for executing the plans. Provide a plan for planning. Separate work plans from techniques. 5. Organize teams and participants. Involve the right people. Allocate adequate time and training. 6. Provide education and training. Internal External missions 7. Define the procedure to approve plans. Process priority High level (top management) commitment 8. Promote technology transfer. Project organization Education/training 9. Communicate plans and progress. Conferences and presentations Recognition and rewards 10. Screen outside assistance. Timetable and expertise

REENGINEERING

AND

DFSS

Reengineering may play a major role in the DFSS because it will focus the process of improvement by • • • • • •

Positioning six sigma in a process management framework Achieving breakthrough improvements in quality, cost, and cycle time Leveraging six sigma with reengineering and vice versa Using the right tool for the right problem Avoiding the traps of six sigma implementation Turning process management and improvement into a way of life

Many executives who so enthusiastically embrace six sigma do not really know what they are getting into, and that is a guarantee of trouble downstream. For many, six sigma methodology has become a corporate panacea, a silver bullet of sorts, and that spells unrealistic expectations and eventual disillusionment. More importantly, many companies are discovering that there is a large gap between learning the techniques of six sigma and realizing its benefits. Some are falling prey to six sigmaitis, symptomized by a vast number of uncoordinated projects that do not support

SL3151Ch10Frame Page 517 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

517

critical corporate goals or customer functionality. If these problems are not addressed quickly, six sigma will become just another corporate fad, companies will not benefit from its power, and institutional cynicism will get yet another boost. We have been down this road before, and we do not need to go down it again. Executives of these companies need a better understanding of what six sigma is and what it is not, what it can do and what it cannot. With six sigma’s focus on problem identification and resolution, can it create breakthrough process designs? Does it have the power to address so-called “big P” (big picture) processes that cross functional boundaries? Is its popularity at least in part a reflection of the fact that six sigma does not shake up an organization, which might make it easier to swallow but limits its impact? Might six sigma actually reinforce, rather than knock down, the silos that impede improved performance? We need realistic answers to these essential questions. We suggest that process management and reengineering are necessary complements to six sigma — especially in the DFSS phase. Six sigma is part of the answer, but it is not the whole answer. Six sigma veterans have learned they need to leverage their six sigma efforts with other improvement techniques — process management, reengineering, and lean thinking. Six sigma is a methodology that uses many tools, is not a religion, and business results are more important than ideological purity. The real winners at DFSS do not limit their arsenal to just one weapon but employ all appropriate techniques, combining them in an integrated program of process redesign and improvement. After all, we already know that over 90% of our problems are systems (process) problems. It would be ludicrous to look the other way when we have a chance to fix them in the design stage. In addition, most of the six sigma projects, discussions, and experts emphasize the manufacturing end of potential improvements. But let us remember that in services, the cost of quality is in the 60 to 70% range of sales. That is an incredible potential of improvement, and reengineering is a perfect tool not only to evaluate but also to help the six sigma initiatives reformulate the processes in such a way that improvements will occur as a matter of course and on an ongoing basis. In this respect, one may even say that the goal of DFSS and reengineering is to create a process in which all its components are managed, measured, and improved from two distinct, yet the same perspectives: 1. Organizational profitability 2. Satisfaction of customer functionality

REFERENCES Chang, R.Y., Improve Processes, Reengineer Them, or Both, Training & Development, Mar. 1994, pp. 54–61. Rupp, R.O. and Russell, J.R., The Golden Rules of Process Redesign, Quality Progress, Dec. 1994, pp. 85–90. Stamatis, D.H., The Nuts and Bolts of Reengineering, Paton Press, Red Bluff, CA, 1997. Weisbord, M., Productive Workplaces, Jossey-Bass, San Francisco, 1987. Wilgus, A.L., The Conference Method of Redesign, Quality Progress, May 1995, pp. 89–95.

SL3151Ch10Frame Page 518 Thursday, September 12, 2002 6:03 PM

518

Six Sigma and Beyond

SELECTED BIBLIOGRAPHY Baer, W., Employee-Managed Work Redesign: New Quality Of Work Life Developments, Supervision, Mar. 1986, pp. 6–9. Cooley, M., Architect or Bee, South End Press, Boston, 1981. Crosby, B., Employee Involvement and Why It Fails, What It Takes to Succeed, Personnel Administration, Feb. 1986, pp. 95–96. Meyer, L., How the Right Measures Help Teams Excel, Harvard Business Review, May/June 1994, pp. 95–104. McElroy, J., Back to Basics at NUMI: Quality Through Teamwork, Automotive Industries, Oct. 1985, pp. 63–64. Pyzdek, T., Considering Constraints, Quality Digest, June 2000, p. 22 Schumann, A.l., Fed Up with Furniture Failure? Office System 85, May 1986, pp. 60–67. Solomon, B.A., A Plan That Proves Team Management Works, Personnel, June 1985, pp. 6–8. Swineheart, D.P. and Sherr, M.A., A System Model for Labor-Management Cooperation, Personnel Administration, Apr. 1986, pp. 87–90. Wall, T., What Is New in Job Design, Personnel Management, 1984, pp. 27–29.

GEOMETRIC DIMENSIONING AND TOLERANCING (GD&T) GD&T is an engineering product definition standard that geometrically describes design intent. It also provides the documentation base for the design of quality and production systems. Used for communication between product engineers and manufacturing engineers, it promotes a uniform interpretation of a component’s production requirements. This interpretation and communication are of interest to those who are about to undertake the DFSS baton. Without the appropriate and applicable interpretation of the design and without the appropriate communication of that design to manufacturing, problems will definitely occur. Therefore, in this section we will address some of the key aspects of GD&T in a cursory manner. We will touch on some of the definitions and principles of general tolerancing as applied to conventional dimensioning practices. The term conventional dimensioning as used here implies dimensioning without the use of geometric tolerancing. Conventional tolerancing applies a degree of form and location control by increasing or decreasing the tolerance. Conventional dimensioning methods provide the necessary basic background to begin a study of geometric tolerancing. It is important that you completely understand conventional tolerancing before you begin the study of geometric tolerancing. When mass production methods began, interchangeability of parts was important. However, many times parts had to be “hand selected for fitting.” Today, industry has faced the reality that in a technological environment, there is no time to do unnecessary individual fitting of parts. Geometric tolerancing helps ensure interchangeability of parts. The function and relationship of a particular feature on a part dictates the use of geometric tolerancing. Geometric tolerancing does not take the place of conventional tolerancing. However, geometric tolerancing specifies requirements more precisely than conventional tolerancing does, leaving no doubts as to the intended definition. This precision

SL3151Ch10Frame Page 519 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

519

TABLE 10.10 GD&T Characteristics and Symbols Geometric Characteristics and Symbols Characteristics Symbol Flatness

Feature

Tolerance

Straightness Individual

Circularity (Roundness) Cylindricity Profile of a Surface

Individual

Profile of a Line

Related

Form

Parallelism Perpendicularity (Squareness) Related

Angularity Circular Runout

Runout

Total Runout Position

Location

Concentricity

Maximum Material Condition

M

Regardless of Feature Size

S

Least Material

L

Projected Tolerance Zone

P

Diameter Basic Datum Feature Symbol Datum Target

.50° A A

may not be the case when conventional tolerancing is used, and notes on the drawing may become ambiguous. When dealing with technology, a drafter needs to know how to properly represent conventional dimensioning and geometric tolerancing. Also, a technician must be able to accurately read dimensioning and geometric tolerancing. Generally, the drafter converts engineering sketches or instructions into formal drawings using proper standards

SL3151Ch10Frame Page 520 Thursday, September 12, 2002 6:03 PM

520

Six Sigma and Beyond

and techniques. After acquiring adequate experience, a design drafter, designer, or engineer begins implementing geometric dimensioning and tolerancing on the research and development of new products or the revision of existing products. Most dimensions in this text are in metric. Therefore, a 0 precedes decimal dimensions less than one millimeter, as in 0.25. When inch dimensions are used, a 0 will not precede a decimal dimension that is less than one inch. For review of decimals and their operations, refer to Volume II of this series. Most dimensions in this text are in the metric International System of Units (SI). The common SI unit of measure used on engineering drawings is the millimeter. The common U.S. unit used on engineering drawings is the inch. (The reader may want to review the discussion and conversions of the SI system in Chapter 20 of Volume II of this series.) The actual units used on your engineering drawings will be determined by the policy of your company. The general note “UNLESS OTHERWISE SPECIFIED, ALL DIMENSIONS ARE IN MILLIMETERS” (or “INCHES”) should be placed on the drawing when all dimensions are in either millimeters or inches. When some inch dimensions are placed on a metric drawing, the abbreviation “IN.” should follow the inch dimensions. The abbreviation “mm” should follow any millimeter dimensions on a predominantly inch-dimensioned drawing. Angular dimensions are established in degrees (°) and decimal degrees (X.X°), or in degrees (°) minutes (′) and seconds (″). The following are some rules for metric and inch dimension units (for a more detailed discussion see Volume II of this series): Millimeters • The decimal point and zero are omitted when the metric dimension is a whole number. For example, the metric dimension “12” has no decimal point followed by a zero. • When the metric dimension is greater than a whole number by a fraction of a millimeter, the last digit to the right of the decimal point is not followed by a zero. For example, the metric dimension “12.5” has no zero to the right of the five. This rule is true unless tolerance values are displayed. • Both the plus and minus values of a metric tolerance have the same number of decimal places. Zeros are added to fill in where needed. • A zero precedes a decimal millimeter that is less than one. For example, the metric dimension “0.5” has a zero before the decimal point. • Examples in ASME Y14.5M show no zeros after the specified dimension to match the tolerance. For example, 24 ± 0.25 or 24.5 ± 0.25 are correct. However, some companies prefer to add zeros after the specified dimension to match the tolerance, as in 24.00 ± 0.25 or 24.50 ± 0.25. Inches • A zero does not precede a decimal inch that is less than one. For example, the inch dimension “.5” has no zero before the decimal point. • A specified inch dimension is expressed to the same number of decimal places as its tolerance. Zeros are added to the right of the decimal point

SL3151Ch10Frame Page 521 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

521

if needed. For example, the inch dimension “.250 ± .005” has an additional zero added to “.25” to match the three-decimal tolerance. • Fractional inches may be used but generally indicate a larger tolerance. Fractions may be used to give nominal sizes such as in a thread callout. • Both the plus and minus values of an inch tolerance have the same number of decimal places. Zeros are added to fill in where needed. • Zeros are added where needed after the specified dimension to match the tolerance. For example, 2.000 ± .005 and 2.500 ± .005 both have zeros added to match the tolerance. The following rules are summarized from ASME Y14.5M. These rules are intended to give you an understanding of the purpose for standardized dimensioning practices. Short definitions are given in some cases: • Each dimension has a tolerance except for dimensions specifically identified as reference, maximum, minimum, or stock. The tolerance may be applied directly to the dimension, indicated by a general note, or located in the title block of the drawing. • Dimensioning and tolerancing must be complete to the extent that there is full understanding of the characteristics of each feature. Neither measuring the drawing or assumption of a dimension is permitted. Exceptions include drawings such as loft, printed wiring, templates, and master layouts prepared on stable material. However, in these cases the necessary control dimensions must be given. • Each necessary dimension of an end product must be shown. Only dimensions needed for complete definition should be given. Reference dimensions should be kept to a minimum. • Dimensions must be selected and arranged to suit the function and mating relationship of a part. Dimensions must not be subject to more than one interpretation. • The drawing should define the part without specifying the manufacturing processes. For example, give only the diameter of a hole without a manufacturing process such as “DRILL” or “REAM.” However, there should be specifications given on the drawing, or related documents, in cases where manufacturing, processing, quality assurance, or environmental information is essential to the definition of engineering requirements. • It is allowed to identify (as non-mandatory) certain processing dimensions that provide for finish allowance, shrink allowance, and other requirements, provided the final dimensions are given on the drawing. Nonmandatory processing dimensions should be identified by an appropriate note, such as “NON-MANDATORY (MFG DATA).” • Dimensions should be arranged to provide required information arranged for optimum readability. Dimensions should be shown in true profile views and should refer to visible outlines. • Wires, cables, sheets, rods, and other materials manufactured to gage or code numbers should be specified by dimensions indicating the diameter

SL3151Ch10Frame Page 522 Thursday, September 12, 2002 6:03 PM

522

Six Sigma and Beyond

•

•

•

•

• •

or thickness. Gage or code numbers may be shown in parentheses following the dimension. A 90° angle is implied where centerlines, and lines displaying features, are shown on a drawing at right angles and no angle is specified. The tolerance for these 90° angles is the same as the general angular tolerance specified in the title block or in a general note. A 90° basic angle applies where centerlines of features are located by basic dimensions and no angle is specified. Basic dimensions are considered theoretically perfect in size, profile, orientation, or location. Basic dimensions are the basis for variations that are established by other tolerances. Unless otherwise specified, all dimensions are measured at 20°C (68°F). Compensation may be made for measurements taken at other temperatures. All dimensions and tolerances apply in a free state condition except for non-rigid parts. Free state condition describes distortion of the part after removal of forces applied during manufacturing. Non-rigid parts are those that may have dimensional change due to thin wall characteristics. Unless otherwise specified, all geometric tolerances apply for full depth, length, and width of the feature. Dimensions apply on the drawing where specified.

To appreciate GD&T, the following definitions must be understood to be successful in your dimensioning practices: Actual size —The measured size of a feature or part after manufacturing. Diameter — The distance across a circle measured through the center. Represented on a drawing with the symbol “0.” Circles on a drawing are dimensioned with a diameter. Dimension — A numerical value indicated on a drawing and in documents to define the size, shape, location, geometric characteristics, or surface texture of a feature. Dimensions are expressed in appropriate units of measure. Feature — The general term applied to a physical portion of a part or object. A surface, slot, tab, keyseat, or hole are all examples of features. Feature of size — One cylindrical or spherical surface, or a set of two parallel plane surfaces, each feature being associated with a size dimension. Nominal size — A dimension used for general identification such as stock size or thread diameter. Radius — The distance from the center of a circle to the outside. Arcs are dimensioned on a drawing with a radius. A radius dimension is preceded by an “R.” The symbol “CR” refers to a controlled radius. A controlled radius means that the limits of the radius tolerance zone must be tangent to the adjacent surfaces, and there can be no reversal in the contour. The use of CR is more restrictive than R (where reversals are permitted). The symbol “SR” refers to a spherical radius.

SL3151Ch10Frame Page 523 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

523

Reference dimension — A dimension, usually without a tolerance, used for information purposes only. This dimension is often a repeat of a given dimension or established from other values shown on the drawing. A reference dimension does not govern production or inspection. A reference dimension is shown on a drawing with parentheses. For example, (60) would indicate a reference dimension. Stock size —A commercial or premanufactured size, such as a particular size of square, round, or hex steel bar. A tolerance is the total amount that a specific dimension is permitted to vary. A tolerance (not to be confused with TOLERANSING) is not given to values that are identified as reference, maximum, minimum, or stock sizes. The tolerance may be applied directly to the dimension, indicated by a general note, or identified in the drawing title block. The limits of a dimension are the largest and smallest numerical value that the feature can be. For example: A dimension is stated as 12.50 ± 0.25. This is referred to as plus-minus dimensioning. The tolerance of this dimension is the difference between the maximum and minimum limits. The upper limit is 12.50 + 0.25 = 12.75 and the lower limit is 12.50 – 0.25 = 12.25. So, if you take the upper limit and subtract the lower limit you have the tolerance: 12.75 – 12.25 = 0.50. The specified dimension is the part of the dimension from which the limits are calculated. The specified dimension of the example above is 12.5. A dimension on a drawing may be displayed with plus-minus dimensioning, or the limits may be specified as 12.75 and 12.25. Many companies prefer this second method because the limits are shown and calculations are not required. This is called limits dimensioning. A bilateral tolerance is permitted to vary in both the + and the – directions from the specified dimension. An equal bilateral tolerance is where the variation from the specified dimension is the same in both directions. An unequal bilateral tolerance is where the variation from the specified dimension is not the same in both directions. A unilateral tolerance is permitted to increase or decrease in only one direction from the specified dimension.

REFERENCES American Society of Mechanical Engineers, Dimensioning and Tolerancing: ASME Y14.5M1994, American Society of Mechanical Engineers, New York, 1994.

SELECTED BIBLIOGRAPHY Anon., GDT Gets Everyone Working on Design Quality, Quality Assurance Bulletin, number 1315, undated. Karl, D.P., Morisette, J., and Taam, W., Some applications of a multivariate capability index in geometric dimensioning and tolerancing, Quality Engineering, 6, 649–665, 1994. Krulikowski, A., Geometric Dimensioning and Tolerancing: A Self Study Workbook. Quality Press, Milwaukee, 1994.

SL3151Ch10Frame Page 524 Thursday, September 12, 2002 6:03 PM

524

Six Sigma and Beyond

Madsen, D.A., Geometric Dimensioning and Tolerancing The Goodheart-Wilcox Company, Inc., South Holland, IL, 1995. Wearing, C. and Karl, D.P., The importance of following GD&T specifications, Quality Progress, Feb. 1995, pp. 95–98.

METROLOGY The two pillars of the six sigma breakthrough methodology are measurement and project selection. In this section we are going to discuss the issue of measurement from a metrology perspective, and in Chapter 13 we will address project selection under the umbrella of project management. We have known for a long time that we are only as good as our measurement will allow us to be. Therefore, improving measurement systems and understanding the protocol of measurement will benefit the DFSS team immensely. It is very frustrating when you know that what you measure with is not sensitive enough, not accurate enough, or not precise enough. As though that is not enough, we compound the problem with variability (incompatibility) between metrology systems. When a project has been selected for a DFSS investigation/improvement, it is imperative to understand the metrology system in place — the advantages as well as the disadvantages/limitations of that system — and plan accordingly.

UNDERSTANDING

THE

PROBLEM

Incompatibility issues are more than just a nuisance. When system A doesn’t work with system B, users have to stock more than one set of spare parts, train operators on multiple software packages, and produce different types and formats of inspection reports. While these problems are significant, the real difficulty that system incompatibility creates is the inability to exchange a common inspection model between measuring devices. The time it takes to reprogram each device to inspect the same part, refixturing time and cost, and the resulting loss of accuracy in the measurement and inspection process create serious inefficiencies in a system that is, theoretically, designed to provide a high level of process control and improved inspection throughput. (In a design format, think for a moment the ramifications of a design control/test if you are not sure of its capability, accuracy, precision and so on.) These data exchange roadblocks also create time-to-market problems. During the product development and introduction cycle, unnecessary time spent reprogramming control/inspection routines and reevaluating data contributes to lengthy product development cycles with a resulting loss of competitiveness. The effects of system incompatibility are growing rapidly. Genest (2001) has estimated that the problem has cost the automotive industry alone about $1 billion. Cost penalties also exist for metrology system suppliers and individual users who absorb the costs of repair and training necessary when incompatible systems must work together. It is imperative that the leadership (Champion, Master Black Belt, and Black Belt) of a designated project under DFSS understands metrology and uses it effectively.

SL3151Ch10Frame Page 525 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

525

Otherwise, the results do not mean very much. Let us then examine in a cursory fashion metrology. Metrology is a Hellenic word made up of metron = measure of length and logos = reason; logic. In its combined form it means the science of or systems of weights and measures. (It was Eli Whitney who between 1800 and 1811 proved interchangeability would be a vital part of manufacturing, thus developing the first use of the metrology system in the United States.)

METROLOGY’S ROLE

IN INDUSTRY AND

QUALITY

In order for metrology to exist we must recognize that we need measurements to 1. Make things: Length, width and height Eliminate one of a kind 2. Control the way others make things: Designing Building 3. Describe scientific …: Worldwide exchange of ideas Dollar Furthermore, we must understand that metrology is based on standards. Without standards and the ability to trace the standards, measurement would not be possible. Therefore, metrology is based on a hierarchical system of standards with a level of accuracy reflecting the level of the standard under question. The relationship of the hierarchy and accuracy is shown as follows: Hierarchy of Standards

Levels of Accuracy

National Standards National reference Primary reference standards Transfer standard Working standards

Highest level

Gages, fixtures, and instruments

Lowest level

In any metrology system there is a control. The control system is one or more of the following items. 1. Calendar elapsed time: Fixed time — e.g., every 3 months. 2. Amount of actual usage: Number of products checked by equipment. 3. Actual operating hours: Meter actual time equipment draws power. Depending on the control used, there is also a calibration requirement that will ensure that the measurement is “what is supposed to be.” Considerations for any calibration system include the following:

SL3151Ch10Frame Page 526 Thursday, September 12, 2002 6:03 PM

526

Six Sigma and Beyond

Establishment and maintenance of a system: • Written description of system • Flow of system — calibration, repair, and maintenance • Intervals identified Established standards and correlation: • Traceable (National Institute of Standards and Technology, NIST) • Levels Identification: • Equipment identified Part, tool or equipment number • Calibration label Calibration date Next due date Responsible party Recall system: • Scheduled or unscheduled Data recording and analysis: • Computerized • Manual • Characteristics Environmental controls: • Temperature • Humidity Just like any other system, a measuring system may develop inaccurate measurements. Some of the sources of inaccuracy are: • Poor contact — Gages with wide areas of contact should not be used on parts with irregular or curved surfaces. • Distortion — Gages that are spring-loaded could cause distortion of thinwalled, elastic parts. • Impression — Gages with heavy stylus could indent the surface of contact. • Expansion — Gages and parts should obtain thermal equilibrium before measuring. • Geometry — Measurements are sometimes made under false assumptions. For example, part being flat when not flat, or concentric when not concentric, or round when lobed. These inaccuracies may be the result of either accuracy and or precision problems due to: • Operator error — The same operator using the measuring instrument on the same product will come up with a dispersion of readings. • Operator to operator error — Two operators using the same measuring instrument on the same product will exhibit differences traceable to operator technique.

SL3151Ch10Frame Page 527 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

527

• Equipment error — Each piece of equipment has built within it its own sources of error. • Material error — In many cases material cannot be retested, such as in cases of destruct test. • Test procedure error — In cases where two procedures may exist which expect the same outcome. • Laboratory error — In cases of two laboratories performing same test. (Here the reader may want to review Gage R&R and the applicability of accuracy, repeatability, reproducibility, stability, and linearity in Volume V of this series — especially Chapters 15 and 16.)

MEASUREMENT TECHNIQUES

AND

EQUIPMENT

There are many measurement techniques. However, the most typical — at least for the DFSS — are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Linear Angular Force and torque Surface and volume Mass and weight Temperature Pressure and vacuum Mechanical Electrical Optical Chemical

To make these measurements, many types of equipment may be used, some traditional and common and some very specific and unique for individual situations. Typical types of equipment include: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Scale Radius gages Plug gage Thread gage Spline gage Parallels Sine plate Surface plate Caliper Hardness tester Indicator Micrometer

SL3151Ch10Frame Page 528 Thursday, September 12, 2002 6:03 PM

528

Six Sigma and Beyond

13. 14. 15. 16. 17.

Comparator Profilometer Coordinate measuring machine Pneumatic gaging Optical

PURPOSE

OF INSPECTION

As surprising as it may sound, inspection sometimes is used as part of the metrology scheme. It is used with the intention of: • • • • • • • •

Distinguishing good product from bad Distinguishing good lots from bad Checking for process change Comparing process to specification Measuring process capability Ensuring product design intent Determining the accuracy of the inspector Determining the precision of measuring equipment

Inspection is used primarily in three areas, called inspection points. They are: 1. Incoming material • Verification of purchase order • Checking for conformance to specification • Verification of quantity received • Acceptance of certification • Identification 2. In-process: • First piece setup • Verification of process change • Monitoring of process capability • Verification of process conformance 3. Finished product • Last piece release • Verification of product process • Preparation of certification of process It is also important to know that there are three kinds of inspection. They are: 1. 100% inspection • Safety product • Lot size too small for sampling • Seventy-nine percent effective manually 2. Sampling • Large lot size • Decision making

SL3151Ch10Frame Page 529 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

529

3. Visual — mechanical, sensory • Fit and function • Lack of standard • Senses — feel, smell, taste, touch.

HOW DO WE USE INSPECTION

AND

WHY?

Even though we know human inspection is very ineffective, sometimes it is the only option we may have. In any case, we use inspection to classify characteristics of importance and to study and evaluate testing. As for the characteristics we are interested in, they can be summarized as follows:

Critical

Major A

Major B

Minor

Will affect safety Personal injury Product not usable Affect consumer confidence Loss to company Possible injury Possible product not usable Affect consumer Loss to company Inconvenient to consumer Visual Possible loss to company Possible visual Inconvenient to company

As for the testing, we are interested in • • • •

Acceptance Reliability Qualification Verification

of the item tested. Perhaps the most important issue in testing for DFSS is verification. We must be sure that the test is reflective of the “real world usage” and that it addresses “customer functionality.”

METHODS

OF

TESTING

1. Destructive testing • Can only be conducted once • Detects flaws in materials and components • Measures physical properties • Is not cost effective

SL3151Ch10Frame Page 530 Thursday, September 12, 2002 6:03 PM

530

Six Sigma and Beyond

2. Nondestructive testing (N.D.T.) • Is a repeatable test • Detects flaws in material and components • Measures physical properties • Is cost effective Obviously, whenever possible methods of N.D.T. should be used. Typical methods are: • • • • • • • • •

Eddy current X-ray Gamma ray Magnetic particle Penetrant dye Ultrasonic Pulse echo Capacitive Fiber optical

INTERPRETING RESULTS

OF INSPECTION AND

TESTING

All inspections and all tests generate reports by their nature. That means someone must evaluate them and take the appropriate action. Typical issues that the DFSS team should be looking at are: 1. 2. 3. 4. 5. 6. 7. 8.

Accuracy and precision Sampling errors Relation to standards Recording documents Tabulation and calculation Reporting of results Analysis and interpretation Corrective action

Another issue in reporting is the level of reporting as well as the level of responsibility. As a team working on a DFSS project, you should be aware of that relationship. That relationship is shown as Levels of reporting

Levels of responsibility

Operator Auditor Technician Engineer Manager

Manager Engineer Technician Auditor Operator

SL3151Ch10Frame Page 531 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

TECHNIQUE

FOR

531

WRINGING GAGE BLOCKS

One of the most common ways to calibrate certain machinery is through gage blocks. It is very important for the DFSS team to be familiar not only with the blocks themselves but how to use them as well. In this section we are going to address both of these issues. The following points are of special importance when working with gage blocks: 1. 2. 3. 4. 5.

Be sure gaging surfaces are clean. Overlap gaging surfaces about 1/8 inch. While pressing blocks lightly together, slip one over the other. Blocks will now adhere. Slip blocks smoothly until gaging surfaces are fully mated.

By wringing gage blocks together, you can obtain accuracy within millionths of an inch. Caution is usually given not to use a circular action because this might cause serious wear or even damage from abrasive dust trapped between surfaces. Gage blocks are calibrated at the international standard measuring temperature of 68°F (20° C). (This is very important to keep in mind, otherwise see below.) When measurements are conducted at this temperature between blocks and parts of dissimilar materials, no correction for different coefficients of expansion is necessary providing the components have had sufficient time to adjust to the environment. If blocks and parts are made of the same material and are at the same temperature, accurate results are possible regardless of whether the temperature is high or low. To determine the correction requirement when blocks and parts are dissimilar and at temperatures other than 68°F, use the following formula: E = L (∆K)(∆T) where E = the measurement error in microinches; L = nominal dimension in inches; ∆K = difference of coefficients in microinches; and ∆T = deviation of temperature from 68°F. Typical coefficients of expansion in microinches per inch of length per degree F are: Hardened tool steel Stainless steel (410) Chrome carbide Tungsten carbide Aluminum Copper

6.4 5.5 4.5 3.0 12.8 9.4

The gage blocks are typically in a set of 81 pieces and they are arranged in the following order: 9 Blocks — 0001” Series: .1001

.1002

.1003

.1004

.1005

.1006

.1007

.1008

.1009

SL3151Ch10Frame Page 532 Thursday, September 12, 2002 6:03 PM

532

Six Sigma and Beyond

49 Blocks — .001” Series: .101 .111 .121 .131 .141

.102 .112 .122 .132 .142

.103 .113 .123 .133 .143

.104 .114 .124 .134 .144

.105 .115 .125 .135 .145

.106 .116 .126 .136 .146

.107 .117 .127 .137 .147

.108 .118 .128 .138 .148

.109 .119 .129 .139 .149

.110 .120 .130 .140

.250 .750

.300 .800

.350 .850

.400 .900

.450 .950

.500

19 Blocks — 1.000” Series: .50 .550

.100 .600

.150 .650

.200 .700

4 Blocks — 1.000” Series: 1.000

2.000

3.000

4.000

LENGTH COMBINATIONS Do not trust trial and error methods when assembling gage blocks into a gaging dimension. The basic rule is to select the fewest blocks that will suit the requirement. To construct a length of 1.3275″ using a typical 81-piece set, the following procedure may be used: 1. Write the desired dimension on a piece of paper: 1.3275 2. Begin the selection at the top of the gage block set. 3. Reduce the last digit of the dimension to zero by selecting –.1005 a block with a 5 in the fourth decimal place. In this case, the .1005 block is selected and its length is subtracted 1.2270 .1005 from 1.3275 to determine a remainder still to be selected. The value .1005 may be written again in an adjacent column for subsequent proof of the selection. 4. Select a block to reduce the third decimal place to zero. –.107 .107 5. Select a block to reduce the second and first decimals 1.1200 to zero. Wherever possible, such double reductions are –.120 .120 desirable to cut down the total number of blocks selected. 1.0000 6. Complete the selection with the 1,000 block. –1.000 +1.000 0.0000 1.3275 There are times when the same gaging dimension must be assembled more than once from single set of blocks. This may unavoidably increase the number of blocks required for the specific length. Assume that a second length of 1.3275″ is required from the 81-piece set: 1. Write the requirement. 2. Select two blocks to reduce the fourth decimal to zero. In this case, .1002 and .1003.

1.3275 –.2005 .1002 .1003 1.1270

SL3151Ch10Frame Page 533 Thursday, September 12, 2002 6:03 PM

Miscellaneous Topics — Methodologies

3. One block may now be selected to reduce three digits. 4. Select two blocks to form the remaining 1.000. In this case, .400 and .600.

533

.127 .127 1.0000 –1.000 .400 +.600 0.0000 1.3275

For some general rules and guidelines on shapes and basic calculations the reader is referred to Volume II, Part II. Also, for an explanation and examples of the SI system see Volume II, Part II. Volume V of this series covers the issue of GR&R and its terminology and therefore here we give only brief definitions of the key terms: Gage accuracy — Difference between the observed average of measurements and the master value. The master value can be determined by averaging several measurements with the most accurate measuring equipment available. Gage repeatability — Variation in measurements obtained with one gage when used several times by one operator while measuring the identical characteristics on the same parts. Gage reproducibility — Variation in the average of the measurements made by different operators using the same gage when measuring identical characteristics on the same parts. Gage stability — Total variation in the measurements obtained with a gage on the same master or master parts when measuring a single characteristic over an extended time period. Gage linearity — Difference in the accuracy values through the expected operating range of the gage.

REFERENCES Genest, D.H., Improving Measurement System Compatibility, Quality Digest, Apr. 2001, pp. 35–40.

SL3151Ch10Frame Page 534 Thursday, September 12, 2002 6:03 PM

SL3151Ch11Frame Page 535 Thursday, September 12, 2002 6:02 PM

11

Innovation Techniques Used in Design for Six Sigma (DFSS) MODELING DESIGN ITERATION USING SIGNAL FLOW GRAPHS AS INTRODUCED BY EPPINGER, NUKALA AND WHITNEY (1997)

The signal flow graph represents a diagram of relationships among a number of variables. When these relationships are linear, the graph represents a system of simultaneous linear algebraic equations. The signal flow graph, as shown in Figure 11.1, is composed of a network of directed branches, which connect at the nodes. A branch jk, beginning at node j and terminating at node k, indicates its direction from j to k by an arrowhead on the branch. Each branch jk has associated with it a quantity known as the branch transmission Pjk. For our modeling processes, the branches represent the tasks being worked (an activity-on-arc representation). The branch transmissions include the probability and time to execute the task represented by the branch: Pjk = pjkz t jk where pjk is the probability associated with the branch; tjk is the time taken to traverse the branch; and z is the transform variable used to connect the physical system (time domain) to the quantities used in the analysis (transform domain). The z transform simplifies the algebra, as it enables us to incorporate the quantities to be multiplied (probabilities) in the coefficient of the expression, and to include the quantities to be added (task times) in the exponent. The resulting system is then analogous to a discrete sampled data system, and the body of literature on this subject can be applied for the analysis thereof. The path transmission is defined as the product of all branch transmissions along a single path. The graph transmission is the sum of the path transmissions of all the possible paths between two given nodes. The graph transmission is also the resulting expression on an arc connecting the two given nodes when all of the other nodes have been absorbed. In particular, we are interested in computing the graph transmission from the start to the finish nodes. Henceforth, graph transmission shall refer to the graph transmission between the start and the finish nodes and is denoted as Tsf. The coefficient of each term in the graph transmission is the probability associated with the path(s) it represents, and the exponent of z is the duration associated

535

SL3151Ch11Frame Page 536 Thursday, September 12, 2002 6:02 PM

536

Six Sigma and Beyond

P43

3 P13 1

P34 P32

P12

4

P23 P24

2

FIGURE 11.1 A typical branching using signal flow graph.

Concept

Product Design

A

Tooling Design

B

FIGURE 11.2 A simple example with signal flow graph.

A

Z2

B

0.4

Start Z3

Finish

0.6Z3

0.3Z2 A

,

0.7

FIGURE 11.3 A hypothetical design process.

with the path(s). The graph transmission can be derived using the standard operations for the signal flow graphs (discussed below). The impulse response of the graph transmission is then a function representing the probability distribution of the lead time of the process. It can be shown that the expected value of the lead time of the process is:

()

E L =

dTsf dz

z−1

SL3151Ch11Frame Page 537 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

0.18Z5

537

0.18Z5 Z5(0.6z3)

Z5

Tsf

B

0.4

B

0.7

z5(0.4+0.42z3) =

1– 0.18z5

FIGURE 11.4 The graph transmission.

Probability

0.42 0.4

0.076 0.072 5

8

10

13

Time

FIGURE 11.5 First few terms of the probability.

NUMERICAL EXAMPLE A simple example is shown in Figure 11.2. The hypothetical design process is represented by the graph shown in Figure 11.3. The two tasks A and B (product design and tooling design) take 3 and 2 units of time respectively. Once task B is attempted, task A is reworked with probability 0.6, and once A is attempted, B is reworked with probability 0.3. Iterative repetitions of A are represented by task A’. The graph transmission is found using the mode elimination technique (discussed below). This graph transmission is given by the equation shown in Figure 11.4. The expected value of the project lead time E(L) is 7.6 units of time. The first few terms of the probability distribution function are represented graphically in Figure 11.5.

SL3151Ch11Frame Page 538 Thursday, September 12, 2002 6:02 PM

538

Six Sigma and Beyond The distribution can be found for this simple example by performing synthetic division on Tsf to obtain the first few terms of the infinite series. The nominal (once through) time for A and B in series is 5 units of time, which occurs with probability 0.4. It is more likely (probability 0.42) that the lead time L of the process will be 8 units of time.

RULES AND DEFINITIONS OF SIGNAL FLOW GRAPHS HOWARD (1971) AND TRUXAL (1955)

AS INTRODUCED BY

To effectively use signal flow graphs, follow four rules: 1. Signals travel along branches only in the direction of the arrows. 2. A signal traveling along any branch is multiplied by the transmission of that branch. 3. The value of any node variable is the sum of all signals entering the node. 4. The value of any node variable is transmitted on all branches leaving that node. A path is a continuous succession of branches, traversed in the indicated branch directions. The path transmission is defined as the product of branch transmissions along the path. A loop is a simple closed path, along which no node is encountered more than once per cycle. The loop transmission is defined as the product of the branch transmissions in the loop. The transmission T of a flow graph is defined as the signal appearing at some designated dependent node per unit of signal originating at a specified source node. Specifically, Tik is defined as the signal appearing at node k per unit of external signal injected at node j. There are a number of ways of computing transmissions.

BASIC OPERATIONS

ON

SIGNAL FLOW GRAPHS

Solution of signal flow graphs requires knowledge of certain of their topological properties. The basic operations of addition, multiplication, distribution and factoring can be used to reduce the number of branches and nodes in the system. At first glance, it might appear that by successive application of such transformations a graph could be reduced to a single branch connecting any two given nodes. However, if the graph contains a closed loop of dependencies, as when modeling iterations, one or more self loops will eventually appear.

THE EFFECT

OF A

SELF LOOP

The effect of a self loop at some node on the transmission through that node is analyzed in Figure 11.6. The node signal at the first node is x, and the signal returning around the self loop is xt. Since the node signal is the algebraic sum of the signals entering that node, the external signal arriving from the left must equal y(1 – t). Hence, the effect of a self loop t is to divide an external signal by the factor (1 – t) as the signal passes through the node. This holds for all t.

SL3151Ch11Frame Page 539 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

539

t x

y

1

1

y

= 1 + t + t 2 + t 3 + ... =

x

x

1/(1-t)

x

1 1–t

FIGURE 11.6 The effect of a self loop.

tzz

txy txy

z

txz x

txy+1 – t zz

txy txy

y

x

y

FIGURE 11.7 Node absorption.

SOLUTION

BY

NODE ABSORPTION

Node absorption corresponds to the elimination of a variable by substitution in the associated algebraic equations. With the aid of the basic transformations and the self loop replacement, any node in a graph can be absorbed and the equivalent expressions for the transmission between two other nodes calculated. Although the branch is no longer shown, its effects are included in the new branch transmission values, as shown in Figure 11.7. To compute the overall graph transmission, all the intermediate nodes are absorbed in turn, yielding the transmission between the start and finish nodes. Reduction of graphs is computationally intensive, and manual solution of graphs of even moderate size can be difficult.

REFERENCES Eppinger, S.D., Nukala, M., and Whitney, D.E., Generalized models of design iteration using signal flow graphs, Research in Engineering Design, 9(2), 112–123, 1997. Howard R., Dynamic Probabilistic Systems, Wiley, New York, 1971. Truxal, J.G., Automatic Feedback Control System Synthesis, McGraw-Hill, New York, 1955.

SL3151Ch11Frame Page 540 Thursday, September 12, 2002 6:02 PM

540

Six Sigma and Beyond

SELECTED BIBLIOGRAPHY Altus, S.S., Kroo, I.M., and G., P.J. (1996). A generic algorithm for scheduling and decomposition of multidisciplinary design problems, Journal of Mechanical Design, Vol. 118(4), 486–489, 1996. Anderson, J., Pohl, J., and Eppinger, S.D., A Design Process Modeling Approach Incorporating Nonlinear Elements, Proceedings of 1998 ASME Design Engineering Technical Conferences, Atlanta, Sept. 1998. Austin, S.A., Baldwin, A.N., and Newton, A. (September 1994). Manipulating the flow of design information to improve the programming of building design, Construction Management & Economics, 12(5), 445–455, 1994. Austin, S.A., Baldwin, A.N., and Newton, A., A data flow model to plan and manage the building design process, Journal of Engineering Design, 7(1), 3–25, 1996. Austin, S.A. et al., Analytical design planning technique: a model of the detailed building, Design Process Design Studies, 20, 279–296, 1999. Baldwin, A.N. et al., Modelling information flow during the conceptual and schematic stages of building design, Construction Management & Economics, 17(2), 155–167, 1999. Browning, T.R., Exploring Integrative Mechanisms with a View Towards Design for Integration, Proceedings of the Fourth ISPE International Conference on Concurrent Engineering: Research and Applications, Rochester, MI, Aug. 20–22, 1997, pp. 83–90. Browning, T.R., Applying the design structure matrix to system decomposition and integration problems: a review and new directions, IEEE Transactions on Engineering Management, 48(3), 292–306, 2001. Cho, S.-H. and Eppinger, S., Product Development Process Modeling Using Advanced Simulation, Proceedings of the 13th International Conference on Design Theory and Methodology (DTM 2001), Pittsburgh, Sept. 9–12, 2001. Eppinger, S.D., Model-based approaches to managing concurrent engineering, Journal of Engineering Design, 2, 283–290, 1991. Eppinger, S.D. et al., A model-based method for organizing tasks in product development, Research in Engineering Design, 6(1), 1–13, 1994. Eppinger, S.D. and Salminen, V., Patterns of Product Development Interactions, International Conference on Engineering Design, Glasgow, Scotland, Aug. 2001. Gebala, D.A. and Eppinger, S.D.,Methods for Analyzing Design Procedures, Proceedings of the ASME Third International Conference on Design Theory and Methodology, 1991, pp. 227–233. Grose, D.L., Reengineering the Aircraft Design Process, Proceedings of the Fifth AIAA/USAF/NASA/ISSMO Symposium on Multidisciplinary Analysis and Optimization, Panama City Beach, FL, Sept. 7–9, 1994. Johnson, E.W. and Brockman, J.B., Measurement and analysis of sequential design processes, ACM Transaction on Design Automation of Electronic Systems, 3(1), 1–20, 1998. Kusiak, A. and Park, K., Concurrent engineering: decomposition and scheduling of design activities, International Journal of Production Research, 28, 10, 1883–1900, 1990. Kusiak, A. and Szcerbicki, E., Transformation from conceptual design to embodiment design, IIE Transactions, 25(4), 6–12, 1993. Kusiak, A. and Wang, J., Decomposition of the design process, Journal of Mechanical Design, 115, 687–695, 1993. Kusiak, A. and Wang, J., Efficient organizing of design activities, International Journal of Production Research, Vol. 31, 753–769, 1993. Kusiak, A., Larson, N., and Wang, J., Reengineering of design and manufacturing processes, Computers and Industrial Engineering, 26(3), 521–536, 1994.

SL3151Ch11Frame Page 541 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

541

Kusiak, A. and Larson, N., Decomposition and representation methods in mechanical design, ASME Transactions: Journal of Mechanical Design, 117(3), 17–24, 1995. Kusiak, A., Engineering Design: Products, Processes and Systems, Academic Press, San Diego, 1999. McCulley, C. and Bloebaum, C., A genetic tool for optimal design sequencing in complex engineering systems, Structural Optimization, 12(2–3), 186–201, 1996. Osborne, S.M., Product Development Cycle Time Characterization Through Modeling of Process Iteration, master’s thesis (mgmt./eng.), M.I.T., Cambridge, MA, 1993. Rogers, J. L., A Knowledge-Based Tool for Multilevel Decomposition of a Complex Design Problem, NASA, Hampton, VA, Technical Paper TP-2903, 1989. Smith, R.P. and Eppinger, S.D., Identifying controlling features of engineering design iteration, Management Science, 43, 276–293, 1997. Smith, R.P. and Eppinger, S.D., A predictive model of sequential iteration in engineering design, Management Science, 43, 1104–1120, 1997. Smith, R.P. and Eppinger, S.D., Deciding between sequential and parallel tasks in engineering design, Concurrent Engineering: Research and Applications, 6, 15–25, 1998. Smith, R.P. and Morrow, J., Product development process modeling, Design Studies, 20, 237–261, 1999. Steward, D.V., Systems Analysis and Management: Structure, Strategy and Design, Petrocelli Books, New York, 1981. Ulrich, K.T. and Eppinger, S.D., Product Design and Development, 2nd ed., McGraw-Hill, New York, 2000. Yassine, A.A., Falkenburg, D.R. and Chelst, K., (1999) Engineering design management: an information structure approach, International Journal of Production Research, 37(13), 2957–2975, 1999. Yassine, A.A. and Falkenburg, D.R., A framework for design process specifications management, Journal of Engineering Design, 10(3), Sept. 1999. Yassine, A.A et al., DO-IT-RIGHT-FIRST-TIME (DRFT) Approach to Design Structure Matrix (DSM) Restructuring, Proceedings of the 12th International Conference on Design Theory and Methodology (DTM 2000), Baltimore, Sept. 10–13, 2000. Yassine, A.A., Whitney, D., and Zambito, T., Assessment of Rework Probabilities for Design Structure Matrix (DSM) Simulation in Product Development Management, Proceedings of the 13th International Conference on Design Theory and Methodology (DTM 2001), Pittsburgh, September 9–12, 2001.

AXIOMATIC DESIGNS Bad design is, well, bad design. Six sigma, tightening tolerances, substituting one material for another and so on only treat the symptoms, not the problem. Also, they may create expensive bad designs. Axiomatic design, a theory and methodology developed at Massachusetts Institute of Technology (MIT; Cambridge, Mass.) 20 years ago, helps designers focus on the problems in bad designs. As Suh (1990) points out, “The goal of axiomatic design is to make human designers more creative, reduce the random search process, minimize the iterative trial-and-error process, and determine the best design among those proposed.” This, of course, applies to designing all sorts of things: software, business processes, manufacturing systems, work flows, etc. The technique can also be used for diagnosing and improving existing designs.

SL3151Ch11Frame Page 542 Thursday, September 12, 2002 6:02 PM

542

Six Sigma and Beyond

SO, WHAT IS

AN

AXIOMATIC DESIGN?

While “MIT” and “axiomatic” might suggest some lofty academic theory, axiomatic design is well grounded in reality. By definition an axiom is universally recognized principle. One of the earliest uses of axioms was by Euclid, who developed Euclidian geometry from a fundamental set of postulates or axioms. Sir Isaac Newton’s laws of mechanics are another example of axioms. Other fields based on axioms include thermodynamics and information theory. So when we talk about axiomatic we are talking about a systematic, scientific approach to design. It guides designers through the process of first breaking up customer needs into functional requirements (FRs), then breaking up these requirements into design parameters (DPs), and then finally figuring out a process to produce those design parameters. [Does this sound familiar? Y = f(x1, x2…xn)]. In MIT language, axiomatic design is a decomposition process going from customer needs to FRs, to DPs, and then to process variables (PVs), thereby crossing the four domains of the design world: customer, functional, physical, and process. The fun begins in decomposing the design. A designer first “explodes” higherlevel FRs into lower-level FRs, proceeding through a hierarchy of levels until a design can be implemented. At the same time, the designer “zigzags” between pairs of design domains, such as between the functional and physical domains. Ultimately, zigzagging between the “what” and “how” domains reduces the design to a set of FR, DP, and PV hierarchies. Along the way, there are these two axioms: the independence axiom and the information axiom. From these two axioms come a bunch of theorems that tell designers some very simple things, which if followed, can make enormous progress in the quality of their product design. The first axiom says that the functional requirements within a good design are independent of each other. This is the goal of the whole exercise: identifying DPs so that each FR can be satisfied without affecting the other FRs. The second axiom says that when two or more alternative designs satisfy the first axiom, the best design is the one with the least information. That is, when a design is good, information content is zero. That is “information” as in the measure of one’s freedom of choice, the measure of uncertainty, which is the basis of information theory. Designs that satisfy the independence axiom are called uncoupled or decoupled. The difference is that in an uncoupled design, the DPs are totally independent; while in a decoupled design, at least one DP affects two or more FRs. As a result, the order of adjusting the DPs in a decoupled design is important. This order is shown in a design matrix — Figure 11.8 — that shows functional coupling between FRs and DPs at a given level of the design hierarchy. Ideally, these FRs and DPs are to be decoupled.

AXIOMATIC

AND

OTHER DESIGN METHODOLOGIES

In the axiomatic design world, zigzagging between adjacent domains, that is between the “what” domain on the left and the “how” domain on the right, will lead to independent, uncoupled (or at least decoupled) design parameters — namely, “good” designs.

SL3151Ch11Frame Page 543 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

Customer: The benefits customers seek

Customer attributes (Cas) Customer domain

Functional: Functional requirements (FRs) are a minimum set of independent requirements that completely characterize the functional needs of the design solution in the functional domain

Functional requirements (FRs) Functional domain

Physical design parameter (DP) are the elements of the design solution in the physical domain that are chosen to satisfy the specified FRs

Design parameters (DPs) Physical Domain

543

Process: Process variables (PVs) are the elements in the process to produce the product specified in terms of DPs.

Process parameters (PPs) Process domain

FIGURE 11.8 Order of design matrix showing functional coupling between FRs and DPs.

Axiomatic design is not quite the Taguchi method, which is a specific application of robust design. It is not quite quality function deployment (QFD). Nor, like many other quality methodologies, is it an after-the-fact approach that looks at results and then traces back to the source of those results. Robust design (Taguchi) and axiomatic design are the only methods that address the design itself, ensuring that the designs are good to start with. Unfortunately, while Taguchi focuses on making a part immune to the error in variation, it focuses on only one requirement at a time. A problem might arise when a design has to satisfy two requirements simultaneously, such as designing a car door to seal completely and close easily. In short, a coupling exists between these two functional requirements. Taguchi method alone sometimes may trap designers into optimizing the wrong function, optimizing a function they do not have ownership of, or optimizing a design parameter that is linked to many functions. Worse, by optimizing one function, designers run the probable risk of degrading other functions. Axiomatic design, on the other hand, avoids all that by breaking the coupling between functional requirements so that they no longer interact with one another. QFD is similar to axiomatic design in that customer requirements are listed along the left side of a matrix and engineering requirements are lined up along the top. From this matrix, designer teams can see conflicts that need to be resolved. However, QFD is very subjective. Nor does QFD show a mathematical relationship between a functional requirement and a design parameter, which axiomatic design does.

APPLYING AXIOMATIC DESIGN

TO

CARS

The automotive industry is fraught with couplings between design parameters, such as in styling versus aero-dynamic/cooling requirements, in styling versus crashworthiness,

SL3151Ch11Frame Page 544 Thursday, September 12, 2002 6:02 PM

544

Six Sigma and Beyond

and in highly complex automatic transmissions designs. At one carmaker, the axiomatic methodology helps design teams optimize elements of a conceptual design before engineering creates the detailed designs. The described benefit is that it helps avoid unintended consequences in design, with the axiomatic method indicating where interactions exist between the various elements and what the optimum sequence is. The point is that axiomatic design is said to be a step beyond Taguchi and worthwhile in a DFSS endeavor. As we already mentioned, an axiomatic design helps designers with both new and existing designs. In both cases, designers are more creative and develop better designs in less time. New Designs By following the process, the designer designs in a systematic way, completing prerequisite tasks before continuing to the next stage. Accordingly, the designer is more creative by: • Understanding a clearly defined problem before design begins • Identifying innovative ways to fulfill the functional requirements • Saving time by: • Avoiding frustrating dead ends • Drastically reducing random searches for solutions • Minimizing or eliminating design iterations The designer uses current design tools more effectively, producing better designs by: • • • •

Selecting the best design among good alternatives Optimizing the design properly Verifying the design against explicit requirements Having a fully documented design for troubleshooting and extensions

Diagnosis of Existing Design For diagnosing an existing design, the use of axiomatic design highlights problems such as coupling and makes clear the relationships between the symptoms of the problem (one or more FRs not being achieved) and their causes (the specific DPs affecting those FRs). While improving the solution, the designer also enjoys the new-design benefits above. Extensions and Engineering Changes to Existing Designs When an existing version needs an engineering change or an upgrade, axiomatic design identifies all of the areas affected by the contemplated changes. As a result, unintended problems are avoided.

SL3151Ch11Frame Page 545 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

545

To summarize, for both new and existing designs, the designer is more creative, turning out better designs quicker. Efficient Project Work-Flow Axiomatic design helps to identify tasks, set a task sequence from the system architecture, and assign resources effectively. This process also allows you to check progress against explicit FRs. Effective Change Management When creating change, axiomatic design uses explicit criteria and allows you to select the best option, identify effects throughout the system, and document changes. Efficient Design Function Axiomatic design enables use of a common language and shared information between design team members, which preserves institutional learning. The designers’ benefits translate into three categories of benefits for the organization: 1. Competitive advantage: The organization gains a competitive advantage when it satisfies its customers’ needs best. With axiomatic design, those needs map to explicit functional requirements and constraints, which the designers strive to meet. (If, for some reason, no design meets the initial set of FRs and Cs, the firm can explain the tradeoffs of specific alternatives to the customer.) Constraints such as cost and weight can be allocated and verified as the design progresses to ensure they are met. Time to market, another source of competitive advantage, is shortened since designers avoid time-consuming iterations and dead ends. 2. Higher profit: The organization can earn more profit by selling more units, commanding a higher price, or reducing cost. Axiomatic design helps in all three areas. With products that meet customers’ needs better than competitive products, the firm gains market share, resulting in higher unit sales. In addition, meeting those needs better means more value to the customer, who is then willing to pay a higher price. Three types of cost can be lowered: research and development (R&D), cost of goods sold (COGS), and support, for the following reasons: a. The R&D cost is less because designers spend less time designing the product initially and making engineering changes after the product is released. b. COGS drops when products are not coupled and therefore are easier to assemble and test. c. Support costs are lower because products that are not coupled install and set up faster and typically require fewer warranty repairs.

SL3151Ch11Frame Page 546 Thursday, September 12, 2002 6:02 PM

546

Six Sigma and Beyond

Customer Domain “structures” (generative) QFD “patterns” (repetitive)

“events” (reactive)

Functional Physical Domain Domain

Process Domain

TRIZ

TRIZ

TRIZ

QFD

QFD

QFD

DoE Systems Robust Design SPC Engineering DFA, DF... FMEA VA/VE FMEA Find Inspection Find & & Fix Fix

FIGURE 11.9 Relationship of axiomatic design framework and other tools.

3. Less risk: Axiomatic design reduces both technical risk and business risk. Axiom 2, the information axiom, ensures that the chosen design has minimum information content, which is defined as the most technically probable to succeed. Business risk is also reduced because: • Products satisfy customers’ needs since FRs are derived from those needs. • Development schedules are shorter and more predictable. • Upgrades can be done quickly and effectively. In sum, axiomatic design provides the designer with the benefit of designing better products faster, and provides the firm with a competitive advantage, higher profit, and less risk. When you follow the axiomatic design process, you continue to use all of your current software design tools. You will find that you will be more creative, turning out better designs faster, since you will minimize iterations and trial and error. You will have complete documentation of all the design decisions and supporting analysis. To facilitate the analysis of axiomatic designs, to our knowledge the Acclaro™ Software allows you to link to all of your tools and to a common database for the entire design team. It is available commercially and may be purchased by contacting Axiomatic Design Software, Inc. There are a number of techniques used today in design such as QFD, TRIZ, and robust design. The use of these techniques and others is completely consistent with axiomatic design. In fact, axiomatic design can help the designer apply these techniques better. Figure 11.9 shows how they all fit together. Some examples of what these techniques can do are: 1. With QFD (Quality Function Deployment, “the voice of the customer”), designers gather information from customers about their requirements and the relative importance of each. This information helps the designer to choose which FRs must be present and which may be safely ignored. 2. When a designer has selected an FR and wants to identify alternative DPs to achieve it, TRIZ (the theory of inventive problem solving) can be helpful in generating alternatives.

SL3151Ch11Frame Page 547 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

547

3. After choosing a DP to satisfy an FR, the designer uses robust design to optimize the design of this particular DP, which helps to reduce the information content of the design. 4. The designer follows the axiomatic design process and uses the various techniques when appropriate. Axiomatic design helps the designer avoid mistakes such as unknowingly attempting to optimize a coupled design. Users of axiomatic designs and applicable software, such as the Acclaro, have found that the process of implementing axiomatic designs is enhanced, in the sense that the designers have more freedom to document every decision and to specify the relationships between FRs and DPs to any level of detail. It also does matrix manipulations, checks for design problems such as coupling, and communicates relevant information to members of the design team. Specifically, the Acclaro software runs on standard PCs and workstations and in addition it links to your software design tools and to your existing database through SQL. No other software is required except for the Java environment, which is available at no charge from Sun Microsystems or from Axiomatic Design Software, Inc.

REFERENCES Suh, N.P., The Principles of Design, Oxford University Press, New York.

SELECTED BIBLIOGRAPHY Black, J.T. and Shroer, B.J., Decouplers in integrated cellular manufacturing systems. Journal of Engineering for Industry, Transactions of A.S.M.E., 110, 77–85, 1988. Creveling, C.M., Tolerance Design: A Handbook for Developing Optimal Specifications, Addison-Wesley, Reading, MA, 1997. Hill, P.H., The Science of Engineering Design, Holt, Rinehart and Winston, New York, 1970. French, M.J., Engineering Design: The Conceptual Stage, Heinemann Educational Books, London, 1971. Kar, K.A., Linking Axiomatic Design And Taguchi Methods Via Information Content in Design, First International Conference on Axiomatic Design, 2000. Kramer, B.M., An Analytical Approach to Tool Wear Prediction, thesis, MIT, 1979. Kramer, B.M. and Suh, N.P., Tool wear by solution: a quantitative understanding. Journal of Engineering for Industry, Transactions of A.S.M.E., 102, 303–339, 1980. Mohsen, H., Thoughts on the Use of Axiomatic Design Within the Product Development Process, First International Conference on Axiomatic Design, 2000. Otto, K. and Wood, K., Product Design: Techniques in Reverse Engineering and New Product Development, Prentice Hall, Upper Saddle River, NJ, 2001. Rinderle, J. R. and Suh, N.P., Measures of functional coupling in design, Journal of Engineering for Industry, Transactions of A.S.M.E, 104, 383–388, 1982. Stoll, H.W., (1986). Design for manufacture: An overview, Applied Mechanics Review, 39, 1356–1364, 1986. Suh, N.P., Development of the science base for the manufacturing field through the axiomatic approach, Robotics and Computer Integrated Manufacturing, 1, 399–455, 1984.

SL3151Ch11Frame Page 548 Thursday, September 12, 2002 6:02 PM

548

Six Sigma and Beyond

Suh, N.P., Trobophysics, Prentice Hall, Englewood Cliffs, NJ, 1986. Suh, N.P. and Rinderle, J.R., Qualitative and quantitative use of design and manufacturing axiom, CIRP Annals, 31,333–338, 1982. Suh, N.P., Bell, A.C., and Gopssard, D.C., On an axiomatic approach to manufacturing systems, Journal of Engineering for Industry, Transactions of A.S.M.E., 100, 127–130, 1978. Tice, W., The Application of Axiomatic Design Rules to an Engine Lathe Case Study, thesis, MIT, 1980.

TRIZ — THE THEORY OF INVENTIVE PROBLEM SOLVING It has been said that TRIZ is one of the components of customer-driven robust innovation. The other two are QFD and Taguchi. TRIZ is a methodology that was developed in Russia in 1926 by Genrich Altshuller and has been growing all over the world since then. (Because TRIZ is the pronunciation in Russian, many names have been given to this methodology. Some of the most common are: TIPS — Theory of Inventive Problem Solving; TSIP — Theory of the Solution of Inventive Problems and SI — Systematic Innovation.) The foundation of the theory is the realization that contradictions can be methodically resolved through the application of innovative solutions. Terninko, Zusman, and Zlotin (1996) have identified three premises that support the theory. They are: 1. The ideal design is a goal. 2. Contradictions help solve problems. 3. The innovative process can be structured systematically. To be sure, TRIZ focuses on innovation, but what is innovation? According to the founder, Altshuller, there are five levels of innovation. They are: Level 1: Refers to a simple improvement of a technical system. It presupposes some knowledge about the system. It is really not an innovation, since it does not solve the technical problem. Level 2: An invention that includes the resolution of a technical contradiction. It presupposes knowledge from different areas within the relevancy of the system at hand. By definition, it is innovative since it solves contradictions. Level 3: An invention containing a resolution of a physical contradiction. It presupposes knowledge from other industries. By definition, it is innovative since it solves contradictions. Level 4: A new technology containing a breakthrough solution that requires knowledge from different fields of science. It is somewhat of an innovation since it improves a technical system but does not solve the technical problem. Level 5: Discovery of new phenomena. The discovery pushes the existing technology to a higher level.

SL3151Ch11Frame Page 549 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

549

For Altshuller (1997, p. 18–19), the benefit of using TRIZ is to help inventors elevate their innovative solutions to levels 3 and 4. To optimize these levels he suggests the following tools: 1. Segmentation — finding a way to separate one element into smaller elements 2. Periodic action — replacing a continuous system with a periodic system 3. Standards — structured rules for the synthesis and reconstruction of technical systems 4. ARIZ — algorithm to solve an inventive problem. The core tool of TRIZ methodology. It provides nine steps, and they are: • Analysis of the problem: Identify the problem in concise, clear and simple language. No jargon. • Analysis of the problem’s model: Identify the conflict in relation to the overall problem. A boundary diagram may facilitate this. The idea of this step is to focus on the conflict. • Formulation of the ideal final result (IFR): Here you identify the physical contradiction. The process is to identify the vague problem and transform it into a specific physical problem. [Another clue for the Y = f(x1, x2,…, xn)] • Utilization of outside sources and field resources: If the problem remains, imaginatively interject outside influences to understand the problem better. • Utilization of informational data bank: The utilization of standards and databases with appropriate information is recommended here to solve the problem. • Change or reformulate the problem: If at this stage the problem has not been solved, it is recommended to go back to the starting point and reformulate the problem with respect to the supersystem. • Analysis of the method that removed the physical contradiction: Check whether or not the quality of the solution provides satisfaction. A key question here is: Has the physical contradiction been removed most ideally? • Utilization of found solution: Here the focus is on interfacing analysis of adjacent systems. It is also a source for identifying other technical problems. • Analysis of steps that lead to the solution: This is the ultimate scorecard. This is where the former process is compared to the current one. The analysis has to do with the new gap. Deviations, obviously, are recorded for future use. To actually use TRIZ in a design situation, the reader must be aware not only of the nine steps just mentioned but also the 40 principles that are associated with the methodology. Here we are going to list them without any further discussion. The reader is encouraged to see Altshuller (1997), Terninko et al. (1996), and other sources in the bibliography for more details.

SL3151Ch11Frame Page 550 Thursday, September 12, 2002 6:02 PM

550

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

Six Sigma and Beyond

Segmentation Extraction Local quality Asymmetry Consolidation Universality Nesting Counterweight Prior counteraction Prior action Cushion in advance Equipotentiality Do it in reverse Spheroidality Dynamicity Partial or excessive action Transition into a new dimension Mechanical vibration Periodic action Continuity of useful action Rushing through Convert harm into benefit Feedback Mediator Self service Copying Dispose Replacement of mechanical system Pneumatic or hydraulic constructions Flexible membranes or thin films Porous material Changing the color Homogeneity Rejecting and regenerating parts Transformation of properties Phase transition Thermal expansion Accelerated oxidation Inert environment Composite materials

SL3151Ch11Frame Page 551 Thursday, September 12, 2002 6:02 PM

Innovation Techniques Used in Design for Six Sigma (DFSS)

551

REFERENCES Altshuller, G., 40 Principles: TRIZ Keys to Technical Innovation, Technical Innovation Center, Worcester, MA, 1997. Terninko, J., Zusman, A., and Zlotin, B., Step-by-Step TRIZ: Creating Innovative Solution Concepts, Responsible Management Inc., 3rd ed., Nottingham, NH, 1996.

SELECTED BIBLIOGRAPHY Altshuller, G.S., Creativity as an Exact Science, Gordon and Breach, New York, 1988. Bar-El, Z., TRIZ methodology, The Entrepreneur Network Newsletter, May 1996. Braham, J., Inventive Ideas Grow on TRIZ, Machine Design, Oct. 12, 1995, 58. Kaplan, S., An Introduction to TRIZ: The Russian Theory of Inventive Problem Solving, Ideation International Inc., Southfield, MI, 1996. Osborne, A., Applied Imagination, Scribner, New York, 1953. Pugh, S., Total Design — Integrated Methods for Successful Product Engineering, AddisonWesley, Reading, MA, 1991. Taguchi, G., Introduction to Quality Engineering, Asian Productivity Organization, Tokyo, 1983. Terninko, J., Systematic Innovation: Theory of Inventive Problem Solving (TRI7lTIPS), Responsible Management Inc., Nottingham, NH, 1996. Terninko, J., Step by Step QFD: Customer-Driven Product Design, Responsible Management Inc., Nottingham, NH, 1995. Terninko, J., Introduction to TRIZ: A Work Book, Responsible Management Inc., Nottingham, NH, 1996. Terninko, J., Robust Design: Key Points for World Class Quality, Responsible Management Inc., Nottingham, NH, 1989. von Oech, R., A Whack On The Side Of The Head, Warner Books, New York, 1983. Zusman, A. and Terninko, J., TRIZlIdeation Methodology for Customer-Driven Innovation, 8th Symposium on Quality Function Deployment, The QFD Institute, Novi, MI, June 1996.

SL3151Ch11Frame Page 552 Thursday, September 12, 2002 6:02 PM

SL3151Ch12Frame Page 553 Thursday, September 12, 2002 6:01 PM

12

Value Analysis/Engineering INTRODUCTION TO VALUE CONTROL — THE ENVIRONMENT

Among the major problems faced by industry today, two are the cost profit squeeze and ineffective communications. Rising wages and materials costs are squeezing profit against a price ceiling, and our communications systems do not seem to be able to help effect a solution to the problem. We cannot buy labor or material for less than the market cost nor can we sell for more than the consumer is willing to pay. What then is the solution? It is necessary to apply every known effective technique to learn how to thoroughly analyze the elements of a product or service so that we can identify and isolate the unknown, unnecessary costs. In short, it is necessary to make a direct attack on the high cost of business. Value control has been proven to be an effective management tool to seek out and eliminate this hidden cost wherever it may be. It can aid in solving both profit and communications problems, and it can have an effect on operations that will be limited only by your understanding of the techniques and management’s willingness to apply them. Many people are highly skilled at cost analysis and problem solving and think that value control is something we do all of the time. There are many who think that value control is part of every engineer’s job. Some also think that it is something we have done for 20 or 30 years, but we did not call it value control. A primary objective of this chapter is to demonstrate that value control is not only different but is a more powerful technique than any used in the past. Value control is not new, in that it has been around for about 25 years, but it has only been within the past five to ten years that it has been widely accepted. It is a broad scope management tool that considers all of the factors involved in a decision. It goes to the heart of the problem, determines the function to be performed, and applies creative problem solving and business operations such as time and motion study, work simplification, feasibility reviews, systems analysis, etc. But, it follows a systematic organized approach that, in addition, applies unique techniques that identify value control as a special approach to profit improvement. Why, after all these years of scientific and specialized management techniques, is it necessary to develop another technique that does what many of the others were supposed to do? Why is value control necessary? It is still possible for one person to know all that is required to operate a small company or design a simple product. However, our increasingly complex society 553

SL3151Ch12Frame Page 554 Thursday, September 12, 2002 6:01 PM

554

Six Sigma and Beyond

and increasingly complex technology have tended to make most of our managers and technical people specialists in a limited area of activity. They have tended to compartmentalize our operations and, to a large degree, our thinking. The more complex the organization, the more the operation becomes fragmented into autonomous units that deal in a small part of the operation and have an effect on only a small part of the profit. In 1927, one of the greatest technological milestones in the history of human development occurred as the result of the knowledge of two men. Charles A. Lindbergh sat on the Coronado Beach in San Diego with Donald Hall, chief engineer of Ryan Airlines, and established the basic criteria for the Spirit of St. Louis. Two people knew all that was needed to develop an advanced product that even today clearly shows their creative thinking. Lindbergh established the requirements, Hall provided the technical knowledge, and the 13 Ryan supervisors and employees provided the understanding, know-how, and enthusiasm to develop a product that was designed, was built, was tested, and won everlasting glory for Mr. Lindbergh, all within 13 weeks. The product was designed to perform a specific function for a specific cost target. There was no communication problem, there was no cost problem since $15,000 was all they had for the entire project, and there was no timing problem — any delay was unthinkable. Consider the design of an advanced aircraft today. The cost in people and materials is almost beyond comprehension. Hundreds of thousands of people in dozens of industries in several states work in vast industrial complexes for years before the product takes to the air. A product such as the automobile has created a similar situation to the degree that it is a basic national industry and affects people in every corner of the country and in many cases abroad. Is it any wonder that value control is developing on the management scene? It is the only technique that is specifically designed to consider all of the factors involved in decision making — product performance, project schedules, and total cost. Value control is a program of involvement. It makes use of experience from engineering, manufacturing, purchasing, marketing, finance — any area and every area that contributes to the development of a product. It can be used to keep cost out of a product and it can be used to get cost out of a product. It can do this because cost is everywhere and everyone in the organization contributes to it. Cost is the result of marketing concepts, management philosophies, standards and specifications, outdated practices or equipment, lack of time, and incomplete, unobtainable, or inaccurate information, along with dozens of other contributing factors. Every company has at least some of these problems, and it often requires completely new ideas to change them. To prevent and eliminate unnecessary cost we must know how to identify cost. We must be able to identify a problem and be willing to improve the situation. This means change — change in habits, change in ideas, change in philosophies. We know we must change to keep up with the world. Value control enables us to take

SL3151Ch12Frame Page 555 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

555

a good look at all of the factors that must go into making a successful change that will be for our benefit. Value control requires special skills. It is not cost analysis, design reviews, or something we do as part of our job. It is different because the basic philosophy is different. It is not concerned with trying to reduce the cost of an item or service; it is concerned with function and methods to provide the function at the lowest overall cost. Value control concerns people and their habits and attitudes. It is to a large degree a state of mind. It accepts changes as a way of life and makes every effort to determine how change can be made to provide the most benefit. It is a function-oriented system that makes use of creative problem solving and team action. The team is designed to provide an experienced, balanced, and broad scope look at a subject without being constrained by past experience. It requires trained people who understand the system and its application.

HISTORY OF VALUE CONTROL Value control was originated at the General Electric Company. In 1947, Harry Erlicher, vice president of purchases, noted that during the war years it had frequently been necessary to make substitutions for critical materials that quite adequately satisfied the required function and often resulted in an improved product. He reasoned that if it was possible to do this in wartime it might be possible to develop a system that could be applied as a standard procedure to normal operations to increase the company’s efficiency and profit. L.D. Miles was assigned to study the possibility, and the result was a systematic approach to problem solving based on function that he called value analysis. The program was so successful that shortly thereafter the U.S. Navy started to use the system to help get more hardware in the face of a rapidly shrinking budget. The Navy called the program value engineering. Value analysis, value engineering, value management, value assurance, and value control are all the same in that they make use of the same set of techniques developed by Miles in 1947. In many cases, however, the title tends to describe how the system is being applied. Value analysis is generally considered to apply to removing cost from a product. Value engineering and value assurance are applied during the program development phase to keep cost out of a product (our focus in design for six sigma [DFSS]). Value management and value control are overall programs that recognize that value techniques can be applied at any stage of a program. They strive to apply value techniques to control value in all areas of operations. The Navy program developed so successfully that it was picked up by the Department of Defense and is now considered to be the key element in the government’s cost reduction program. In addition, value techniques are now being used in industry and government throughout the free world. They are being applied to aircraft, engines, automobiles, washing machines, dryers, TV sets, and all sorts of consumer and industrial products as well as construction projects and management planning. In addition, several states are applying value techniques to increase efficiency of operations.

SL3151Ch12Frame Page 556 Thursday, September 12, 2002 6:01 PM

556

Six Sigma and Beyond

VALUE CONCEPT The value process is a function-oriented system. It makes use of team action and creative problem solving to achieve results and is specifically designed to simultaneously consider all of the factors involved in decision making — performance, schedule, and cost. In order to obtain an experienced, balanced, broad-scope look at all facets of the project, a carefully selected team is organized to satisfy the specific requirements needed. Selection of the team must consider personality as well as technical competence of the candidates. The team must not only have the technical know-how required, but the members must be compatible and know how to work together. The value manager acts as a coach to guide the team members through the system to obtain maximum benefit from their activities.

DEFINITION The Society of American Value Engineers defines the term “value engineering” as follows: Value engineering is the systematic application of recognized techniques which identify the functions of a product or service, establish a monetary value for that function and provide the necessary function at the lowest overall cost.

The program described by this definition is not a cost reduction program — it is a profit improvement program in that it recognizes all of the factors contributing to product cost. It recognizes that the lowest cost product may induce high warranty problems that may adversely affect profit. In order to increase profit it may, therefore, be necessary to increase product cost in some cases. Overall cost is the prime concern. A value program is implemented by applying all of the known techniques of problem solving and cost reduction plus a large body of special skills. The primary objective is to identify and remove unnecessary cost. Unnecessary cost is cost that can be removed without affecting product function. It has been estimated that the average consumer product may include over 30 percent unnecessary cost. This unintentional cost is the result of habits, attitudes, and all other human factors, and everyone in an organization contributes to it.

PLANNED APPROACH Value control achieves results by following a well-organized planned approach. It identifies unnecessary cost and applies creative problem-solving techniques to remove it. The three basic steps brought to bear are: 1. Identify the function (what does it do?) 2. Evaluate the function (what is it worth?) 3. Develop alternatives (what else will do the job?)

SL3151Ch12Frame Page 557 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

557

FUNCTION Function is the very foundation of value control. The concern is not with the part or act itself but with what it does; what is its function? It may be said that a function is the objective of the action being performed by the hardware or system. It is the result to be accomplished, and it can be defined in some unit term of weight, quantity, time, money, space, etc. Function is the property that makes something work or sell. We pay for a function, not hardware. Hardware has no value; only function has value. For example, a drill is purchased for the hole it can produce, not for the hardware. We pay to retrieve information, not to file papers. Defining functions is not always easy. It takes practice and experience to properly define a function. It must be defined in the broadest possible manner so that the greatest number of potential alternatives can be developed to satisfy the function. A function must also be defined in two words, a verb and a noun. If the function has not been defined in this way, the problem has probably not been clearly defined. Function definition is a forcing technique that tends to break down barriers to visualization by concentrating on what must be accomplished rather than the present way a task is being done. Concentrating on function opens the way to new innovative approaches through creativity. Some examples of simple functions are as follows: produce torque, convert energy, conduct current, create design, evaluate information, determine needs, restrict flow, enclose space, etc. There are two types of functions, basic and secondary. The basic function describes the most important action performed. A secondary function supports the basic function and almost always adds cost.

VALUE After the functions have been defined and identified as basic or secondary, we must evaluate them to determine if they are worth their cost. This step is usually done by comparison with something that is known to be a good value. This means the term value must be understood. Aristotle described seven classes of value: economic, political, social, aesthetic, ethical, religious, and judicial. In value control, we are primarily concerned with economic value. Webster defines value in terms of worth as follows: Value: (1) A fair return or equivalent in goods, services, or money for something exchanged; (2) the monetary worth of something; marketable price; monetary value.

Webster in turn defines worth in terms of value: Worth: The value of something measured by its qualities or by the esteem in which it is held.

If we set a value, we determine its worth. If we determine the worth of something, we set a value on it. The two terms can be used interchangeably, and value is defined

SL3151Ch12Frame Page 558 Thursday, September 12, 2002 6:01 PM

558

Six Sigma and Beyond

for our purpose as follows: Value is the lowest overall cost to reliably provide a function. By overall we mean all costs that affect the function such as design and development expenses as well as manufacturing, warranty, service, and other costs. Value (V-E) is broken down into three kinds, each with a specific meaning: 1. Use Value (V-U): A measure of the properties that make something satisfy a use or service. 2. Esteem Value (V-e): A measure of the properties that make something desirable. 3. Exchange Value (V-x): A measure of the properties of an item that make it possible for us to exchange it for something else. These measures may be in dollars, time, or any other measurable quantity. However, on occasion it is necessary to rank a series of functions by their relative value, one to another. Value is not constant; it changes to satisfy circumstances at a given time. As circumstances change, values may change. This is usually true of monetary, moral, and social values. Value can therefore be expressed as the relationship of three kinds of value: VE = Vu + Ve + Vx Do not confuse cost with value. Cost is a fact; it is a measure of the time, labor, and material that go into producing a product. We can increase cost by adding material or labor to a product, but this will not necessarily increase its value. Value is an opinion based on the desirability or necessity of the required qualities or functions at a specific time. The relationship of cost (C) to value provides an index of performance (P). P = (VE)/C = (Vu + Ve + Vx)/C It can be seen that maximum performance of our resources can be obtained when cost is low and value is high or when P is greater than one. However, it is usually found that P is less than one, and this is an indication that we are not getting good value for our expenditure of funds. It is a direct indication of unnecessary cost.

DEVELOP ALTERNATIVES Function has been defined as the property that makes something work or sell and value as the lowest cost to reliably provide a function. The performance index (P) has identified the problem. Now, what else will do the job? We need to develop alternative ways to perform the function. In order to develop alternatives, we make maximum use of imagination and creativity. This is where team action makes a major contribution. The basic tool is brainstorming. In brainstorming, we follow a rigid procedure in which alternatives are developed and tabulated with no attempt to evaluate them. Evaluation comes

SL3151Ch12Frame Page 559 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

559

later. At this stage, the important thing is to develop the revolutionary solution to the problem. Free use of imagination means being free from the constraints of past habits and attitudes. A seemingly wild idea may trigger the best solution to the problem in someone else. Without a free exchange of ideas, the best solution may never be developed. A skilled leader can produce outstanding results by brainstorming and by providing simple thought stimulations at the proper time.

EVALUATION, PLANNING, REPORTING,

AND IMPLEMENTATION

The creative phase does not usually result in concrete ideas that can be directly developed into outstanding products. The creative phase is an attempt to develop the maximum number of possible alternatives to satisfy a function. These ideas or concepts must be screened, evaluated, combined, and developed to finally produce a practical recommendation. This requires flexibility, tenacity, and visualization and frequently the application of special methods designed to aid in the selection process. The process is carried out during the evaluation and planning phases of the job plan and is covered in detail in those sections. The recommendation must be accepted as part of a design or plan to be successful. In short, it must be sold. It must show the benefits to be gained, how these benefits will be obtained, and finally proof that the ideas will work. This takes time, persistence, and enthusiasm. Details of a recommended procedure are covered in later sections of the text.

THE JOB PLAN These are the basic features that make value control an effective tool. All are applied in a stepwise approach to a value study. The approach, called the job plan, demands specific answers to the following questions: What is it? What does it do? What does it cost? What is it worth? Where is the problem? What can we do? What else will do the job? How much does that cost? The plan is broken down into six steps: 1. 2. 3. 4. 5. 6.

Information phase Creative phases Evaluation phase Planning phase Reporting phase Follow-up phase

SL3151Ch12Frame Page 560 Thursday, September 12, 2002 6:01 PM

560

Six Sigma and Beyond

$ Savings potential

Potential savings

Cost to change

Concept

Design

Development

Production

------------------------------------------➠ time FIGURE 12.1 Relationship of savings potential to time.

Each step is designed to lead to a systematic solution to the problem after consideration of all of the factors involved.

APPLICATION Although indoctrination workshops are usually conducted with existing hardware or systems, the greatest opportunity for savings is in the prevention of unnecessary cost. The function techniques apply, but modifications are required that depend on the user’s understanding and ingenuity in applying them to conceptual ideas. Systems, procedures, manufacturing methods, and tool design are some of the areas other than hardware where functional techniques have been used successfully. Figure 12.1 shows how the overall savings varies with time in the application of function analysis techniques from concept to hardware, system, plan, or any other type of project implementation. The figure also indicates that the cost to change increases and the net savings decreases as a project develops. Once a product or service is in production or use, the cost of tools, hardware, forms, and time necessary to achieve the product stated at any given time cannot be recovered. In order to be effective, value control needs trained people working as a team. A team needs a coach who, in this case, is the value manager. The team provides the technical expertise necessary, and the value manager provides the know-how to apply this knowledge for effective results. Value control also requires management to provide the necessary support and a creative environment. In short, success is up to everyone in an organization. People must be trained; they must understand the system; they must understand the application; they must be aware of cost and how to handle it, and management must support their activity by active participation. Success means change: change in methods, change in procedures, change in attitudes. With this approach, value control will become an effective profit maker.

SL3151Ch12Frame Page 561 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

561

VALUE CONTROL — THE JOB PLAN The six steps in the job plan involve the following questions and activities: 1. Information phase What is it? Collect all data, drawings, blueprints, costs, parts, flow sheets, process sheets. Talk with people, ask questions, listen, develop. Become familiar with the project. Discuss, probe, analyze. What does it do? Define functions. Determine basic function(s). Construct FAST diagram. What does it cost? Conduct function/cost analysis. What is it worth? Establish a value for each function. Determine overall value for the product or source. Where is the problem? Analyze the diagram. Locate poor value functions. Pinpoint the areas for creativity. What can we do? Set goal for achievement. 2. Creative phase What else will do the job? Brainstorm the poor value target functions — use imagination — create alternatives, develop unique solutions, combine or eliminate functions. Look for revolutionary ideas. Do not overlook discoveries obtained by serendipity. 3. Evaluation phase Select best ideas. Screen all creative ideas. Evaluate carefully for useful solutions. Combine best ideas. Categorize into basic groups. Screen for best ideas. How much does it cost? Generate relative costs. Analyze potential. Anticipate roadblocks.

SL3151Ch12Frame Page 562 Thursday, September 12, 2002 6:01 PM

562

Six Sigma and Beyond

4. Planning phase Develop best ideas. Develop practical solutions. Obtain accurate costs. Review engineering and manufacturing requirements. Check quality, reliability. Talk with people. Resolve anticipated roadblocks. Develop alternative solution. Plan your program to sell. Show the benefits. 5. Reporting phase Present ideas to management. Show before and after costs, advantages and disadvantages, non-recurring costs of development and implementation, scrap, warranty, and other forecasts and net benefit. Plan your recommendation to sell. Make recommendation for action. 6. Implementation Ensure proper implementation. Be certain that the change has been made in accordance with the original intent. Audit actual costs.

VALUE CONTROL — TECHNIQUES VERSUS JOB PLAN TECHNIQUES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Define functions. Identify/overcome roadblocks. Use specialty products/processes. Bring new information. Construct FAST diagram. Cost/evaluate FAST diagram. Use accurate costs. Establish goals. All info from best source. Use good human relations. Get all the facts. Blast-create-refine. Get $ on key tolerance. Put $ on main idea. Use your own judgment. Spend company $ as if own. Use company’s services.

SL3151Ch12Frame Page 563 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

18. 19. 20. 21. 22.

563

Specifics not generalities. Use standards. Use imagination. Challenge requirements. Use supplier services.

Items in bold type indicate techniques that apply at every phase of the job plan — as well as in most other activities. Job Plan Information phase What is it? What does it do? What does it cost? What is it worth? Where is the problem? What can we do? Creative phase What else will do the job? Brainstorm Create alternatives Evaluation phase Review suggestions Refine results Evaluate carefully Planning phase Develop best ideas Develop alternative solutions Plan program to sell Reporting phase Present ideas to management Make recommendation for action Implementation phase Ensure proper implementation

Applicable Techniques 1,2,4,5,6,7,8,9,10,11,13,14,15,16,17,18,19,21,22

10,12,15,20,21

2,3,7,8,10,11,13,14,15,16,17,18,19,20,21,22

2,3,5,7,8,10,11,13,14,15,16,17,18,19,20,21,22

2,10,15,18

2,10,15,16,17,18,19,20,22

INFORMATION PHASE DEFINE

THE

PROBLEM

The first phase of the job plan is the information phase. It is broken down into three distinctly separate parts: 1. Information development 2. Function determination 3. Function analysis and evaluation They are all part of the information phase because in reality, they are part of the problem resolution. The work done in the information phase is the basis for the

SL3151Ch12Frame Page 564 Thursday, September 12, 2002 6:01 PM

564

Six Sigma and Beyond

TABLE 12.1 Project Identification Checklist 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Assembly and part drawings Quantity requirements per assembly and annual usage Sample of assembly (project) Sample of each part in assembly (raw cost of stamping blanks if practical) Cost data (material, labor, overhead) Tooling cost (special instructions) Planning sheets (sequence of manufacturing operations including detailed cost breakdown) Specifications (materials, manufacturing, engineering) Required features (special instructions) Name of project engineer

development of alternative methods to perform the required functions. If the functions have not been properly defined and evaluated, the correct questions will not be generated, and the most satisfactory problem solution is not likely to be developed. Information Development Information Collection The first part of the information phase is the development of all available information concerning the project. This includes drawings, process sheets, flow diagrams, procedures, parts samples, costs, and any other available material. Discuss the project with people who are in a position to provide reliable information. Check to be certain that honest wrong impressions are not being collected, that is, information that may have been fact at one time but is no longer valid. It is very important that good human relations be used during this data and information collecting phase. Get the person responsible for the project or development in the first place to help by showing that individual how he or she will be able to profit from successful results of the completed study. The project identification pre-workshop checklist — Table 12.1 — details all of the information required for study. If the data or information are not on hand, it will be necessary to obtain them. A basic information data sheet that should be filled out as a first step to identify the project is shown in Figure 12.2. A brief description of the project should be written under “Operation and performance” to be certain all of the team members are in at least basic agreement as to the product or process operation. If available, a schematic or a picture should also be drawn in this area. Cost Visibility The next step towards a problem solution is to complete the cost visibility section — Figure 12.3 — of the cost-function worksheet as detailed in the cost visibility sheet — Figure 12.4. P.F. costs are estimated as follows: Manufacturing cost = Material + Labor + Burden P.F. cost or Total cost = Manufacturing cost + Other

SL3151Ch12Frame Page 565 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

565

Date: ____________ Drawing or Part number: _____________________

Part Name:

Number required per assembly:

_______________________________

__________________________

Used on major assembly (Name and Number): _______________________________________ Team Number:

Task Force and date:

_____________________

_________________________________

Team members:

Department:

_____________________

________________

Phone: _________________

Present cost:____________________________ Cost Estimate:___________________ Total cost: _________________

Material: ______________

Labor: _________

Overhead: ________

Operation and performance: _______________________________________________ ______________________________________________________________________ Additional comments: ____________________________________________________ ______________________________________________________________________

FIGURE 12.2 Project identification sheet.

Review these cost data in accordance with the preset goals of your project, and make a preliminary judgment of the potential profit improvement. Consider the factors involved and set a goal for achievement that will provide a profitable position. The target should indicate a 30 to 100% cost reduction to be practical. It may seem improbable that this can be achieved; however, it is a target to work toward. A check against this target will be made at the completion of the information phase. Project Scope It is now possible to make a preliminary determination of the project scope. Consider the new project as outlined on the project identification sheet, the present cost and target for improvement, and the time available for the study. After evaluating these factors, define the scope. Limiting or expanding the scope of a study depends on the objective and the time allowed for the study. In project work, the analysis of function should first be performed upon the total assembly or process. If the objectives of the value control study are not achieved at that level, the next lower level

SL3151Ch12Frame Page 566 Thursday, September 12, 2002 6:01 PM

566

Six Sigma and Beyond

Cost Visibility

Team No:

Assembly Part Name:

Sheet

Date:

Assembly Part No:

Determine Manufacturing

Determine Cost Elements

Cost Material $ Item

Reg.

#

Part

Raw

Name

Material

Labor $ Labor $

Burden

Burden $ Other

$

Total

Cost

Component

per unit

Total cost FIGURE 12.3 Cost visibility sheet.

List all functions and separate from constraint Verb

Basic

Second

Remarks

Noun

FIGURE 12.4 Cost function worksheet.

should be studied and so on down to the lowest level of indenture. The lower the level of indenture, the more detailed and complex the study will become. This may require additional time in the present study or future studies to consider segments identified by function analysis.

SL3151Ch12Frame Page 567 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

567

For existing projects seeking to improve performance or utility in general, there are usually existing designs that have cost data available. This will mean that thinking can be focused on a chosen part of a complex assembly. Everything up to and beyond the basic part should be accepted as necessary. Therefore, the initial scope will have been defined. However, as the study progresses, it may be necessary to redefine the scope to ensure completion of the project within the available time and conditions of the workshop. Function Determination The information on hand together with an analysis of costs will tend to define the initial scope of the project. The product or process has been defined and its cost evaluated by use of the cost visibility techniques and a target set. It is now possible to start to define the functions to be performed or that are being performed by the system. Start with the function or functions of the assembly or total system, then break the system down into each part and define the functions of each. Remember to strive to define the functions in two words, and also keep in mind that the definition must not constrain thinking. It is the function definition that will help to visualize new ways to satisfy the function. If it is too constraining, it will tend to restrain thinking. Figure 12.4 should be used for this effort. Most simple products will have at least 20–25 functions. Detailed information on defining functions is covered further on in this section. Function Analysis and Evaluation After the functions have been determined, identify the basic function or functions. The basic function is the function that cannot be eliminated unless the product can be eliminated. There may be more than one, but an effort should be made to determine the one most likely basic function. Determining the basic function is the first step in the construction of a FAST diagram. A detailed discussion on the construction of FAST diagrams is to be found further on in this section. The FAST diagram makes it possible to complete the cost function worksheet. A typical cost function sheet lists all functions versus all parts of a product or actions of a system, procedure, or administrative activity. The objective is to convert product cost to function cost. The cost of each piece of hardware or action is redistributed to the function performed. This proportional redistribution of cost to function requires information, experience, and judgment, and all team members must contribute their expertise. After the cost of each part or action has been redistributed to the functions performed, the cost columns are totaled to obtain the function cost. This cost is then placed on the FAST diagram. The FAST diagram then becomes a very valuable tool. It tells what is happening, why, how, when, and what it costs to perform the function. It is now possible to evaluate the functions to determine if they are worth what is being paid for them. In other words, a value must be set on each function. Determining the value of each function is a subjective process. However, it is a key element in the value process. Comparing the function cost to function value

SL3151Ch12Frame Page 568 Thursday, September 12, 2002 6:01 PM

568

Six Sigma and Beyond

provides an immediate indication of the benefit being obtained for expended funds. The ratio of value cost to function cost is the performance index. The sum of all values is the value of the system or the lowest cost to reliably provide the basic function. It should be compared to the preliminary goal set earlier. It may be that the new goal is considerably higher than the original. If this is the case, an evaluation of the diagram will indicate what must be done to achieve the original goal. It may indicate that an entirely new concept is required, or it may be that it will be acceptable to settle for less. It is often the case that the original goal and the new value are close. An analysis of the function costs will again indicate necessary action. This analysis clearly defines the task for product improvement. It breaks the problem down to functions that must be improved, revised, or eliminated to achieve the goal. It is now possible to proceed to the second phase of the job plan — the creative phase.

COST VISIBILITY Experience has shown, for example, that the automotive industry is a price leadership industry. Experience has also shown that in spite of the tremendous leverage of the industry, it cannot control the prices it must pay for the basic materials required for production. It is then quite clear that we cannot buy materials for less and we cannot sell our products for more. Consequently, only one avenue remains open to increase profit, and this is to identify the areas of high and unnecessary costs and to find ways to reduce or eliminate these costs. In the past, tremendous effort has been made to keep our products at a competitive level. The intent is to add value control as another tool to aid in achieving the desired function of a product at the best cost. Cost visibility techniques are the first to be applied in the value control job plan. Cost visibility techniques are well ordered and range from very simple to highly complex. These techniques do not tell us where unnecessary costs are; they tell us where high costs are. This is important because they identify a starting point. Definitions Since the techniques of cost visibility are concerned with all types of costs, each type will be defined so there is no misunderstanding: Actual cost — Costs actually incurred during the performance of a manufacturing process. They include labor, material, and burden applied in accordance with local ground rules. Allowance — Costs other than material, labor, and burden that must be included in the total cost of a product, such as: packaging materials, scrap, inventory losses, inventory costs, etc. Burden — Includes all cost incurred by the company that cannot be traced directly to specific products. The accounting department determines burden rates. These are assigned to individual operations on a formula basis.

SL3151Ch12Frame Page 569 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

569

Burden consists of both fixed and variable categories, and separate rates are often established for each. The method of assigning burden differs from industry to industry and even from one company to another within an industry. Any quantifiable product factor may serve as a basis for assignment of burden as long as consistent use of the factor across the entire product line results in full and equitable burden distribution. Fixed burden — Includes all continuing costs regardless of the production volume for a given item, such as salaries, building rent, real estate taxes, and insurance. Variable burden — Includes costs that increase or decrease as the volume rises or falls. Indirect materials, indirect labor, electricity used to operate equipment, water, and certain perishable tooling are also included in this classification. Cost — The amount of money, time, labor, etc. required to obtain anything. In business, the cost of making or producing a product or providing a service. Design cost — The sum of material, labor, and variable burden. An understanding of the elements of design is essential for an understanding of cost visibility techniques. Fixed cost — Cost elements that do not vary with the level of activity (insurance, taxes, plant, and depreciation). Incremental cost (sometimes called a marginal cost) — Not all variable costs vary in direct proportion to the change in the level of activity. Some costs remain the same over a given number of production units, but rise sharply to new plateaus at certain incremental changes. The costs thus effected are incremental or marginal costs. Labor — Manpower expended in producing a product or performing a service. Labor may be direct or indirect. Direct labor — Labor that can be traced directly to a specific part. Wages paid the stamping press operator would be classified direct labor. Indirect labor — Labor that is necessary in the manufacturing process but is not directly traceable to a specific part (material handling, inspection, receiving, shipping, etc.) is generally included in burden. Manufacturing cost — The sum of material, labor, and variable burden. An understanding of the elements of manufacturing is essential for an understanding of cost visibility techniques. Material — All hardware, raw (steel, zinc, plastic powders) and purchased (instrument panel knobs, decals, rivets, screws, etc.) items consumed in manufacturing a part. Material may be direct or indirect. Direct material — Raw and purchased material which becomes an integral part of an end item. (The cost of the metal from which a fender is formed would always be a direct material.) Indirect material — Material that is necessary in the manufacturing process but is not directly traceable to a specific part (lubricants, wiping cloths, marking pens, etc.) is generally included in burden.

SL3151Ch12Frame Page 570 Thursday, September 12, 2002 6:01 PM

570

Six Sigma and Beyond

Profit — Amount earned in producing a part or a service. It is usually applied as a percentage of manufacturing cost. Standard cost — A theoretical manufacturing cost developed by engineers and accountants. It is based on manufacturing processes, work measurements, and material sizes and weights developed by engineers and historical and current actual costs furnished by accountants. Standard costs are used to measure the amount of materials, labor, and overhead factors that enter into the manufacture of a product. Dollar values are assigned to these theoretical costs as a common denominator. Standard costs are important as control aids — although they are based primarily on historical data. Total cost — Includes manufacturing cost plus a profit and other expenses. The following expenses are usually added to manufacturing cost by sales and/or accounting departments to make up the total cost: a. Administrative and Commercial Costs — Costs incurred in the administration of the company, research, and selling of the product. They are usually a factor represented as a percentage of manufacturing cost. b. Freight Costs — Costs incurred in getting materials, sub-assemblies and purchased parts to manufacturing or assembly plants. Sources of Cost Information The application of cost visibility techniques begins with an analysis of total cost, progresses through an analysis of cost elements, and finally analyzes component or process costs. To perform these steps the best cost information available is required. This information will be available from sources such as: Accounting — Current and historical costs (actual costs) Purchasing — Cost of purchased items and tools, tool breakdowns, operation line-ups, and material weights, both gross and net Operating facilities cost planning — Estimated costs of parts and tools, process sheets and material weights. Suppliers — Estimates and/or quotations, costs, process information, and material prices. Feasibility and value guidance — Manufacturing feasibility and cost trends. All requests to previously mentioned sources should be channeled through an appropriate Value Guidance Group to coordinate and follow. Cost Visibility Techniques These techniques are not necessarily used in chronological order. We must always use our judgment, not only in utilizing the techniques that indicate high cost, but also in utilizing all the other tools of value control. Cost visibility analysis is based on the information shown in Figure 12.3. Based on the information gathered, the team makes the appropriate recommendations.

SL3151Ch12Frame Page 571 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

571

Technique 1 — Determine Manufacturing Cost First look at manufacturing costs and, basing your judgment on experience or comparison, determine if the part is worth the cost. This technique is simple and obvious — but perhaps it has been overlooked as an indicator as to where high costs may be. If this technique tells us that the cost is high, it is necessary to go to one or more of the other techniques to find out specifically where that high cost is located. Technique 2 — Determine Cost Element If the objective is to locate the high cost, use the second technique which involves determining cost elements. Here, the total direct material, total direct labor, and total burden are broken out of the total manufacturing cost. The elements of total cost again offer a basis for comparison: Determine Cost Elements Material $______________ Labor $______________ Burden $_________ Compare material content with labor dollar content. Compare these elements of cost with those for another similar manufactured item. If the elements of cost vary, it is an indication one may be high in cost, and the reason for the difference must be found. This technique can also be used to arrive at a normal distribution of cost. Accounting can usually determine the normal distribution of cost in material, labor, and overhead for a specific department or profit center. Every part can then be compared to the distribution cost to determine if the cost elements are high or low. Again, comparison is being used to find high cost. The cost breakdown may show that $10 worth of material and $.10 worth of labor are being expended on a certain part. If this is the case, it can be asked if we are in business to spend $.10 on labor for $10 worth of material. Perhaps the material supplier should be asked to perform the labor operation. This could eliminate the labor which may be used more productively elsewhere. Conversely, it may be found that $.10 worth of raw material requires $10 worth of labor. If this is the case, the overhead should be broken down into variances, setups, tooling, direct labor, indirect labor, etc. The manufacturing area should be questioned about methods and processes, profit centers being used, overhead, capital equipment, labor grades, etc. Technique 3 — Determine Component or Process Costs The third technique goes one step further in breaking down material, labor, and overhead. To determine component and element costs as they occur in the manufacture of a part, break down each component as shown in Figure 12.3. Figure 12.3 shows the components broken down in elements. From this list, you examine the reasonable costs versus the unreasonable. The process may sound very subjective at this stage, but it is important to differentiate an item that does not “fit”

SL3151Ch12Frame Page 572 Thursday, September 12, 2002 6:01 PM

572

Six Sigma and Beyond

the pattern of other items. When the examination ends, more than likely you have identified a “most probable” high cost item. Circle this amount, and examine it in detail. Determine why this cost is so far out of line with other operations. This technique gives a very precise and accurate cost visualization. It shows where costs are being created on a component and element basis. Almost every analysis would include the use of techniques 1, 2, and 3. Now think of the third technique in more depth. If we study technique 3 in depth, we will see that it can be used to analyze parts being assembled into a major sub-assembly, major sub-assemblies being put together into a final assembly, and a number of final assemblies being put together to make the total product. Good judgment must be used in the application of this technique, and it will also dictate the way the techniques should be used. Technique 4 — Determine Quantitative Costs This technique analyzes cost on the basis of some measurable unit such as time, weight, size, area, etc., and then makes a comparison with the cost per unit of a known good value. It is sometimes surprising how seemingly complex products will fall into a pattern. One of the most convenient ways to use this technique is to build a cost curve for the product under study. A comparison to the curve will indicate whether the product is high or low. Techniques 1, 2, and 3 can then be used to zero in on the specific cause of the cost deviations. Cost per period of time — This is good for high volume production. It can also be used to describe cost per similar product class. Simply determine the number produced in a convenient time period, minute, hour, day, etc. This can then be compared to a similar unit. A simple example would be the cost per unit of a specific class and size fastener. Cost per pound — This is a basis for comparison usually applied to castings, weldments, or forgings, but it can be applied to anything that will plot on a graph. Determine the cost per pound of each item and plot these on a graph, and the high cost items will be immediately apparent. Again, this is a basis for comparison — another way to find meaningful cost basis. Remember, even though weight may not be an important design criterion, it still costs money to ship every pound of unnecessary weight. Cost per dimension — Some examples of the use of this unit would be as follows: the cost per unit length for a simple extrusion, the cost per unit volume in a tank, the cost per unit length of wiring, The cost per square foot of area covered by a high-cost epoxy paint. These are convenient cost figures to have available as a basis for comparison. Cost per unit of length, area, and volume are the key words of this technique. Cost per functional property — Determine the actual amount spent per functional property. For example, in wiring harnesses, what is the cost per ampere conducted, on a mechanical component, per pound of weight supported or per inch pound of torque transmitted? This gives a basis for a direct comparison. The function can then be evaluated by comparison. This is a basic value control technique.

SL3151Ch12Frame Page 573 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

573

Part Name: Functional

Present cost

Hi

Low

area X Y Total cost

FIGURE 12.5 A form that may be used to direct effort.

The use of these cost analysis techniques will literally explode costs in such a way that a circle can be drawn around the areas that show where work is required. The functional approach techniques can be used to study the high cost area. It does not follow automatically that high cost is unnecessary cost. High cost may be unnecessary cost, but we must use other value tools to find out if it really is. Technique 5 — Determine Functional Area Costs One purpose of this technique is to help answer the question, “where should effort be applied?” If the study item is a part or a simple assembly (two or three parts), then the scope is already defined. If the project is a complex assembly which could have its principle of operation changed by a new design concept, questions such as available time, savings potential, type of improvements, stage of product maturity, etc., should be considered. Divide the present cost into functional areas to define the project scope. Division of cost into functional areas will pinpoint high cost differently than usual cost visibility analysis, and will help to broaden or narrow the scope of study. This will direct effort to more profitable areas. An example is shown in Figure 12.5.

FUNCTION DETERMINATION Function analysis is the foundation of value control. A product or system is not analyzed from a part or action standpoint, it is analyzed from a function standpoint to break down the barriers to visualization for improved creativity and the development of the maximum number of practical alternatives. The objective is to obtain the maximum benefit possible; cost reduction is often as high as 30 to 100 percent. Function analysis makes it possible to set high cost reduction goals and to meet them. This can be done because basic functions are identified and isolated, and other methods to perform them are developed through the use of applied creativity. The function approach requires that certain definitions, ground rules, techniques, and relationships be understood.

SL3151Ch12Frame Page 574 Thursday, September 12, 2002 6:01 PM

574

Six Sigma and Beyond

Experience has shown that function analysis combined with the systematic approach of the job plan will almost invariably produce desired cost reductions. However, the goal of eliminating all unnecessary costs is dependent upon the skill, training, dedication, and organizational support attained. What Is Function? Function is the property that makes something work or sell. Function states what the product or system does. It is the objective of the action, the result to be accomplished, and can be defined in some unit of measure such as weight, quantity, time, money, space, or some other practical measure. Functions are expressed in two words, a verb and a noun. The use of only two words forces a brief or terse definition of the necessary characteristics. The use of two words avoids the possibility of combining functions and of attempting to define more than one simple function at a time. The two word requirement aids in achieving the broadest level of abstraction. It is a forcing technique that causes a struggle to clarify understanding and aid visualization for creativity. Proper identification of function involves a point of view. The function must be identified in such a way that is stripped of all restrictions that would inhibit development of new and better ways to provide the function. For example, consider the fastening of a simple nameplate to a part. One might describe the function that applies as “attach nameplate.” It would be far better to describe the function as identify product, because a nameplate is only one of many ways to achieve the desired function. Nameplates might be riveted, welded, or cemented. However, it is also possible to identify products by etching, stamping, molding, or printing on the part, thereby eliminating the nameplate altogether. Some examples of functions are: Support weight Transmit torque Enclose part Conduct current Amplify voltage

Improve appearance Establish style Increase prestige Decrease cost Create style

Create design Evaluate information Develop plan Survey market Change attitude

Identifying the function in broadest terms provides the greatest potential for value improvement because it allows greater freedom to creatively develop better value alternatives. Further, it tends to overcome any preconceived ideas of the manner by which the function is to be accomplished. Basic and Secondary Functions Basic Functions There are two types of functions: basic and secondary. Basic function is the specific purpose for which a device is designed and made. Or stated another way: basic function is the performance feature that must be attained if the total item or system is to work or sell. Consider a screwdriver. “Transfer torque” is the basic function. If this function is eliminated, the screwdriver will not work.

SL3151Ch12Frame Page 575 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

575

A clear understanding of the user’s need is necessary if a satisfactory basic function definition is to be developed. For example, if the desired application is to pry open paint cans, the function would be defined in terms of the transfer of a linear force. A screwdriver could perform this function but the “transfer force” function may be provided at lower cost if “transfer torque” is eliminated. A plain, flat strip will transfer force without the costly handle. Make sure your study item has a basic function; otherwise, it can be eliminated. Secondary Functions Secondary functions are the result of performance features of a system or item that have been added because of the method chosen to accomplish the basic function. They may help the product work a little better and sell better; in other words, they support the basic function. In the case of the screwdriver, its secondary functions would be: Transmit information Multiply torque Resist corrosion

Upgrade appearance Prevent slip Increase leverage

Resist force Reduce wear Insulate user

Can you determine what parts perform these functions in a typical screwdriver? Secondary functions are sometimes unwanted or unnecessary. An example would be “make noise.” We have a complete sound laboratory trying to eliminate or control noise on our cars. On the other hand, money was added to the turn signal flasher to “increase noise” and then later to “control noise.” In the automobile business, styling is a major factor. Styling features may be basic or secondary. However, whether they are basic or secondary is more subjective than in a mechanical part. For this reason, good supporting marketing data are required to guide and advise the stylist of the consumer’s attitude and requirements.

FUNCTION ANALYSIS

AND

EVALUATION

There are six distinct function evaluation techniques to help clarify problems and identify unnecessary cost. The problem will dictate which techniques are needed. The order in which they are given here has no particular significance; skill develops through application. Practice will eventually provide a methodology that will best fit problem needs. The six techniques are: Technique Technique Technique Technique Technique Technique

1 2 3 4 5 6

Identify and evaluate function Evaluate principle of operation Evaluate basic function Theoretical evaluation of function Input output method Function analysis system technique

Technique 1 — Identify and Evaluate Function This is a simple technique that asks the question “what must the part or assembly do?” It applies to all projects and requires a clear determination of all use and esteem functions. Each function should be expressed in two words — see Figure 12.4.

SL3151Ch12Frame Page 576 Thursday, September 12, 2002 6:01 PM

576

Six Sigma and Beyond

After all functions have been listed, classify them as basic or secondary — refer to definitions of basic and secondary in the ground rules. This technique clarifies the function, prevents combining of functions, and reveals the relationship of basic and secondary functions. Technique 2 — Evaluate Principle of Operation This technique is essentially the same as technique 3, except the emphasis is on principle of operation. This technique requires a detailed examination of the physical laws or effects upon which the function could be based, to allow a simpler, more reliable, and less costly operation. For example, to provide data on auto engine temperature, a “transmit information” function based on laws and effects that respond to heat might be replaced by one based on magnetic principles. This approach has broad application on new items and can be a useful tool in the advance departments or research departments on developmental items. For example, in the development stage, the decision to provide mechanical, electrical, vacuum, or other means to provide automatic temperature control would have an effect on the system design. Technique 3 — Evaluate Basic Function This technique imposes the strictest discipline and requires the acceptance of a forcing assumption: “only basic function has value.” The assumption is made as a mental step in order to force our thinking to search for new and simpler designs that will provide the basic function in such a way that the least number of secondary functions is required to make it work and sell. This technique is best applied to assemblies; however, it can be modified to single parts. The blast-create-refine technique as described in detail in the creative phase is an example of a special case of this technique. The value, as it is developed here, is the combined result of individual judgment, creativity, and past experience that expresses what the function should cost based on the work it performs (and the way it could be done). There are many variations to this technique. One is to expand the scope of study and eliminate imposed functions by revising each listed function determined in technique 2 and asking the question, “Is this function performed this way as a result of the basic design concept?” Redesign to eliminate imposed functions means expanding the scope thereby causing adjoining components to dictate new limits. Some of the largest savings, 50 to 80%, will come using this technique. Technique 4 — Theoretical Evaluation of Function The theoretical evaluation of function places a precise value on a function by using appropriate mathematical relationships. It applies to measurable parameter functions only, such as “create heat” and “resist bending,” as opposed to functions that provide appearance, maintain decor, etc. For example, if we were to plot the cost in cents per foot against the torque carrying capacity for various materials, we would see that the graph instantly will

SL3151Ch12Frame Page 577 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

577

highlight the cost required to satisfy the function “transmit torque.” This approach takes value engineering from an art to a science and opens the door for value research. While the basic concept is still the same, equating cost to function, a considerable grasp of basic value techniques and mathematics is required. Technique 5 — Input Output Method This technique is useful in highlighting the basic function of a product by viewing it as a black box item that receives certain inputs and transforms them into outputs. These inputs and outputs are not functions and therefore do not have to be defined in terms of two words. The function itself is a result of the input and it causes the output; hence, the function is positioned between the input and the output. In the example below, 6 volts DC is the input to the transformer and 12 volts DC is the output. The function that fits between the input and the output is “amplify voltage.” Additional examples of this technique are listed below. Item

Input

Function

Output

Transformer Hot water Heater Pipeline

6 volts DC Cold water Power Fluid

Amplify voltage Heat water Convert energy Transmit fluid

12 volts DC Hot water Heat Fluid

It should be noted that any item may have more than one input or output, and that unless inputs are transformed into outputs, the item has no value. Since function is the key link between input and output, this is equivalent to stating that only function can have value. Technique 6 — Function Analysis System Technique This technique is the primary function analysis technique used in most cases. This system was developed to assist in performing function analysis on a complete system. The use of determination logic helps to identify and verify the basic functions and also helps identify higher and lower level functions and supporting systems. The technique requires construction of a FAST (Functional Analysis System Technique) diagram by the use of determination logic questions: How? Why? and When? The steps necessary to complete a FAST diagram are: 1. List all functions performed by the assembly or system on Figure 12.4. Be sure to identify each function by a verb and noun. Review and check proper columns to identify basic and secondary functions. This is actually technique 2. 2. Prepare a 1″ × 2″ card (or a Post-it note) for each function listed in Step 1. Take a close look at the functions and indicate the relationship of all functions to each other. This requires determination of the next higher level function for each known function. In other words, find the functions that cause other functions to be performed. In order to do this, ask three

SL3151Ch12Frame Page 578 Thursday, September 12, 2002 6:01 PM

578

Six Sigma and Beyond

HOW

WHY Modulate air Achieve comfort

Control air Direct air

FIGURE 12.6 Second step in the FAST diagram block process.

questions about each of the functions listed in Step 1 to identify the functions that will link other functions together. Each question must be answered specifically. The logic questions are: How? How is this function accomplished? Why? Why is this function performed? When? When is this function performed? Select the function you think is the basic function and apply the logic questions to the right and left of the basic function. Ask how the function is performed to determine the function to the right. Ask why this function is performed to determine the function to the left. It may be necessary to select more than one of the functions to get the correct basic function. Ask why? < ----------------------Control Air ----------------------> Ask how? In the example, the function “control air” is selected as the basic function. How is air controlled? The reply is “direct air” and “modulate air.” Both answer the “how” question. Do they both answer the “why” question? Why is the air modulated? Why is the air directed? The answer is to “control air.” The logic questions are satisfied and we can add the next FAST diagram block — see Figure 12.6. Now the question “why do we control air?” must be answered. The reply is “achieve comfort.” How do we achieve comfort? Control air. So the basic logic questions have been satisfied for the basic function “control air”. The basic function has been isolated, and the rest of the primary path functions can be determined. These primary path functions become the basic framework for developing a complete FAST diagram. The “how” and “why” logic questions must now be applied to every function. Each must satisfactorily answer the question relative to its position in the diagram. For example: If we take the function “modulate air,” we can further analyze it into vary opening, direct force, apply torque, apply effort.

SL3151Ch12Frame Page 579 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

HOW

579

WHEN Meet Specs Modulate Air

Scope

WHY

And so on

Scope Apply Torque

Achieve Comfort

Apply Effort

Control Air Increase Concept Control Assembly

Direct Air

And so on

And so on

FIGURE 12.7 A partial cost function FAST diagram.

Whenever these questions are answered satisfactorily, the position of a known function is established within the FAST diagram. In some cases a new function is discovered. Then the primary path questions must be asked of the new function. This step identifies the relationship between a low-level and a high-level function, with the highest-level functions on the left. It identifies functions that are the result of other functions and functions that cause other functions. Unless you understand these relationships, it will not be possible to develop a FAST diagram, which is necessary to stimulate creativity and to clarify the relationships of parts or actions. When the primary path has been selected and positioned on the chart, position all secondary functions that did not fit into the primary path by applying the “when” question and adding them below the primary path. All of the functions listed may not be functions; some may be specifications or objectives. Show the objectives and specifications on your FAST diagram in phantom blocks in the upper left corner of the diagram. This completes the construction of the FAST diagram but does not complete the information that can be added to it to provide the total assembly picture. The function cost worksheet (Figure 12.3) can now be completed by listing all functions horizontally and all parts and process costs as determined from the detailed cost data. Remember: the cost information should be for a specific function. A partial cost function FAST diagram is shown in Figure 12.7. The parts that perform each function can also be added to the FAST diagram. This step will define the high cost areas and point out where to concentrate creative effort. By analyzing the FAST diagram, you can find interesting creative relationships. The function to the right of the selected function tells how this function is

SL3151Ch12Frame Page 580 Thursday, September 12, 2002 6:01 PM

580

Six Sigma and Beyond

performed. The function to the left tells why this function is performed. The function below or above tells when, and that listed immediately below the function tells what performs this function. These simple words — how, why, when, and what — stimulate creativity. The answers also keep your thinking close to the area in which a change is being sought. In further utilization of your FAST diagram, try incorporating secondary functions into existing parts by modification to the part. You will have the most success if the functions are next to each other or happening at the same time. This technique may be applied to existing or proposed designs, concepts, procedures, processes, documents, or any type of software. The primary purpose is to identify functional relationships to stimulate creativity. Cost Function Relationship The FAST diagram clearly identifies functions and their relationship to each other. The techniques of cost visibility identify high cost areas. These techniques can now be combined to clearly identify the relationship between cost and function. This will make it possible to identify the areas of unnecessary cost for the application of creative problem-solving techniques. The cost function worksheet (Figure 12.3) is the basic tool. The functions are listed across the top from the FAST diagram. The parts, processes, or actions are listed vertically with their actual costs — see format in Figure 12.7. It is now necessary to determine the actual cost of each function by applying the cost for the part or action that causes the function to be performed. In many cases, it may be necessary to break the cost down into several functions. For example, say in a foldout sample, the thumbwheel costs $.0957. This is distributed over three functions: provides decor, apply torque, limit rotation. The percentage of cost applied to each is a matter of qualified judgment unless a detailed breakdown can be obtained. In our example, let us assume that the function “modulate air” is made up of the cost of three items totaling .1100. This is the cost of the function “modulate air.” In order to find the cost of the system to modulate air, all of the functions in the critical path plus the supporting functions must be totaled. This cost is, say, $0.2699 or X% of the total assembly cost. In other words, if the modulate air function could be eliminated, $.2699 could be removed from the assembly. These function costs can be applied to the FAST diagram for convenience. This enables a ready determination of what can be accomplished by eliminating or combining functions to provide a less costly assembly. When used in conjunction with the FAST diagram, the cost function worksheet provides an accurate function cost. This can then be evaluated in terms of its value or worth. By the application of creative techniques, new ways to perform the desired function can be developed. Evaluate the Function After a FAST diagram is complete, with part or action costs or assigned to the proper functions, values can be assigned to each of the functions. By assigning cost first,

SL3151Ch12Frame Page 581 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

581

the task force members become familiar with detail costs and are therefore better prepared to assign values to the functions by comparison. Value is defined as the lowest cost to reliably perform a function. In evaluating a function, the value or worth used must be the intrinsic value, not the result or effect of that function, and it does not include other functions on the FAST diagram. (During this phase of the job plan, the team must be optimistic, just as in the creative phase; if not now, when will the team be optimistic?) One of the easiest ways to determine value of a function is by comparison to another method to perform the function at a lower cost. For example, in a given design, the “support weight” function was performed by a columnar part of its attachment, for a function cost of 23 cents. The team assigned a value of 5 cents for the “support weight” function because the team members reasoned that the specified load could be supported in suspension for that amount. At this time, the team did not have a solution to the problem, but during the brainstorming session the team generated proposed changes that were developed to accomplish the overall target. In many cases, function values cannot be assigned by comparison, and other means must be used such as: 1. Apply the test for value — How much of my own money would I pay for that function? 2. Rate function numerically — Apply ratios to function cost to arrive at new values. 3. Apply VE techniques for lower cost —Set a goal or target for functions (percentage reduction). 4. Others — Make use of other information, such as noticeable differences, value standards, and mathematical comparisons. The sum of the individual function values establishes a product or total system value; this becomes the team’s new target. Now the team knows which functions to attach during their creative sessions — the high cost and low value functions. Once these values have been established by the team, place this assigned value in the upper right-hand corner of the function box. The team has isolated the problem and set its own goal(s) for improvement. Remember these ten tests for value: 1. 2. 3. 4. 5. 6. 7.

Can we do without it? Does it need all of its features? Does it cost more than it is worth? Is anyone buying it for less? Is there something better that can do the job? Can it be made by a less costly method? Can a standard item be used? Does it cost more than the total of reasonable costs for material, labor, burden, and profit? 8. Can a less costly tooling method be used, considering the quantities involved?

SL3151Ch12Frame Page 582 Thursday, September 12, 2002 6:01 PM

582

Six Sigma and Beyond

9. Can another dependable supplier provide it for less? 10. Would you pay the price if you were spending your own money?

CREATIVE PHASE The creative phase requires the use of your imagination to develop alternative solutions to the functions defined in Phase I. The systematic value control approach makes use of “brainstorming” as a principal technique; however, the “blast-createdefine” technique must frequently be used in conjunction with others. Brainstorming is defined as the combined effort of two or more people to determine all possible methods for performing the required functions. There is no attempt at evaluation; this will come later. The requirement is to develop any and all ideas that may include the outstanding alternative to satisfy the required functions. It is necessary to become free from the constraints of past habits and attitudes and apply thought needlers — see Table 12.2 — to increase the ideas when they begin to slow down. Refer to specialty processes, products, or materials for ideas. Apply the use of standards. Seek ideas from plant specialists and supplier representatives. Use catalog files such as Thomas Register and Sweets. Remember: • Ideas come from every place and anybody. Do not restrict your thinking! • Conduct a brainstorming session on each required function. List all ideas. • Try to eliminate or combine functions. Be as flexible as possible. There is no end to change. Change is in fact necessary to survival; therefore, people must constantly advance. Our concern in value control is advancement in engineering and manufacturing in a creative and productive sense. Although we would all readily agree with the comments in the preceding paragraph, the individual effort necessary to expand our creative contributions is not automatic but rather requires concentrated effort and deliberate practice. There are several stifling factors that prevent creative productivity from being as free as it could be. For one, customs and traditions that have become a part of our everyday life bind us whether we realize it or not. Second, habits that can be good or bad, depending on the situation, can limit creative productivity. One way to control habits is to first realize that much of what we do and observe in others is determined by habit, and then make a conscious effort to appraise the value of our problem solving habits and attempt to discard those that minimize creative thinking. Unless we progress in this effort, we can become enclosed in a prism of complacency. Inappropriate habits in problem solving can also build a wall of pride about the way we are currently doing things and completely smother our will. In addition, factors that stifle our own creativity are present in those around us. The attitudes of others can be encouraging and stimulating. On the other hand, the attitudes of associates can be stifling when our creative efforts are met with complacency or defensive reactions. Some individuals who have presented good original ideas and then encountered dogma, inertia, minimizers, rationalizers, complacency, apathy, negativism, autocracy,

SL3151Ch12Frame Page 583 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

583

TABLE 12.2 Idea Needlers or Thought Stimulators • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

How much of this is the result of custom, tradition or options? Why does it have this shape? How would I design it if I had to build it in my home workshop? What if this were turned inside out? Reversed? Upside down? What is this were larger? Higher? Wider? Thicker? Lower? Longer? What else can it be made to do? Suppose this were left out? How can it be done piecemeal? How can it appeal to the senses? How about extra value? Can this be multiplied? What if this were blown up? What if this were carried to extremes? How can this be made more compact? Would this be better symmetrical or asymmetrical? In what form could this be? Liquid, powder, paste, or solid? Rod, tube, triangle, cube, or sphere? Can motion be added to it? Will it be better standing still? What other layout might be better? Can cause and effect be reversed? Is one possibility better than the other? Should it be put on the other end or in the middle? Should it slide instead of rotate? Can you demonstrate or describe it by what it is not? Has a search been made of the patent literature? Trade journals? Could a supplier supply this for quicker assembly? What other materials would do this job? What is similar to this but costs less? Why? What if it were made lighter or faster? What motion or power is wasted? Could the package be used for something afterward? If all specifications could be forgotten, how else could the basic function be accomplished? Could these be made to meet specifications? How do competitors solve problems similar to this?

or other stifling conditions will “freeze” creative thought. Others will transfer their creativity to other parts of their lives: home, church, recreation, any place but the job. If we are to encourage creative productivity, we must eliminate any idea that the instant an idea is proposed it must be bitten, broken, or kicked. In order to break ineffective habits and overcome stifling environments, a technique that is helpful is to firmly commit ourselves to a goal to our associates, superiors, or even the general public. In actual process, this technique resolves itself into the establishment of firm deadlines and numerous subdeadlines in the course of a project. You will experience that process during every value control experience. Another technique is the inversion technique. It is used to solve the what-causesit type problem. This technique concentrates on inverting the problem. For example, if the problem is how to cut cost, the technique would ask how you increase the cost effectively.

SL3151Ch12Frame Page 584 Thursday, September 12, 2002 6:01 PM

584

Six Sigma and Beyond

Yet another technique for breaking through our judgment controls of creative expression is that of “blast, create, and refine.” This technique is extremely helpful in reaching value objectives. For years, we have been trying to reduce cost by 5, 10, or 15% through normal cost reduction procedures (material, fabrication methods, etc.) This has become more and more difficult. If we try to take out a larger percentage, say 50%, we are immediately forced to take a new approach to the problem. The blast, create, and refine (BCR) technique combines the function approach with creativity and evaluation of ideas in order to find new, more effective ways to accomplish the required function in products, processes, or procedures. There are several reasons to use the BCR approach; however, the three major ones are: 1. It makes possible more creative problem solutions by eliminating details of the existing product and freeing the mind for thought that could lead to more productive solutions. 2. It directs thinking to basic considerations. 3. It provides a mechanism for building on these basic considerations to develop a final product satisfying all necessary requirements. Intense study of any product shows that it is, to greater or lesser degree, the result of a chain of happenings (evolution). Even the new products that value engineering may bring forth will, to some extent, also exhibit this type of evolution. Therefore the search for better value requires that we ask the following vital questions: How can this chain of influence be stopped? How can we objectively look at a function? The technique of blasting, creating, and then refining is especially directed toward accomplishing these objectives. Its application is in three phases, which are:

PHASE 1. BLAST This phase consists of specifically identifying that portion of the problem under study that does, in fact, perform the basic function (or part or most of it). Next, we blast that portion out of the problem (isolate it) so that we can think about it clearly and specifically. The basic function is the first block in the FAST diagram.

PHASE 2. CREATE In this phase we try to answer the question: What do I have to add to that which I isolated, in the blast phase, to make it capable of performing the required functions or to have it work and sell? Alternatives are developed and costs are put on each one. Make no attempt to evaluate alternatives at this time.

PHASE 3. REFINE We evaluate the ideas developed in the create phase and through an objective process of refining, develop an approach which will meet all the performance, cost, and delivery parameters required.

SL3151Ch12Frame Page 585 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

585

EVALUATION PHASE Evaluating the ideas developed during the creative phase is a critical step in the job plan. The ideas generated will include practical suggestions as well as wild ideas. Each and every idea must be evaluated without prejudice to determine if it can be used or what characteristics the idea has that may be useful. Proper evaluation of the ideas is a critical step. Remember, if an idea is discarded without thorough evaluation, the key to a successful solution may be lost. The time to create ideas is in the creative phase. If an idea is discarded, there may not be another opportunity to develop it again. Evaluation processes can range from the simple to the complex. The methods selected depend to some degree on the number and quality of the ideas generated. (It is not uncommon to have several hundred ideas to evaluate.) In the evaluation process, do not be too critical. Look for the good rather than the bad and do not present unnecessary roadblocks. The initial screening will weed out worthless ideas and sometimes generate new ideas or variations of the present ones. The initial screening will also begin to classify the ideas into basic groups that, in effect, constitute a second stage in the screening process. After the initial screening, it may be necessary to resort to systems designed to aid the process. Two favored, because of their simplicity, are paired comparison and Pareto voting. When the initial list of ideas has been screened and evaluated and reduced to a choice between several alternatives, evaluate the good and bad features of each alternative. Watch out for roadblocks, and try to determine if they can be eliminated and how they may be eliminated. Experience has shown that this evaluation process is a difficult task. The impulse to quickly screen through the list to zero in on the best ideas must be controlled. The mass of data must be handled systematically to obtain maximum benefit from the creative phase. Careful screening is essential to isolating the best concept to carry over into the planning phase where the idea will be developed into a practical recommendation for action.

SELECTION

AND

SCREENING TECHNIQUES

A difficult problem that frequently confronts decision makers is the need to organize a large amount of data so that one or several of the most important items may be identified. It may be required to determine which of several alternatives appears to be the best, or it may be necessary to select a number of items so that they may be ranked and weighted by order of importance or some other criteria. Experience has shown that most people are not able to handle this task quickly and effectively. For this reason, it was decided to develop a simple method that would be applicable in most cases. More complex situations may require more sophisticated methods. However, experience has shown that a combination of two simple methods, Pareto voting and paired comparisons, will satisfy a majority of requirements. Pareto Voting Pareto voting is based on Pareto’s law of maldistribution. Vilfredo Pareto (1846–1923), a political economist, observed a common tendency of wealth and

SL3151Ch12Frame Page 586 Thursday, September 12, 2002 6:01 PM

586

Six Sigma and Beyond

power to be unequally distributed. This observation has been refined to the degree that it can be said that there is an 80/20 percent relationship between similar elements. For example, twenty percent of the parts in an assembly contain eighty percent of the cost. This is most useful information in cost estimating; however, the relationship holds for many diverse examples such as the following: • Twenty percent of the states use eighty percent of the fuel. • Twenty percent of the activities create eighty percent of the budgeted expense. • Twenty percent of the items sold generate eighty percent of the profit. In value engineering, it is frequently necessary to select the best ideas, the highest value functions, the highest potential projects, or any of a number of other requirements. It has been found that the application of Pareto voting can help to simplify the list and will in most cases ensure that the most important items have been selected. It also produces results quickly and can be incorporated into the value engineering process to allow continuous operations without undue disruptions. Pareto voting is conducted by requesting each team member to select what he or she believes are the items or elements that have the greatest effect on the system. This list of items is limited to twenty percent of the total number of items. For example, each team member would be allowed to select six items out of a list of 30. The vote is on an individual basis to obtain as much objectivity as possible. The resultant lists are then compared and arranged into a new consolidated list in descending order by the number of votes each item received. Usually, several items will have been selected by two or more team members. The top 10 to 15 items are then ranked and weighted in a second step by using paired comparisons. Paired Comparisons Paired comparisons, or numerical evaluation as it is sometimes called, compares a list of items to rank and weight them in order of importance or some other criteria. Ranking is the assignment of a preferred order of importance to a list of items. Weighting is the determination of the relative degree of difference between items. In paired comparisons, each item is compared to every other item on the list in turn, using a simple matrix. It is most convenient for up to 15 items; however, the limit is only for convenience. In most cases, ranking and weighting of long lists may be more practically done by direct magnitude estimation (DME). A comparative decision is made between any two items on a two-level basis. There is either a great difference or a minor difference. The decision can be made based on the length of time it takes to decide. If there is no question as to which item to select, there is a great difference. If thought must be put into the decision, it would then be a minor difference. A major difference is weighted a 2, a minor difference a 1. The paired comparison worksheet provides for the list to be evaluated and the evaluation grid. Start by transferring the list of items to the worksheet. Now compare A to B, A to C, etc. comparing A to each item of the list. A is then dropped and B

SL3151Ch12Frame Page 587 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

587

TABLE 12.3 The Worksheet for Setting the List Key Letter A B C D

Alternatives

Weight

Majorca Florida Colorado Greek island cruise

TABLE 12.4 Evolution Summary

A

B

C

D

A2 B

A2 C2 C

A1 D2 D1

compared to C, to D, etc. on through the list. B is then dropped and C is compared to each item on the list until every item has been compared to every other item. The following example will illustrate the process. It is desired to select a vacation from among the following areas: Majorca, Florida, Colorado, or Greek island cruise. The first step is to list the locations on the evaluation summary area of the worksheet as shown in Table 12.3. The second step is to begin to compare the items. Evaluation Summary From the evaluation summary list, compare A to B, Majorca to Florida, and place the selected location letter in the A-B box of the evaluation grid. If the difference is major or clearly in favor of A, place a suffix 2 after the letter A. The A-B box should read A2. Now compare A to C. If the selection is A, place an A in the A-B box. If the difference is great, again add the suffix 2. Now compare A to C. If A is again the selection, place the A in the A-C box. If it requires thought to make the decision, the numerical suffix should be 1, minor. Drop the A and now compare B to C and B to D. Lastly, drop the B and compare C to D — see Table 12.4. To determine the ranking and weighting, add up the As, Bs, Cs, etc. In the example the result is as shown in Table 12.5. This analysis shows Majorca to be the most desirable. It is 40 percent more desirable than a Greek island cruise and 60 percent more desirable than Colorado. Matrix Analysis Although Pareto voting and paired comparison satisfy the screening and evaluation process in most cases, there are times when a more detailed analysis is required.

SL3151Ch12Frame Page 588 Thursday, September 12, 2002 6:01 PM

588

Six Sigma and Beyond

TABLE 12.5 Ranking and Weighting Key Letter A B C D

Alternatives

Weight

Majorca Florida Colorado Greek island cruise

5 0 2 3

TABLE 12.6 Criteria Affecting Car Purchase XXXX — Paired Comparison

A

B

C

D

E

F

G

Coding & Results

A1 B

A1 B1 C

D1 B1 C1 D

E1 E1 E1 E1 E1

F1 F1 F1 F1 F1 F

A1 G1 G1 G1 G1 F1

F – Cost 6 G – Economy 4 E – Image 4 A – Styling 3 B – Comfort 2 C – Reliability 1 D – Selection 1

TABLE 12.7 Criteria Weighing Criteria

Weight Alternatives Ford Chrysler Chevy Honda Audi

A

B

C

D

E

F

G

3

2

1

1

4

6

4

Total

Rank

Two such cases could be when a decision involves large financial outlays or when serious consequences could result from a change. In these cases, every effort must be made to base a decision on the most objective data possible. For many of these decisions, there is a need to rank and weigh a number of alternatives against a series of specific criteria. By doing this, we learn which trade-offs must be made for the

SL3151Ch12Frame Page 589 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

589

various requirements of the project, enabling us to make the best decision. In these cases, a combinex method is recommended. Combinex was developed by Fallon (1971) and is based on comparing a number of alternatives to a series of criteria. Each alternative is compared to the criteria in turn and given a specific numerical rating. The resultant analysis clearly ranks and weighs each alternative against each criterion, which allows for trade-offs based on clearly defined data. This makes it an excellent tool in decision making. Example To illustrate the process, a typical problem familiar to most people will be used. The problem is to select an automobile for purchase. The criteria for selection have been taken from a list of factors affecting the sale of most products. The criteria selected will have a different value for each individual and have been chosen to illustrate several points. The selection criteria are: A. B. C. D.

Styling Comfort Reliability Selection (models available)

E. F. G.

Image Cost Economy (mi/gal)

In other instances, the criteria used could be the factors affecting the purchase of manufacturing equipment, location of a plant, construction of various types of facilities, or any other requirement involving a series of criteria for selection. The alternatives to be considered for purchase are the XXXX models listed below along with their fictitious base prices. The analysis was made in April XXXX. The same analysis made in September XXXX might have resulted in a different conclusion as time and opinions change. Alternatives 1. 2. 3. 4. 5.

Ford Plymouth Chevrolet Honda Audi

$ $ $ $ $

14,000 13,600 14,500 15,000 28,000

Rank and Weigh Criteria

The first step in the process is to decide the importance of the various criteria since each does not have an equal weight or bearing on the selection. In other words, the selection criteria must be ranked and weighed. To do this we will use the method of paired comparisons. A team of five persons applied paired comparisons as seen in Table 12.6. The result of the group’s analysis is their opinion. Another group would apply their own values and probably produce a different result. This group’s ranking and weighing shows cost to be the most important criterion. Cost was six times more important than comfort and 50 percent more important than economy.

SL3151Ch12Frame Page 590 Thursday, September 12, 2002 6:01 PM

590

Six Sigma and Beyond

TABLE 12.8 Criteria Comparison Criteria

Weight Alternatives Ford Chrysler Chevy Honda Audi

A

B

C

D

E

F

G

3

2

1

1

4

6

4

2/ 4/ 4/ 4/ 3/

4/ 4/ 4/ 3/ 4/

3/ 4/ 3/ 3/ 3/

4/ 5/ 4/ 3/ 3/

3/ 3/ 3/ 3/ 4/

3/ 3/ 3/ 3/ 1/

4/ 4/ 4/ 5/ 3/

Total

Rank

TABLE 12.9 Criteria Weight Comparison — Completed Matrix Criteria

Weight Alternatives Ford Chrysler Chevy Honda Audi

A

B

C

D

E

F

G

3

2

1

1

4

6

4

2/6 4/12 4/12 4/12 3/9

4/8 4/8 4/8 3/6 4/8

3/3 4/4 3/3 3/3 3/3

4/4 5/5 4/4 3/3 3/3

3/12 3/12 3/12 3/12 4/16

3/18 3/18 3/18 3/18 1/6

4/16 4/16 4/16 5/20 3/12

Total

Rank

67 75 73 74 57

4 1 3 2 5

Evaluate Each Alternative

The criteria values are entered into the combinex scoreboard as illustrated in Table 12.7. Next, the team compares each alternative, in turn, to each of the criteria. A value is then placed in the upper section of its respective box. These values are based on the criteria weighing scale shown below. 5 4 3 2 1

Superior Good Average Fair Poor

In this example, the comparison was made as shown in Table 12.8. How does the Ford satisfy the styling criteria in the opinion of the selection team? The team decided it was fair and rated it a 2. For reliability, the team said the Ford was average and weighed it a 3. After the Ford was compared to each

SL3151Ch12Frame Page 591 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

591

criterion in turn, the second alternative, the Chrysler, was compared. In each case, each team member expressed an opinion individually. In some instances, it was necessary to develop an average. In other cases, the decision was unanimous. This was done until each alternative was compared to each criterion. The third step of the process is to multiply the criteria weight by the comparison value as shown in Table 12.9. For example, the Ford styling weight of 3 was multiplied by the value of 2. The resultant product of 6 is inserted in the lower section of the box. After completion of each individual weighing, the score is summed under the total column. The total score is shown in the column at the right, and the choices are ranked in the far right column. This analysis shows the first choice to be the Chrysler and the last choice to be the Audi, as illustrated in the complete combinex scoreboard (Table 12.11). Analyze Results

An analysis of the table shows that although the Audi was a poor fifth in the selection process, the primary reason was cost. If the cost had been average, the additional 12 points would have raised Audi’s total above that of the Ford. The table also shows that if the Ford styling had been rated as good, 4, this car would have been ranked second with a score of 73. Although styling was originally ranked fourth in importance with a 3, other factors may now be considered. An improvement in reliability would not have a major effect on the overall rating, but a reduction in cost or an improvement in economy could have. Cost could be negotiated; economy would require some basic product changes.

IMPLEMENTATION PHASE The objective of a value engineering study is the successful incorporation of recommendations into the product or operations. However, a successful project often starts back at the beginning. Each project must be thoroughly analyzed to determine its potential for benefit and the probability of implementation. This is as important as the knowledge and skill required to apply the system to attain successful results. An excellent idea is worthless unless it can be properly implemented. If it is not implemented, no one will obtain any benefit. It must also be implemented in the manner intended. Unfortunately, there have been many cases on record where the idea could not be implemented because of the high cost to make the change. There are other cases where the recommendations were not properly understood and implementation resulted in increased cost. This often results in disillusionment or the feeling that value engineering does not work on our problems. Actually, in most cases the real problem was that the problem was not properly diagnosed. It was not that value engineering does not work; it was a matter of inefficient preliminary analysis and preparation. It does not seem reasonable to expend the effort and funds required to make a value study without first having done the necessary work to ensure that the project is practical, that it can be implemented, and that the necessary funds and people will be available.

SL3151Ch12Frame Page 592 Thursday, September 12, 2002 6:01 PM

592

Six Sigma and Beyond

Selection of projects is a part of the entire value engineering implementation process. Many times, management will assume that any project will prove profitable. This is not always the case. The project must be practical in relationship to its effect on the organization — see the discussion on SIVE. To aid in the selection of projects, development of people, implementation of projects, and all the other aspects necessary to successfully achieve the stated objective, we have prepared some guidelines. They are guidelines, not rules, as every organization is different and successful value engineering efforts must be integrated into operations to become part of the day-to-day decision-making process of the company. To begin with, we will look at the overall organization and implementation of value engineering operations. Then we will look at some of the details that make for success.

GOAL

FOR

ACHIEVEMENT

What do we want to get from value engineering? What will be the objective? This is the first question to answer. Value engineering can increase productivity, reduce product cost, improve quality, reduce administrative costs, or produce a number of other benefits that may be critical to operations. Whatever the goal, it should be defined in specific terms, such as increase productivity by a specific percent, reduce product cost by a specific number of dollars per unit, and so on. Whatever the initial goal may be, it can be revised and broadened as skill in application and implementation of the process develops and understanding and credibility increase. Value engineering is a people-oriented program designed to help people to do a better job by aiding them to break down constraints to understanding. It provides some very specific methods and systems to achieve results. Since people perform a wide range of jobs in an organization, it is certainly logical to expect that if they can be provided with a system that can help them to do a better job, anything that they are expected to do can be improved. In the end it is people who do the thinking. If they can improve their performance, everyone will benefit. This has been our experience. Many people who are highly skilled in their jobs have developed new insights that have created breakthroughs in technology as well as major organizational and operational improvements. The goal for achievement should be known to everyone. It can be product oriented or directed towards manufacturing or administrative operations. It need not be company wide. However, the scope can be broadened at any time. Once the goal has been determined, the means to achieve the objective can be developed.

DEVELOPING

A

PLAN

There are five steps to incorporating value engineering into operations. They are as follows: 1. Evaluate the system. 2. Define an objective. 3. Develop plan and organization to achieve objective.

SL3151Ch12Frame Page 593 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

593

4. Understand the principles. 5. Implement the plan. Each step can be approached in a number of different ways. However, certain specific problems must be considered, and pitfalls must be avoided in each. Understanding the problems and pitfalls rather than outlining a specific method or procedure should provide the necessary guidelines for an effective operation. In many cases, a consultant can aid in the initial stages and support each step of the process by providing a broad range of experience for the client to build upon. However, it is important that the consultant have the type and quality of experience to ensure success.

EVALUATION

OF THE

SYSTEM

Evaluation of value engineering can be a very tricky process. Some companies have spent large sums of money for educational training/seminars and are not using the systems in any way. Some companies did not understand the principles and, when they tried to apply them, found that they had neither the skill nor the discipline to achieve success. There are still others who feel a highly organized cost reduction program is value engineering. As a result, there are some who feel that value engineering works but not on their product. There are others who feel that value engineering is nothing new; it is the same thing they have done for years under a different name. And, of course, there are some who ask the classic question, “Who has the time for all this?” Evaluation of the benefits to be obtained from value engineering should therefore be based on at least some prior knowledge of the methods and disciplines so questions can be asked to determine what is being done. Are the principles of function analysis and evaluation being applied? Is the function analysis system technique (FAST) used? How is the creative stage handled? How are the projects selected and organized? How is the team approach used? What authority does a value engineering team have to implement projects? How are teams selected? How is the operation organized? These are key questions that are required to evaluate whether the company actually has been using value engineering based on the principles established by Miles (1961) and supported by the Society of American Value Engineers (SAVE). A major element of the evaluation process should be a one-day orientation for key management — those who will be required to support operations with time, manpower, and funds. The orientation should be presented by one who has had successful experience conducting value engineering operations within the constraints and limitations of daily operations. Preferably, the person should be certified as a Value Specialist (CVS) by SAVE. To just understand the principles is often not enough. How to make them work in an operating environment is frequently at least of equal importance. As in everything, future success is based on a firm foundation.

UNDERSTANDING

THE

PRINCIPLES

Very early in the plan to introduce value engineering into operations, high-level and operating management must be introduced to the system. The intention is not to

SL3151Ch12Frame Page 594 Thursday, September 12, 2002 6:01 PM

594

Six Sigma and Beyond

teach them value engineering but to demonstrate the benefits to be achieved and how they are produced. This establishes the need to apply the process and defines the necessary commitments for success. Those who should attend would be everyone who will be expected to support operations with time, manpower, and funds. It is difficult for a large group of high-level people to attend a one-day seminar. However, it is essential for successful operations. Attendance also broadcasts the message of importance to all levels of the organization. In addition, the managers attending often derive substantial benefit from the session that can lead to immediate results. The one-day orientation should be a case study, so participants can try the various methods and systems. The result will be understanding of the system and how it may be applied to various projects. It will identify the organizational and operational pitfalls and in many cases define projects for future workshops. Completion of the management orientation will create a need for a decision to determine how operations will proceed from this point. If a consultant has been brought in to aid in progressing to this point, the consultant will now be able to assist in getting down to brass tacks. If one has not been brought in, now would be the time. The consultant’s experience can ensure success from the start and increasingly successful performance as skill develops. At this point there are two ways to go. However, in the long run, the same objective will be achieved. One approach is a large multi-team workshop or series of workshops directed towards indoctrinating a large group of people (30–40) in the system at one time. These people would learn the process while applying the methods and systems to projects of current interest to the company. These workshops usually develop substantial monetary benefit for the company. The second approach is one or two teams working on a specific project. Both methods can be successful. However, the first is better suited to very large organizations with large amounts of manpower. The second can be used in both large and small organizations and produces substantial benefit that can be used for further development. In many cases, a combination of the two plus a series of orientations can be used effectively. The specific plan depends entirely upon the organization and should be tailored to fit.

ORGANIZATION The first step is to determine the objective, as was discussed earlier. The second should be to develop a plan to achieve the objective and set up the necessary organization. The third step is implementation of the plan; the fourth follow-up and audit operations. The essential elements are: 1. 2. 3. 4.

Define the objective Develop the plan Implement the plan Conduct follow-up and audit operations

Upon completion of the evaluation and the making of a decision to implement value engineering operations, the first step should be to appoint a coordinator. A

SL3151Ch12Frame Page 595 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

595

brief outline of factors to be considered in selecting a value engineering coordinator or manager is: Primary purpose of position • Establish the value engineering business discipline as part of the fiber and decision-making process of the company to increase the opportunity to maximize the profitability of all products marketed by the company. • Plan, staff, and direct a value engineering program to provide maximum product value by the application of recognized techniques to identify and eliminate unnecessary cost in products and operations. • Develop and implement a program to educate key employees, management, and suppliers in the value engineering approach to problem solving with particular emphasis on function and value. • Publicize and demonstrate the use of value engineering techniques to company management and suppliers to develop support and participation in the use of value engineering and in the implementation of recommendations. Knowledge and skills requirements • Degree in engineering, business, or economics with a thorough understanding of technical aspects of product design and development, business operations and economic factors involved • Value engineering training • Three or more years in value engineering program operations and a thorough understanding of the techniques and methodology as applied to both product development and manufacturing operations • Minimum of ten years combined experience in product management, project engineering, manufacturing management, or product development with a thorough understanding of procurement practices, systems analysis, cost, estimates, or any of a number of other broad rather than specialized product areas • Creativity and flexibility in planning and thinking, with demonstrated leadership abilities necessary to organize and guide persons of widely divergent backgrounds into an effective team • Ability to communicate effectively in both oral and visual techniques The coordinator will develop and organize a plan for management approval. Inherent in the plan should be education and application programs for all who will be involved in operations. The coordinator should be required to select a consultant, develop an educational plan, aid in organizing and conducting workshops, and identify people who may be developed into value specialists. The extent of these programs will depend upon the size and scope of the company. From what we have noted here, it is obvious that the problem is complex from the standpoint of options. However, successful operations do not have to be extensive. Starting small and developing successfully is preferred to a lot of noise and a big crash because of poor planning.

SL3151Ch12Frame Page 596 Thursday, September 12, 2002 6:01 PM

596

Six Sigma and Beyond

ATTITUDE One of the most important factors in value engineering is attitude — attitude on the part of both management and people on task teams. A positive, cooperative, supportive attitude is required. In many cases, value engineering actually requires a new management style. It cuts across organizational lines, looks at taboo aspects of a problem, and recommends drastic changes compared to the past. To accept these disruptions to the old way of doing business requires faith and understanding — a positive attitude. In many cases, whenever a new idea is presented to an American management team the initial reaction is negative. The first remarks are, “It is interesting but let me tell you what is wrong with it.” The best approach to this reaction is to listen carefully. The managers may have some ideas you overlooked. After all negative reaction has run out, be prepared to ask some specific positive questions of the group that will develop positive responses. For example, “I understand your difficulty in producing this part in the plant. What do you think we would have to do to make this practical? Do you see any changes we might make to satisfy our methods?” This will usually work to a positive result. Never argue. In many cases it is beneficial to solicit negative ideas, but be prepared to develop positive questions. Our attitude is that we must begin to ask “What’s good about this idea?” “How will it help us to do a better job?” Changing people’s attitudes is difficult and may never happen, but understanding the reasons behind the negative reaction should make it possible to persuade most people that they can benefit from success. Remember, there is a risk of failure in new ideas. New ideas require change, and they may not work. People want proof that something will work before they will support it. However, maybe you can show that the benefits are greater than the risk. The best way to change people’s attitude is to show that top management is interested in value engineering and expects participation and results in achieving the stated goals.

VALUE COUNCIL The value council is a small group of high-level executives who oversee operations. In a small company, it might be chaired by the president; in a large company, by a division manager. The council should be staffed with people who have the authority to make decisions relative to acceptance and/or rejection of proposals, authorizing funds, and manpower. They set the attitude, develop the environment, break the bottlenecks, and by their interest and visibility create credibility to participation and provide authority to operations. It is important that members of the council make every effort to attend council meetings except in cases of dire emergency. A member who is unable to attend should authorize a key assistant to act on his or her behalf. If the council attendance degenerates, the message sent is that we are losing interest.

SL3151Ch12Frame Page 597 Thursday, September 12, 2002 6:01 PM

Value Analysis/Engineering

597

The council should be made up of five to six people. Their duties are as follows: • • • • •

Set objectives Guide operations Monitor progress Eliminate roadblocks Recommend/approve projects

AUDIT RESULTS There are two reasons to audit results. The first is to determine the actual benefit received. Is it in accordance with expectations? If not, why not? The second is to determine how to improve operations. A periodic status report on a project tends to move it along. This is especially true of cost reductions.

PROJECT SELECTION 1. Develop awareness to potential Products Operations Planning Investments 2. Selection methods Intuitive Scientific 3. Considerations Noncompetitive product Low volume High warranty Quality problems Vendor problems Manufacturing difficulties Capital investments High manpower requirements Bottlenecks Potential market Government regulations 4. SIVE analysis List potential projects Potential saving Implementation cost Confidence factor Project priority R = S/C × F

S C F R

SL3151Ch12Frame Page 598 Thursday, September 12, 2002 6:01 PM

598

Six Sigma and Beyond

Confidence factor Poor Questionable Fair Good Very good 5. Example 1 S= $60,000 C= $10,000 F= 1 R= 6

F 1 2 3 4 5 2 $20,000 $10,000 5 10

3 $2000 $500 4 16

CONCLUDING COMMENTS This is a very brief outline of some of the factors to be considered to implement value engineering operations in your organization. The complete subject would require an entire book but even then there would be many exceptions. Value engineering is a task force type system. Set up the group, get the job done, dissolve the group, get on with the next problem. It is people oriented; it is designed to get maximum performance from the individual and capitalize on that person’s performance by supplementing it with the group. Of course, there must be some type of staff, and they must be skillful in application or know-how to get the people who can produce results. Remember, success is based on the three As: attitude, awareness, application. There must be a positive attitude in the organization — an awareness of the need to change and the skills to apply systems for effective results. If these guidelines are followed, it has been proven that the benefits will be almost immediate and far greater than the usually expected results. They are often outstanding.

REFERENCES Fallon, C., Value Analysis to Improve Productivity, Wiley, New York, 1971. Miles, L., Techniques of Value Analysis and Engineering, McGraw-Hill, New York, 1961.

SELECTED BIBLIOGRAPHY Fowler, T.C., Value Analysis in Design, Van Nostrand Reinhold, New York, 1990. Mendelson, S. and Greenfield, H.B. Taking value engineering/value analysis into the twentyfirst century, Cost Engineering, Vol. 37, No. 8, August, pp. 33–34, 1995. Mudge, A.E., Numerical evaluation of functional relationships, Proceedings, Society of American Value Engineers, Vol. 2, pp. 111–123, 1967. Penza, P., Measuring Market Risk with Value Risk, Wiley, New York, 2000. Shillito, M.L. and DeMerle, D.J., Value: Its Measurement, Design and Management, John Wiley & Sons, New York, 1992. Stakgold, I., Green’s Functions and Boundary Value Problems, 2nd ed., John Wiley & Sons, New York, 1997.

SL3151Ch13Frame Page 599 Thursday, September 12, 2002 6:01 PM

13

Project Management (PM)

Project management (PM) is the application of knowledge, skills, tools, and techniques in order to meet or exceed stakeholder requirements from a project. Meeting or exceeding stakeholder requirements means balancing competing demands among: 1. Scope, time, cost, quality, and other project objectives 2. Stakeholders — customers — with differing requirements 3. Identified requirements and unidentified requirements (expectations) Knowledge about project management can be organized in many ways. In fact, the official Guide to the Project Management Body of Knowledge (PMBOK) has identified 12 subsections (Duncan, 1994). They are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Project management The project context The process of project management Key integrative processes Project scope management Project time management Project cost management Project quality management Project human resource management Project communications management Project risk management Project procurement management

It is beyond the scope of this book to cover the entire discipline of project management. However, this chapter will address PM as it may be used in six sigma and design for six sigma (DFSS) initiatives within an organization. Towards that end, this chapter will discuss some of the basic concepts of project management and how the methodology of project management may be used.

WHAT IS A PROJECT? Projects are tasks performed by people, constrained by limited resources, describable as processes and subprocesses, that are planned, executed, and controlled within definite time limits. Above all, they have a beginning and an end. Projects differ from operations primarily in that operations are ongoing and repetitive while projects 599

SL3151Ch13Frame Page 600 Thursday, September 12, 2002 6:01 PM

600

Six Sigma and Beyond

are temporary and unique. A project can thus be defined in terms of its distinctive characteristics — it is a temporary endeavor undertaken to create a unique product or service. Temporary means that every project has a definite ending point. Unique means the product or service is different in some distinguishing way from all similar products or services. Projects are undertaken at all levels of the organization. They may involve a single person or many thousands. They may require less than 100 hours to complete or over 10 million. Projects may involve a single unit of one organization or may cross organizational boundaries as in joint ventures and partnering. Examples of projects include: 1. 2. 3. 4. 5.

Developing a new product or service Effecting a change in structure, staffing, or style of an organization Designing a new product Developing a new or modified product or service Implementing a new business procedure or process

Temporary means that every project has a definite ending point. The ending point is when the project’s objectives have been achieved, or when it becomes clear that the project objectives will not or cannot be met and the project is terminated. Temporary does not necessarily mean short in duration. It means that the project is not an ongoing task, therefore is finite. This point is very important, since many undertakings are temporary in the sense that they will end at some point, but not in the same sense that projects are temporary. For example, assembly work at an automotive plant will eventually be discontinued, and the plant itself decommissioned. Projects are fundamentally different because the project ceases work when its objectives have been attained, while nonproject undertakings adopt a new set of objectives and continue to work. The temporary nature of the project may apply to other aspects of the endeavor as well: The opportunity or market window is usually temporary — most projects have a limited time frame in which to produce their product or service. The project team seldom outlives the project — most projects are performed by a team created for the sole purpose of performing the project, and the team is disbanded and members reassigned when the project is complete. A project or service is considered unique if it involves doing something that has not been done before and is therefore unique. The presence of repetitive elements does not change the fundamental uniqueness of the overall effort. Because the product of each project is unique, the characteristics that distinguish the product or service must be progressively elaborated. Progressively means “proceeding in steps; continuing steadily by increments,” while elaborated means “worked out with care and detail; developed thoroughly” (American Heritage Dictionary, 1992). These distinguishing characteristics will be broadly defined early in the project and will be made more explicit and detailed as the project team develops a better and more complete understanding of the product.

SL3151Ch13Frame Page 601 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

601

Progressive elaboration of product characteristics must not be confused with proper scope definition, particularly if any portion of the project will be performed under contract. In contrast to a project, there is also a program. A program is a group of projects managed in a coordinated way to obtain benefits not available from managing them individually (Turner, 1992). Most programs also include elements of ongoing operations, as well as a series of repetitive or cyclical undertakings. (It must be noted, however, that in some applications program management and project management are treated as one and the same; in others, one is a subset of the other. It is precisely this diversity of meaning that makes it imperative that any discussion of program management versus project management must have a clear, consistent, and agreed-upon definition of each term.)

THE PROCESS OF PROJECT MANAGEMENT The process of project management is an integrative one. The interactions may be straightforward and well understood, or they may be subtle and uncertain. These interactions often require trade-offs among project objectives. Therefore, successful project management requires actively managing these interactions, so that the appropriate and applicable objectives may be attained within budget, schedule and constraints. A process from a project management perspective is the traditional dictionary definition, which is a “series of actions bringing about a result” (American Heritage Dictionary, 1992). In the case of a project, there are five basic management processes: 1. Initiating: Recognizing that a project should be begun and committing to do so 2. Planning: Identifying objectives and devising a workable scheme to accomplish them 3. Executing: Coordinating people and other resources to carry out the plan 4. Controlling: Ensuring that the objectives are met by measuring progress and taking corrective action when necessary 5. Closing: Formalizing acceptance of the project and bringing it to an orderly end Operational management — the management of ongoing operations — also involves planning, executing, and controlling. However, the temporary nature of projects requires the addition of initiating and closing. To be sure, these processes occur at all levels of the enterprise, in many different forms, and under many different names. However, even though there are many variations, it is imperative to understand that operational management is an ongoing activity with neither a clear beginning nor an expected end. Finally, it must be understood that these processes (initiating, planning, executing, controlling and closing) are not discrete, one-time events. They are overlapping activities that occur at varying levels of intensity throughout each phase of the project. In addition, the processes are linked by the results they produce: the result

SL3151Ch13Frame Page 602 Thursday, September 12, 2002 6:01 PM

602

Six Sigma and Beyond

TABLE 13.1 Key Integrative Processes Project Plan Development

Project Plan Execution

Overall Change Control

1. TABInputs 1. TABInputs 1. TABInputs Outputs of other processes Project plan Project plan Historical information Supporting detail Progress report Organizational policies Organizational policies Change request Constraints and assumptions 2. TABTools and techniques 2. TABTools and techniques 2. TABTools and techniques Project planning methodology Technical skills and Change control system knowledge Progress measurement Stakeholder skills and knowledge Work authorization system Additional planning Project management information Status review meetings Computer software systems Project management Reserves information system Organizational procedures 3. TABOutputs 3. TABOutputs 3. TABOutputs Project plan Work results Project plan updates Supporting detail Change requests Corrective action Lessons learned

or outcome of one becomes an input to another. Among the central processes, the links are iterated — planning provides executing with a documented project plan early on and then provides documented updates to the plan as the project progresses. It is imperative that the basic process interactions occur within each phase such that closing one phase provides an input to initiating the next. For example: closing a design phase requires customer acceptance of the design document. Simultaneously, the design document defines the product description for the ensuing implementation phase. For more information on this concept see Duncan (1994), Kerzner (1995), and Frame (1994).

KEY INTEGRATIVE PROCESSES In project management, the key integrative processes are: • Project plan development — taking the results of other planning processes and putting them into a consistent, coherent document • Project plan execution — carrying out the project plan by performing or having performed the activities included therein • Overall change control — coordinating changes across the entire project Although the processes seem to be discrete from each other, that is not the case in practice. In fact, they overlap and interact in ways that are beyond the scope of this book. A typical summary of key integrative process is shown in Table 13.1.

SL3151Ch13Frame Page 603 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

603

PROJECT MANAGEMENT AND QUALITY As we have seen, project management is a problem-solving methodology. On the other hand, both six sigma and DFSS are a “process project” that require total acceptance for improvement. For that improvement to occur, six sigma and DFSS commitment must be understood and implemented in the entire organization as a culture change first and then for the project itself. As such, it fits the profile of project management. Every component of it is designed to facilitate the solving of complex problems. It uses teams of specialists. It makes use of a powerful scheduling method. It tightly tracks costs. It provides a mechanism for management of total improvement and customer satisfaction. It depends on the integration of several skills and disciplines. It encourages monitoring of processes and depends on feedback for evaluation. It requires leaders with clear vision and doable objectives. It requires knowledge of appropriate and applicable tools. And it plans for success. Project management makes and at the same time facilitates change(s). By definition, projects have a start, work accomplished, and a finish. The finish comes when the objectives for the project are satisfied. Project objectives always address changes that will be made in some current situation. If an organization does not want to make a change, then project management is not an appropriate management method. This does not imply that changes should not be made there, only that there is no motivation for change. In such an organization, the introduction of project management would have little support and may even encounter resistance. For a discussion on change and when change actually takes place see Stamatis (1996). Since the implementation of both six sigma and DFSS is a project, with a beginning, work changes, and an end, project management is indeed a method that can be used in the implementation process. (It is very important to differentiate the concept of six sigma and DFSS, which are philosophical in nature, and the implementation of six sigma and DFSS, which is a project. Here we are talking about the physical implementation of both six sigma and DFSS projects.)

A GENERIC SEVEN-STEP APPROACH TO PROJECT MANAGEMENT Much has been written about how to use project management in a variety of industries and specific situations. Many articles and books have proclaimed specific approaches for the best results in a given situation. Rather than dwell on a particular approach, we will present a summary discussion of a generic seven-step approach for using project management in a quality orientation for any organization. The seven steps are based on the four-phase cycle of any project.

PHASE 1. DEFINE

THE

PROJECT

Step 1. Describe the Project Describing a project is not as simple as it might seem. In fact, this step may be the most difficult and time consuming. To be successful, the project description should

SL3151Ch13Frame Page 604 Thursday, September 12, 2002 6:01 PM

604

Six Sigma and Beyond

include: simple specifications, goals, projected time frame, and responsible individuals, as well as constraints and assumptions. Capturing the essence of highly complex projects in a few words is an exercise in focus and delineation. However, we must be vigilant about avoiding becoming too simple and in the process failing to convey the scope of the project. On the other hand, a detailed, complex description may cloud the big picture. The key is clarity without an excess of volume or jargon. Step 2. Appoint the Planning Team After describing the project, begin to identify the right players. Too many people on a team can stifle the decision-making process and reduce the number of accomplishments. Cross-functional teams are among the most difficult to appoint. Except in the pure project organization, where the team is solely dedicated to completing the project, roles and priorities can cause conflict. In cross-functional teams the project leader must seek support from the functional managers and identify team goals. Step 3. Define the Work Once the planning team is in place, team members must define the work. Since each member hails from a different department, there will be many different concepts of the project’s work content. There are many ways to divide the work for convenient use in planning. Two common ways are the process flow diagram and the work breakdown structure (WBS). The method should be chosen to reflect the most useful division and summarization for the situation. After all, the objective of this step is to define the tasks to be done, not the order of doing them.

PHASE 2. PLAN

THE

PROJECT

Step 4. Estimate Tasks Before a project schedule is created, each task must be evaluated and assigned an estimate of duration. There are essentially two ways of looking at this process. The first way is to establish the duration of the task by estimating the time it takes to complete the task with given resources. The second way is to estimate the type and amount of resources needed and the effort in terms of resource hours that is necessary to complete the task. Step 5. Calculate the Schedule and Budgets The next step is to construct a network logic diagram or a performance evaluation review technique (PERT) and a budget. The focus of the logic diagram or the PERT is to develop appropriate scheduling datelines and more importantly to define the critical path. The focus of the budget is to estimate the costs of the project based on all activities. The identification of the critical path will zero in on the bottleneck areas as well as opportunities for improvement. Tasks not on the critical path may have a float that can be calculated and may be used to facilitate the efficiency and utilization of resources without affecting the project final date.

SL3151Ch13Frame Page 605 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

PHASE 3. IMPLEMENT

605

PLAN

THE

Step 6. Start the Project The kick-off of the project can really make an impact on project team members’ attendance, performance, and evaluation. Kick-off meetings should convey the following ideas: • • • •

This is a new project. Project management is going to be used to manage the project. A plan exists, open to all, which is going to be followed. The focus will be on the starts of activities (ends cannot happen without starts). • Realistic status is needed to allow timely decisions. • The focus will always be on forecasting and preventing problems.

PHASE 4. COMPLETE

THE

PROJECT

Step 7. Track Progress and Finish the Project The essence of this step is to bring the project to closure. That means that the project must be officially closed, and all deliverables must be handed over to the stakeholders — customers. In addition, a review of the lessons learned must take place, and a thank you for the project team is the appropriate etiquette. Key questions of this step are: • • • • • • • • • •

Where are we? Where should we be? What do we have to do to get there? Did it work? Where are we now? Can the process employees take over? Can the process employees maintain the new system? What have we learned from the successes in this project? What have we learned from the failures in this project? What would we have done differently? Why? Why not?

A GENERIC APPLICATION OF PROJECT MANAGEMENT IN IMPLEMENTING SIX SIGMA AND DFSS Project management brings together and optimizes (the focus is always on allocation of resources) rather than maximizes (concentrating on one thing at the expense of something else; maximization leads to suboptimization) resources, including skills, cooperative efforts of teams, facilities, tools, information, money, techniques, systems, and equipment.

SL3151Ch13Frame Page 606 Thursday, September 12, 2002 6:01 PM

606

Six Sigma and Beyond

TABLE 13.2 The Characteristics of the DFSS Implementation Model Using Project Management Phase 1

Phase 2

Phase 3

Management Commitment

Structure Setup

Implementation

Establish a six sigma Capture company and DFSS objectives implementation Define: team of one person Mission Values from each Goals functional area Strategy Train those selected in the six sigma and Focus on continual improvement DFSS requirements Develop policies and procedures Reconfirm quality management commitment

Make goal of six sigma and DFSS total improvement Examine internal structure and compare it to the goals of six sigma and DFSS Determine departmental objectives Review structure of the organization Review job descriptions Review current processes Review control mechanisms Review training requirements Review all communication methods Review all approval processes Review supplier relationship(s) Review risk considerations and how they are addressed Review all outputs Review all action plans

Phase 4 Working with Employees Provide applicable and appropriate training Prepare the organization for both internal and external audits Provide and/or develop appropriate and applicable methodology for corrective action Continue focus on improvement

Why should project management, as opposed to other management principles, be used in the six sigma and DFSS implementation process? There are at least two reasons. First, project management focuses on a project with a finite life span, whereas other organizational units expect perpetuity. Second, projects need resources on both part-time and full-time bases, while permanent structures require resource utilization on a full-time basis. The sharing of resources may lead to conflict and requires skillful negotiation to see that projects get the necessary resources to meet objectives throughout the project life. Since we already have defined the process of both six sigma and DFSS implementation as a project, indeed then, project management will ensure successes of the implementation process by following the generic four phases of a project’s life. A typical approach is shown in Table 13.1. Tables 13.2 and Table 13.3 show the characteristics of the six sigma and DFSS implementation model and process using project management.

SL3151Ch13Frame Page 607 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

607

TABLE 13.3 The Process of Six Sigma/DFSS Implementation Using Project Management Phase 1

Phase 2

Phase 3

Management Commitment

Structure Setup

Implementation of plan

Initiate Project

Understand Process

Provide Six Sigma and DFSS Training

Management planning and goal setting Departmental commitment Quality team selection and active participation Training philosophy and tools of quality Process definition and selection Identification of critical processes and characteristics

Team flow charting for process understanding and analysis Cause and effect analysis Critical in-process parameters identified Standard operating procedures review, equipment repair, preventive maintenance, and calibration Process input and measurement evaluation Static process data collection Process evaluation

THE VALUE

OF

PROJECT MANAGEMENT

Executive training Departmental training Identification of shortcomings in the system of quality (specific areas) Definition of boundaries of responsibility Definition of limitations of resources Review of system for completeness

IN THE IMPLEMENTATION

Phase 4 Working with Employees and Suppliers Monitor Progress Worker/operator control in process Define quality system as it relates to current policies and practices (quality manual, procedures, instructions, and so on) Internal audits conducted Definition of key characteristics and monitoring of process variables Application of statistical process control in all key processes Initiation and follow up of corrective action

PROCESS

Project management is a tool that helps an organization to maximize its effort in implementing a project. Since the process of implementing both six sigma and DFSS — or any other quality initiative — is a project, the value of project management can be appreciated in at least two areas: 1. Planning the process 2. Setting reliable, realistic and obtainable goals Planning the Process Four steps define the planning process from a project management perspective. They are:

SL3151Ch13Frame Page 608 Thursday, September 12, 2002 6:01 PM

608

Six Sigma and Beyond

1. Identify and prioritize the customer base by contribution to current and future organizational profits. 2. Identify and weight criteria key customers use in selecting organizations and assess what changes to criteria or weighting are likely to occur in the future. 3. Assess the organization’s competitive advantages and disadvantages in each area important to decision makers. 4. Establish long-term strategic objectives by identifying where the biggest gap exists between what is important to key customers and the organization’s own strengths and weaknesses relative to competition. To optimize the output of these four steps the following questions may be raised: • Is there a true management commitment for the project? • Does the project address needs of the organization’s top priority customer groups? • Does the project address important needs of the customer? • Is the organization far ahead of competition in this area already? • Does this project truly offer the organization a good chance of making an improvement large enough to change customer behavior? • Will the project require investment large enough to wipe out potential gain? • How does the project rank on the above criteria in relation to other possible projects? • Once the project is selected, is the team continuously assessing whether or not the project is the best one to move the department and organization toward their goals? Goal Setting There are three basic steps in goal setting from a project perspective. They are: 1. Translate corporate strategy into concrete organizational goals that are attainable within a reasonable time. 2. Involve department managers in internal audit and benchmarking exercises to identify problem areas related to the goals. 3. With department managers, set specific improvement goals for each department and each team.

PM

AND

SIX SIGMA/DFSS

Harry (1997, p. 21.14) posed five questions in reference to projects. They are: 1. What do you want to know? 2. How do you want to see what it is that you need to know?

SL3151Ch13Frame Page 609 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

609

3. What type of tool will generate what it is that you need to see? 4. What type of data are required of the selected tool? 5. Where can you get the required type of data? These questions, upon further probing, will deliver some very impressive results. However, the concern remains: How would a Black Belt or even a Master Black Belt go about getting the correct answers to these questions? We believe the answer lies with strategic planning and persistence to the basic format of PM. That is, in the language of PM identify: • • • • • • • • •

Work breakdown structure Work packages Time-scheduled network diagrams Responsibility-assignment matrix Risk analysis and quantification Earned value analysis Project integration management plans Resource costing Project change management

And, in the language of six sigma/DFSS: • • • • • • • • • • • • • • •

Select key project/product. Define performance variables. Create the SIPOC model. Measure current performance and capability. Conduct a benchmarking. Identify and evaluate gap. Identify success factors and goals of project. Select the performance variables. Evaluate new performance. Confirm causal variables. Establish operating limits and verify performance improvement. Validate control system. Implement control system. Audit. Monitor.

Ultimately, all projects in the six sigma/DFSS world are managed in the following four categories: 1. 2. 3. 4.

Project justification and prioritization techniques Project planning and estimation Monitoring and measurement of project activity Project documentation and related procedures

SL3151Ch13Frame Page 610 Thursday, September 12, 2002 6:01 PM

610

Six Sigma and Beyond

Project Justification and Prioritization Techniques Justification and prioritization of projects are based upon the following methods: Benefit-cost analysis: Return on investment (ROI) Internal rate of return (IRR) Return on assets (ROA) Payback period Net present value (NPV) Decision analysis and portfolio analysis as applied to project decisions Benefit-Cost Analysis Project benefit-cost analysis is a comparison to determine if a project will be (or was) worthwhile. The analysis is performed prior to implementation of project plans and is based on time-weighted estimates of costs and predicted value of benefits. The benefit-cost analysis is used as a management tool to determine if approval should be given for the project go-ahead. The actual data are analyzed from an accounting perspective after the project is completed to quantify the financial impact of the project. The sequence for performing a benefit-cost analysis is: • • • • • •

Identify the project benefits. Express the benefits in dollar amounts, timing, and duration. Identify the project cost factors including materials, labor, resources. Estimate the cost factors in terms of dollar amounts and expenditure period. Calculate the net project gain (loss). Decide if the project should be implemented (prior to start) or if the project was beneficial (after completion). • If the project is not beneficial using this analysis, but it is management’s desire to implement the project, what changes in benefits and costs are possible to improve the benefit-cost calculation?

Return on Assets (ROA)

Johnson and Melcher (1982) give an equation for return on assets (ROA) as:

ROA =

Net Income Total Assets

where net income for a project is the expected earnings and total assets is the value of the assets applied to the project. Return on Investment (ROI)

ROI =

Net Income Investment

SL3151Ch13Frame Page 611 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

611

where net income for a project is the expected earnings and investment is the value of the investment in the project. There are several methods used for evaluating a project based on dollar or cash amounts and time periods. Three common methods are the net present value (NPV), the internal rate of return (IRR), and the payback period methods. Project risk or likelihood of success can be incorporated into the various benefit-cost analyses as well. Net Present Value (NPV) Method

Weston and Brigham (1974) and Johnson and Melcher (1982) give the following equations: n

NPV =

∑ (1 + r ) CFt

t

t =0

where n = the number of periods; t = the time period; r = the per period cost of capital for the organization (also denoted as i if annual interest rate is used); and CFt is the cash flow in time period t. Note that CF0, the cash flow in period zero, is also denoted as the initial investment. The cash flow for a given period, CFt is calculated as: CFt = CFB,t – CFC,t where CFB,t is the cash flow from project benefits in time period t and CFC,t is the project costs in the same time period. The standard convention for cash flow is positive (+) for inflows and negative (–) for outflows. The conversion from an annual percentage rate (APR) per year, equal to i, to a rate r for a shorter time period, with m periods per year, is: 1

R = (1 + i ) m − 1 If the project NPV is positive, for a given cost of capital, r, the project is normally approved. Internal Rate of Return (IRR) Method

The internal rate of return (IRR) is the interest or discount rate, i or r, that results in a zero net present value, NPV = 0, for the project. This is equivalent to stating that time weighted inflows equal time weighted outflows. The equation is n

NPV = 0 =

∑ (1 + r ) CFt

t

t =0

The IRR is that value of r that results in NPV being equal to 0 and is calculated by an iterative process. Once calculated for a project, the IRR is then compared with

SL3151Ch13Frame Page 612 Thursday, September 12, 2002 6:01 PM

612

Six Sigma and Beyond

that for other projects and investment opportunities for the organization. The projects with the highest IRR are approved, until the available investment capital is allocated. Most real projects would have an IRR in the range of 5 to 25% per year. Managers given the opportunity to accept a project that has calculated values for IRR higher than the company’s return on investment (ROI) will normally approve, assuming the capital is available. The above equations for net present value and internal rate of return have ignored the effects of taxes. Some organizations make investment decisions without considering taxes, while others look at the after-tax results. The equations for NPV and IRR can be used with taxes, if the cash flow effect of taxes is known. Payback Period Method

The payback period is the length of time necessary for the net cash benefits or inflows to equal the net costs or outflows. The payback method generally ignores the time value of money, although the calculations can be done taking this into account. The main advantage of the payback method is the simplicity of calculation. It is also useful for comparing projects on the basis of quick return on investment. A disadvantage is that cash benefits and costs beyond the payback period are not included in the calculations. Organizations using the payback period method will set a cut-off criterion, such as 1, 1½, or 2 years maximum for approval of projects. Uncertainty in future status and effects of projects or rapidly changing markets and technology tend to reduce the maximum payback period accepted for project approval. If the calculated payback period is less than the organization’s maximum payback period, then the project will be approved. (Quite often, in the six sigma/DFSS world, the payback is figured on a preset project savings rather than time. The most common figure floating around is a $250,000 per-project savings.) Project Decision Analysis In addition to the benefit-cost analysis for a project, the decision to proceed must also include an evaluation of the risks associated with the project. To manage project risks, first identify and assess all potential risk areas. Risk areas include: Business risks

Insurable risks

Technology changes Competitors Material shortages Health and safety issues Environmental issues

Property damage Indirect consequential loss Legal liability Personnel

After the risk areas are identified, each is assigned a probability of occurrence and the consequence of risk. The project risk factor is then the sum of the products of the probability of occurrence and the consequence of risk. Project Risk Factor =

∑ {(probability of occurrence) * (consequence of risk)}

SL3151Ch13Frame Page 613 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

613

Risk factors for several projects can be compared if alternative projects are being considered. Projects with lower risk factors are chosen in preference to projects with higher risk factors. A more extensive description of risk management is found in Kerzner (1995).

WHY PROJECT MANAGEMENT SUCCEEDS The single most important characteristic of project management is the consistent ability to get things done. It is a results- or goal-oriented approach, where other considerations are secondary, so the single-minded concentration of resources greatly enhances prospects for success. This also implies that the results, success or failure, are quite visible. Integrative and executive functions of the project manager provide another inherent advantage in the project management approach that improves the likelihood for success because of the single point of responsibility for those functions. Specific advantages of the single point integrative characteristic include: • Placing accountability on one person for the overall results of the project • Assurance that decisions are made on the basis of the overall good of the project, rather than the good of one or another contributing functional department • Coordination of all functional contributors to the project • Proper utilization of integrated planning and control methods and the information they produce Advantages of integrated planning and control of projects include: • Assurance that the activities of each functional area are being planned and carried out to meet the overall needs of the project • Assurance that the effects of favoring one project over another are known • Early identification of problems that may jeopardize successful project completion, to enable effective corrective action to prevent or resolve the problem Project management is a specialized management form. It is an effective management tool that is used because something is gained by departing from the normal functional way of doing things in terms of people, organizations, and methods. Conflict, confusion, and additional costs are often associated with significant changes of this nature. Poorly conceived or poorly executed project management can be worse than no project management at all. Project management should be used well or not at all. Executives should not permit a haphazard, misunderstood use of project management principles. Although simple in its concepts, project management can be complex in its application. Project management is not a cure-all intended for all projects. Before project management can succeed, the application must be correct. Executives should not use project management unless it appears to be the best solution. The use of project management techniques seems most appropriate when:

SL3151Ch13Frame Page 614 Thursday, September 12, 2002 6:01 PM

614

Six Sigma and Beyond

1. 2. 3. 4. 5.

A well-defined goal exists. The goal is significant to the organization. The undertaking is out of the ordinary. Plans are subject to change and require a degree of flexibility. The achievement of the goal requires the integration of two or more functional elements or independent organizations.

Even though project management may not be feasible, good principles have contributed to the success of thousands of small and medium-sized projects. Many managers of such projects have never heard of project management but have used the principles. A wider application of these principles will also help achieve success in smaller projects. Executives play a key role in the successful application of project management. A commitment from top management to ensure it is done right must be combined with the decision to use this approach. Top management must realize that establishing a project creates special problems for the people on the project, for the rest of the organization, and for top managers themselves. If executives decide to use this technique, they should expend the time, decision-making responsibility, and executive skills necessary to ensure that it is planned and executed properly. Before it can be executed properly, sincere and constructive support must be obtained from all functional managers. Directives or memos are not enough. It takes personal signals from top management to members of the team and functional managers to convey that the project will succeed and that team members will be rewarded by its success. In addition, necessary and desirable changes in personnel policies and procedures must be recognized and established at the onset of the project. The human aspect of project management is both one of its greatest strengths and one of its most serious drawbacks. In order for project management to succeed, it requires capable staff. Only good people can make a project successful. In the long run, this is true for any organization. Good people alone cannot guarantee project success; a poorly conceived, badly planned, or inadequately resourced project has little hope for success. Great emphasis is placed on the selection of good people. The project leader, more than any other single variable, seems to make the difference between success and failure. Large projects will require one person to be assigned the full-time role of project manager. If a number of projects exist but not enough project managers are available for full-time assignment to a project, assign several projects to one full-time project manager. This approach has the advantage that the individual is continually acting in the same role, that of a project manager, and is not distracted or encumbered by functional responsibilities. To conclude, project management is an effective management tool used by business, industry, and government, but it must be used skillfully and carefully. In review, the following major items are necessary for successful results from project management in the field of quality: • Executives provide wholehearted support and commitment when the decision is made to use this approach. • Project management is the best solution or right application for implementing any quality program.

SL3151Ch13Frame Page 615 Thursday, September 12, 2002 6:01 PM

Project Management (PM)

615

• Emphasis is placed on selecting the best people for staff, especially the project leader. • Good principles of project planning and control are applied. Effective use of project management will reduce costs and improve efficiency. However, the main reason for the widespread growth of project management is its ability to complete a job on schedule and in accordance with original plans and budget.

REFERENCES American Heritage Dictionary of the English Language, 3rd ed., Houghton Mifflin, Boston, 1992. Duncan, W.R., A Guide to the Project Management Body of Knowledge, Project Management Institute, Upper Darby, PA, 1994. Harry, M., The Vision of Six Sigma: A Roadmap for Breakthrough, 5th ed., Vol. II, Tri Star Publishing, Phoenix, AZ, 1997. Johnson, R.B. and Melicher, R.W., Financial Management, 5th ed., Allyn and Bacon, Inc., Boston, 1982. Kerzner, H., Project Management: A Systems Approach to Planning, Scheduling and Controlling, 5th ed., Van Nostrand Reinhold, New York, 1995. Stamatis, D.H., Total Quality Service, St. Lucie Press, Delray Beach, FL, 1996. Turner, J. R., The Handbook of Project-Based Management, McGraw-Hill, New York, 1992. Weston, J. F. and E.F. Brigham, Essentials of Managerial Finance, 3rd ed., Dryden Press, Hinsdale, IL, 1974.

SELECTED BIBLIOGRAPHY Frame, J.D., The New Project Management, Jossey-Bass, San Francisco, 1994. Geddes, M., Hastings, C., and Briner, W., Project Leadership, Gower, Brookfield, VT, 1993. Lock, D., Gower Handbook of Project Management, Gower, Brookfield, VT, 1994. Michael, N. and Burton, C., Basic Project Management, Singapore Institute of Management, Singapore, 1993. Stamatis, D.H., TQM Engineering Handbook, Marcel Dekker, New York, 1997. Stamatis, D.H., Total Quality Management and Project Management. Project Management Journal, Sept. 1994, pp. 48–54.

SL3151Ch13Frame Page 616 Thursday, September 12, 2002 6:01 PM

SL3151Ch14Frame Page 617 Thursday, September 12, 2002 5:59 PM

14

Limited Mathematical Background for Design for Six Sigma (DFSS)

EXPONENTIAL DISTRIBUTION AND RELIABILITY EXPONENTIAL DISTRIBUTION f(t)

δ

δ e - δt 0.37 δ

0

t

µt = 1/ δ

F(t) 1

1 - e - δt 0.63

0

µt = 1/ δ

t

617

SL3151Ch14Frame Page 618 Thursday, September 12, 2002 5:59 PM

618

Six Sigma and Beyond

Probability Density Function and Cumulative Distribution Function Probability Density Function (Decay Time) δe − δ t  f (t ) = E x (t; δ ) =  0

; t 〉0,δ 〉0 ; elsewhere

Cumulative Distribution Function (Rise Time) t

∫

F (t ) ≡ δ e − δ t dt = 1 − e − δ t 0

1 some use η = 1  δ δ

Mean Time:

µt =

Variance:

 1 σ =   δ

One parameter:

2

2 t

δ

Reliability Problems Exponential distribution is used in reliability problems. Exponential distribution can describe the probability of a failure prior to some specified time t assuming that failure occurs at a constant rate δ ( = λ) over time. Reliability, the chance of no failure in time t, is expressed as R(t ) = e −δ t Failure is a complement of cumulative probability of reliability: F (t ) ≡ 1 − R(t ) = 1 − e − δ t F(t) is used to compute the probability of failure prior to t. The derivative of the cumulative distribution function is the probability density function (pdf). f (t ) =

dF (t ) = δ e−δ t dt

SL3151Ch14Frame Page 619 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

Mean Time to Failure (MTTF) = TMF = 1

619

δ

Failure Rate: δ = 1 TMF Fail  (Bad)

AT

Pass (Good)

0 t

Time, T

0

t+∆t

CONSTANT RATE FAILURE Exponential function: A e−δ t Evaluate at any time t, the time rate of decrease in amplitude is constant: dAe− δ t = −δ A e − δ t dt If we consider equal time increments ∆t, then exponential has consistent amplitude ratio between increments − δ t +∆ Ae ( )

Ae ( ) −δ t

=

− δ t + n∆t ) Ae (

Ae

( ( ) )

− δ t + n−1 ∆t

= e −δ ∆ t

Example Data from 100 pumps demonstrated an average life of 5.75 years and that failures followed an exponential distribution. 1. 2. 3. 4. 5.

Determine the probability of failure during the first year. Determine the probability of failure during the first 3 months. Determine the probability of failure prior to the average life. Determine the probability of reliably operating for at least 10 years. Plot the reliability curve and compare with the pdf curve.

SL3151Ch14Frame Page 620 Thursday, September 12, 2002 5:59 PM

620

Six Sigma and Beyond

Solutions: Given: MTTF = TMF = 1/δ = 5.75 years Compute Failure Rate: δ = 1/TMF = 0.174 per year Exponential pdf: f (t ) = δ e − δ t = 0.174 e − 0.174t Failure cdf: F (t ) = 1 − R(t ) = 1 − e − δ t 1. Probability of failure during the first year: 16%

( ) (

F t −1 = 1 − e − 0.174t

)

t =1

= 1 − e

( )  = 1 − 0.84 = 0.16 

− 0.174 1

2. Probability of failure during the first 3 months or ¼ year:

(

)

(

)

− 0.174 ( 0.25)  F t = 0.25 = 1 − e  = 1 − 0.957 = 0.043

3. Probability of failure prior to the average life; MTTF = TMF = 5.75

(

(

)

)

− 0.174 ( 5.75)  − 1.0 F t = 5.75 = 1 − e = 1 − 0.368 = 0.632  = 1−e

4. Probability of reliably operating for at least 10 years.

(

)

R t = 10 = e

( ) = 0.176

− 0.174 10

5. Plot the reliability curve and compare with the pdf curve.

f(t) ; R(t) f(t) = δe-δτ R(t) = e-δτ

1.0

R(t) 0.50 0.42

δ = 0.174

f(t)

0

5

10

15

t , years

SL3151Ch14Frame Page 621 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

PROBABILITY

OF

621

RELIABILITY

The exponential distribution as the basis of the reliability function is based on the probability of samples of an event that describes a general physical situation; i.e., time to a (bad) occurrence.

CONTROL CHARTS Continuous Time Waveform Fail  (Bad)

AT

Pass (Good)

0 t

Time, T

0

t+∆t

Discrete Time Samples

tk = k∆t ∆t

AT

0

Fail  (Bad)

Pass (Good)

t1t2 0

tk t

tn t+∆t

Time, T

SL3151Ch14Frame Page 622 Thursday, September 12, 2002 5:59 PM

622

Six Sigma and Beyond

Digital Signal Processing

tk = k∆t ∆t

AT

0

Fail  (Bad)

Pass (Good)

t1t2

tn

tk

0

Time, T

t+∆t

t

Uniform sampling time or time increment: ts = ∆t Time increment ∆t small so as to include only one sample. Total sampling time: t n = nts = n∆t Total number of samples taken: n = t n ts Current time sample: t k = kts = k∆ t Designate RV event as occurrence of bad sample X. Probability of a bad sample only in the n-th interval.

SAMPLE SPACE

Good

Sample, n n=g+b

G g

b B

Bad

Sample Space: n = g + b

Population, N N=G+B

SL3151Ch14Frame Page 623 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

623

Bad Sample

Good Samples

1

2

k

n

Sample “Bad” when outcome exceeds a limit AT Sample “Good” when outcome less than limit AT Random variable for this experiment is X. “Good” or “Bad” are the only two possible states for X. Assign X = 0 for “Good” X = 1 for “Bad”

tk = k∆t ∆t

AT

0

Fail  (Bad)

Pass (Good)

t1t2

tn

tk

0

Time, T

t+∆t

t Sample Space: n = g + b

Set Bad Samples: {b}

Set Good Samples: {g}

● 1

● 2

●

●

●

●

● k

●

●

●

●

●

 n

One bad sample {b} is assumed to occur exactly on the n-sample. {b} = {Xn = 1} This single bad sample is preceded by a sequence of (n – 1) good samples {g}. {g} = {Xn – 1 = 0, Xn – 2 = 0, …, X1 = 0}

SL3151Ch14Frame Page 624 Thursday, September 12, 2002 5:59 PM

624

Six Sigma and Beyond

If each sample is independent of the proceeding sample then

({ } { })

{ }{g} P({g}) = P({b}) P({g})

P b • g = P b

ASSIGNING PROBABILITY

TO

SETS

Assume only one sample can be measured in any interval ∆t.

[

]

• The probability that one bad sample occurs in interval t ≤ T ≤ t + ∆ t is assumed to be constant p = δ ∆t

({ }) (

)

P b ≡ P X = 1; t ≤ T ≤ t + ∆ t ≡ δ ∆ t = p • Conversely, the probability of one good sample in this interval is

(

)

(

)

P X = 0 ; t ≤ T ≤ t + ∆ t ≡ 1 − P X = 1; t ≤ T ≤ t + ∆ t = 1 − δ ∆ t = q • The probability of (n – 1) good samples in the range [0 ≤ T ≤ t] is

({ }) (

)

P g ≡ P X = 0 ; 0 ≤ T ≤ t ≡ R(t ) = Reliability • Probability of total set:

({ }) ({ }) (

) (

P g P b ≡ P X = 0 ; t ≤ T ≤ t P X = 1; t ≤ T ≤ t + ∆ t

)

Note: There are two types of probabilities or variables, one when X = 0 for set {g} and one when X = 1 for set {b}. To establish an “equation,” we need to deal with only one variable. Assume the sample of the increment t ≤ T ≤ t + ∆ t where also “good” then we could write directly:

[

(

]

) ( ) ( = P( X = 0 ; 0 ≤ T ≤ t ) [1 − δ ∆ t ]

P X = 0; 0 ≤ T ≤ t + ∆ t ≡ P X = 0; 0 ≤ T ≤ t P X = 0; 0 ≤ T ≤ t + ∆ t

)

Differential equation form (take limit as ∆t → dt):

(

) (

) (

P X = 0; 0 ≤ T ≤ t + d t − P X = 0; 0 ≤ T ≤ t = P X = 0; 0 ≤ T ≤ t

) [ − δ d t]

SL3151Ch14Frame Page 625 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

625

Dividing by dt puts the LHS into the form of a derivative:

(

) (

P X = 0; 0 ≤ T ≤ t + d t − P X = 0; 0 ≤ T ≤ t

)

dt

=

[ − δ] P ( X = 0 ; 0 ≤ T ≤ t )

First order differential equation (homogeneous):

(

P X = 0; 0 ≤ T ≤ t + d t dt

) = − δ P( X = 0 ; 0 ≤ T ≤ t )

[ ]

which can be conveniently expressed in terms of reliability:

( ) = δ R(t )

dRt dt

()

Solution: R t = e

−δ t

+ C1

() R(t ) = e

Initial condition at t = 0: R 0 ≡ 1 = 1 + C1 and C1 = 0 Hence, the reliability is:

−δ t

1

GAMMA DISTRIBUTION The probability that the nth event (e.g., failure) will occur exactly at the (end) time t, when the events are assumed to occur at a constant rate δ. The idea of a constant event rate δ is the same assumption used for both exponential and Poisson distributions. The variable time, t, is said to have a gamma distribution.

GAMMA DISTRIBUTION (PDF)  δ n t n −1   n e − δ t; t 〉 0 f t = G t; n, δ ≡    0; elsewhere

() (

Two parameters: Scale parameter: Shape parameter:

()

)

n (changes scale not shape) δ (changes shape not scale) ∞

Gamma Function

( ) ∫x

n =

0

n −1

e − x dx

SL3151Ch14Frame Page 626 Thursday, September 12, 2002 5:59 PM

626

Six Sigma and Beyond

Mean: µ =

n δ

Variance: δ 2 =

n δ2

n −1 δ If n ≤ 1 non-modal shape with mode at: m = 0 If 1 < n unimodal shape with mode at: m =

GAMMA FUNCTION ∞

( ) ∫x

n =

n −1

e − x dx

0

Properties of Gamma Functions If n = positive integer,

() (

)

 n = n −1 !

() ( )( (1) = (0)! = 1

)

 n = n −1  n −1

 1   = Π  2 Degrees of freedom (e.g., see chi-square distribution) If n = ν = degrees of freedom (always a positive integer) If n = even integer:  n  n    =  − 1 !  2  2  If n = odd integer:   n  n   n   =  − 1  − 2 …   2  2   2

 5  3 Π  2  2 2   

SL3151Ch14Frame Page 627 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

627

Gamma Distribution and Reliability

tk = k∆t ∆t

AT

0

Fail  (Bad)

Pass (Good)

t1t2

tn

tk

0

Time, T

t+∆t

t

Gamma distribution is used in reliability where a number of partial failures n must occur before a system or item completely fails. The time to the n-th failure is estimated assuming that the times to individual (partial) failures are exponentially distributed. Two parameters have the following interpretation: n is the number of partial failures per complete failure δ is the failure rate Limiting case: When total system failure occurs at the time of the first partial failure n = 1 and the gamma distribution reduces to the exponential distribution. δe − δ t  f (t ) =  0

EXAMPLE 1: TIME

TO

; 0≤t ;t 〈0

TOTAL SYSTEM FAILURE

To ensure reliability, an important computer system is controlled by a set of four switches. Each switch has a constant failure rate of two per year. The computer system is said to totally fail when there has occurred a total of three switch failures.

Time to total failure, t Two partial failures

dt Final failure n

SL3151Ch14Frame Page 628 Thursday, September 12, 2002 5:59 PM

628

Six Sigma and Beyond

State the parameters and plot the pdf with its mean and mode for the time to total system failure. Solution: The two parameters are: δ and n Failure rate: δ = 2 failure/year Number of partial failures to total system failure: n = 3 Mean time to system failure:

TMT = µ T =

[

]

[

3 failures n = = 1.5 years δ 2 failures per year

[

]

]

Mode:

[

]

[

2 failures n −1 = = 1.0 years δ 2 failures per year

[

]

]

Gamma Distribution and Reliability 1. Probability density function  δ nt n − 1 − δ t e ;t〉 0  f (t ) = G(t; n; δ ) =  n  ; elsewhere 0 For case of: n = 3 [failures] and δ = 2 [failure/year] Gamma function: (3) = (3 – 1)! = 2! = 2 [f] 1 [f] = 2 [fail]2  δ nt n − 1 − δ t 2 3 t 2 − 2 t f (t ) =  e = e = 4 t 2 e −2t 2  n Time (years) pdf

t = 0:

f (0)

=

0.0000

t = 0.5

f (.05)

=

0.3679

t = 1.0

f (1.0)

=

0.5416

SL3151Ch14Frame Page 629 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS) t = 1.5

f (1.5)

=

0.4481

t = 2.0

f (2.0)

=

0.2931

t = 2.5

f (2.5)

=

0.1685

t = 3.0

f (3.0)

=

0.0892

t = 4.0

f (4.0)

=

0.0215

629

0.6 0.5 n= 3 δ=2

f(t)

0.4 0.3 0.2 0.1 0 0

1

2

3

4

t

2. The case when the failure rate is reduced to 1 per year and the total system failure occurs after only two switches fail. Failure rate: δ = 1 failure/year Number of partial failures to total system failure: n = 2 Mean time to system failure: µ=

[

]

[

]

[

]

[

]

2 failures n = = 2.0 years δ 1 failures per year

[

]

Mode: 1 failures n −1 = = 1.0 years δ 1 failures per year

[

]

SL3151Ch14Frame Page 630 Thursday, September 12, 2002 5:59 PM

630

Six Sigma and Beyond

Gamma function: (2) = (2 – 1)! = 1! = 1 [fail]  δ nt n − 1 − δ t 1 2 t 1 − 1t f (t ) =  e = e = te − t  n 1  Time (years) pdf t = 0:

f (0)

=

0.0000

t = 0.5

f (.05)

=

0.3033

t = 1.0

f (1.0)

=

0.3679

t = 1.5

f (1.5)

=

0.3347

t = 2.0

f (2.0)

=

0.2707

t = 2.5

f (2.5)

=

0.2052

t = 3.0

f (3.0)

=

0.1494

t = 4.0

f (4.0)

=

0.0733

0.6 n=2 δ=1

0.5

f(t)

0.4 0.3 0.2 0.1 0 0

1

2

3

4

t

3. For special case of n = 1; Gamma → Exponential Distribution δ e − δ t  f (t ) = E x(t; δ ) ≡  0 

;t≤ 0 ;t 〈 0

SL3151Ch14Frame Page 631 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

Consider case of: n = 1 [failures] and δ = 2 [failure/year] Gamma function: (1) = (1 – 1)! = 0! = 1 [unitless]

()

f t = δ e − δt = 2e − 2 t Time (years) pdf t = 0.2:

f (0)

=

1.3406

t = 0.5

f (.05)

=

0.7358

t = 0.6

f (0.6)

=

0.6024

t = 1.0

f (1.0)

=

0.2707

t = 1.5

f (1.5)

=

0.0996

t = 2.0

f (2.0)

=

0.0366

t = 2.5

f (2.5)

=

0.0135

t = 3.0

f (3.0)

=

0.0050

t = 4.0

f (4.0)

=

0.0007

f(t) 0.7 – 0.6 –

● ● n=1 δ=2

0.5 – 0.4 – 0.3 –

●

0.2 –

●

0.1 – 0

● 0

1

● 2

● 3

4 t

631

SL3151Ch14Frame Page 632 Thursday, September 12, 2002 5:59 PM

632

Six Sigma and Beyond

Reliability Relationships

tk = k∆t ∆t



AT

 

Fail (Bad)

Pass (Good)

0

t 1t 2

tn

tk

0

Time, T

t+∆t

t

Sample data approach: n – identical systems are placed in operation at t = 0 g(t) – good; number operating or survive to time t b(t) – bad; number to fail or survive to time t where b(t) = n – g(t) Random variable approach: dt Final failure

Time to total failure, t Two partial failures

n

T is a continuous random variable representing the time to failure or failure time of a system. Reliability Function Sample Space: n = g(t) + b(t) H Set bad samples: {b(t) = x} • Set good samples: {g(t) = n – x}

1

2

...

k

...

n ∆t

0

t

SL3151Ch14Frame Page 633 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

633

Reliability is the probability that a system can operate successfully over the time interval from 0 to t. Reliability can also be viewed as the probability that the system will survive beyond a given point in time, t. gt ) (n ) = n −nb(t ) = 1 − b(nt )

() (

R t =P T >t =

As time increases, chance of failure increases and reliability decreases; as t → ∞ or n → ∞, then R(t) → 0.

DATA FAILURE DISTRIBUTION Probability of systems failure as a function of time.

() (

)

Q t = P T > t = 1 − R(t ) =

FAILURE RATE

OR

b(t ) n

DENSITY FUNCTION

Failure rate is ratio of the number of failures occurring in time interval ∆t to the size of original population, divided by time interval.

f (t ) =

(

)

b t + ∆ t − b(t ) n ∆t

[n − g(t + ∆ t )] − [n − g(t )]

= lim

n ∆t

∆t → 0

= lim

() (

g t − g t + ∆t n ∆t

∆t → 0

)

()

1 d gt n dt

=−

A measure of the overall speed at which failures are occurring Alternatively, can express the failure rate as the probability that a failure occurs in any given time interval ∆t by taking the time derivative (rate) of the failure distribution Q(t).

()

() Q(t + ∆t ) − Q(t ) = lim

f t =

d Qt dt

∆t →0

= lim

∆t →0

(

∆t

P T < t ≤ t + ∆t ∆t

)

SL3151Ch14Frame Page 634 Thursday, September 12, 2002 5:59 PM

634

Six Sigma and Beyond

HAZARD RATE FUNCTION Hazard rate is ratio of the number of failures occurring in time interval ∆t to the number of survivors at the beginning of the time interval t, divided by time interval.

(

)

b t + ∆ t − b(t )

h(t ) = lim

() [n − g(t + ∆ t )] − [n − g(t )] = lim g(t ) ∆ t g(t ) − g(t + ∆ t ) 1 d = lim =− g(t ) g(t ) ∆ t g(t ) dt nf (t ) f (t ) f (t ) = = = g t ∆t

∆t → 0

∆t → 0

∆t → 0

g(t )

g(t ) / n

R(t )

Hazard rate as the conditional probability of failure in the interval (t, t + ∆t] given that the system has survived up to the time t:

h(t ) = lim

(

P T < t ≤ t + ∆t T > t ∆t

∆t → 0

= lim

(

P T < t ≤ t + ∆t

(

PT 0 

yields a Poisson distribution probability density function (a = δt):

(δ t ) e R( x ; t ) ≡ P( X = x ; 0 ≤ T ≤ t ) = x

−δt

x!

; x = 0, 1, 2, K

Example An average of four (4) private aircraft landings per hour in a small local airport, δ = 4L/h. What is probability that six (6) aircraft will land in one hour? The Poisson distribution is

( )

PO x ; a = One parameter: Mean: µ = a Solution:

(a ) x e − a x!

a ( = np = δ t)

; x = 0, 1, 2, K

Variance: δ2 = a

Number of aircraft landing in one hour: X = 6L Average number of landings in one hour: a = 4L That is, a = δ t = 4 L/h • 1h = 4L

SL3151Ch14Frame Page 640 Thursday, September 12, 2002 5:59 PM

640

Six Sigma and Beyond

The probability of six aircraft in one hour is then given by

( )

PO 6 ; 4 =

( 4)6 e −4 = 0.1042 6!

There is a 10.4% chance that six aircraft will land in one hour.

WEIBULL DISTRIBUTION One of the more popular models for time-to-failure (TTF), Weibull distributions take many shapes and are typically identified as in the following illustration.

f(t) Weibull Distribution δ=1

a=4

a=1 a=2 a = 0.5 0

1

2

3

t

Weibull probability density function (pdf)  a δ δ t f (t ) = W t ; a, δ ≡  0 

(

)

( )

e ( ) ;0≤t

a−1 − δ t a

;t < 0

Cumulative distribution t

∫ ( )

F (t ) ≡ a δ δ t 0

Two parameters:

− δt e ( ) dt = 1 − e ( )

a−1 − δ t a

a

4

SL3151Ch14Frame Page 641 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

641

Shape parameter: a (changes shape not scale) Scale parameter: δ (changes scale not shape) Some authors define δ = 1/η and a = β In a typical Weibull distribution shown below, there are some general characteristics f(t) Weibull Distribution δ=1

a=4

a=1 a=2 a = 0.5 0

• Mean: µ =

1 δ

1

2

3

t

4

 1  1 +   a

• Variance: σ 2 =

 1 1   2 2  1 + −     1 + a   δ 2   a    

• 1/δ also referred to as “characteristic life” or “time constant,” the life or time at which 63.2% of population has failed. • If a = 1, the Weibull reduces to the exponential distribution. • If a = 2, the Weibull reduces to the Rayliegh distribution. • If a ≈ 3.5, the Weibull approximates the normal distribution. • For a < 1, reliability function decays less rapidly. • For a > 1, reliability function decays more rapidly. • A useful model for the failure time (or length of life) distributions of produces and processes. • Does not assume that the failure rate, δ, is a constant as do the Exponential and Gamma distributions. • Has the advantage that the distribution parameters can be adjusted to fit many situations; because of this adaptability it is widely used in reliability engineering. • The cumulative distribution has closed form expression that can be used to compute areas under the Weibull curve.

SL3151Ch14Frame Page 642 Thursday, September 12, 2002 5:59 PM

642

Six Sigma and Beyond

• Estimates of the two parameters, δ and a, can be obtained when ranked sample data are plotted on scale adjusted cumulative percentile (See Probability Plots). Note: Characteristic life t = 1/δ corresponds to the 63.2%

xxxx

99 Mileage (miles)

95 90 70 Occurrence (CDF)

60 50 40

°

30 20

°1

°

80

° °

° 10 Eta Value Beta Value r2 Value

5

2

xxxx

xxxx

xxxx

n/s where n = samples and s = suspended samples xx/yy

xxxx xxxx xxxx

1 10000

100000 Mileage

x

1000000

(miles)

• Weibull reliability or survival function:

(

∞

) ∫ f (t() dt(

R(t ) ≡ P T > t =

t

∞

( = aδ δt

∫ ( )

( ( e ( ) dt ; let u = δt

a−1 − δ t( a

( )

a

t

∞

=

∫

u (t )

e− u d u = −e− u

∞

= e − (δ t )

a

u(t )

• Weibull failure distribution: (same as cumulative distribution)

SL3151Ch14Frame Page 643 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

t

Q(t ) ≡ P(T ≤ t ) =

(

643

(

∫ f (t ) dt 0

= 1 − R(t ) = 1 − e − (δ t )

a

• Weibull hazard rate function:

( )

a−1

( )

f (t ) a δ δ t e − δ t h(t ) ≡ = a R(t ) e − δt

( )

a

( )

= aδ δt

a−1

• The shape parameter a, can be used to adjust the shape of the Weibull distribution to allow it to model a great many life (time) related distributions found in engineering.

f(t) Weibull Distribution δ=1

a=4

a=1 a=2 a = 0.5 0

1

2

3

t

4

THREE-PARAMETER WEIBULL DISTRIBUTION If failures do not have the possibility of starting at t = 0, but only after a finite time tO, a time-shift variable can be used to redefine the Weibull reliability function: R(t ) = e

( ( ))

− δ t −tO

a

where the time tO is called the failure free time or minimum life.

SL3151Ch14Frame Page 644 Thursday, September 12, 2002 5:59 PM

644

Six Sigma and Beyond

TAYLOR SERIES EXPANSION Determines the value of a function f(x) at any x from the value of the function and all its derivatives at a given location xo (provided no discontinuities occur). ∞

∑

f ( x) =

n =0

( x − x0 ) (n) f ( x0 )

n

n!

( )

= f x0 +

()

df x

(x − x ) + 0

dx

( ) ( x − x ) + L + d f ( x) ( x − x )

d2 f x

n

0

dx 2

dx n

2!

x0

n!

x0

() ( )

( )(

n

0

x0

)

( )

f x = f x0 + slope x0 ⋅ x − x0 + curvature x0

(x − x ) ⋅

2

0

2

+L

( x − x0 ) + L (1) (2 ) f ( x ) = f ( x0 ) + f ( x0 ) ⋅ ( x − x0 ) + f ( x0 ) ⋅ 2

2

f(x) f(x) f(x) f( 2) (xo) [x-xo ]2 /2 f( 1) (xo ) [x-xo]

f(x) f(xo )

O

f(xo )

(x - xo ) xo

O

x

x

f(xo ) xo [x -xo ] x

Taylor series expansion — evolves into a power series 1. Series about x0 ∞

f ( x) =

∑ n =0

( x − x0 ) (n) f ( x0 )

n

n!

(2 ) f ( x0 ) 2 1) ( = f ( x0 ) + f ( x0 )( x − x0 ) + ( x − x0 ) + L 2!

+L+

f

()x ( 0) n

n!

(x − x ) 0

n

+L

x

SL3151Ch14Frame Page 645 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

645

2. Series about origin x0 = 0 ∞

f ( x) =

∑ n =0

( n ) ( x) f ( 0)

n

n!

(2 ) (n) f ( 0) 2 f ( 0) n 1) ( = f ( 0) + f ( 0)( x ) + x +L+ x +L 2!

n!

= a0 + a1 x + a2 x + L a n x + L n

Observations: 1. An arbitrary function f(x) can be expressed as a power series: an =

()

n f( ) 0

n! 2. Coefficients of power series are related to the derivative of the function evaluated at origin. 3. A linear function consists of only the first two terms: f x = a0 + a1x

()

TAYLOR SERIES EXPANSION To establish linear relationship about ambient state: • • • •

Stress Pressure Voltage Input

↔ ↔ ↔ ↔

Strain constitutive relation in elasticity Density equation of state Current about quiescent point Output

Linear implies: “input” disturbance (x – x0) small enough that “output”

() ( )

f x ≈ f x0 +

( ) ( x − x ) = f ( x ) + m( x − x ) dx

df x

0

0

0

x0

Linear response about xo

f(x) f(b) f(xo )

output input

f(a)

O

x a

xo

b

SL3151Ch14Frame Page 646 Thursday, September 12, 2002 5:59 PM

646

Six Sigma and Beyond

Output is a linear function of input: x Input

f(x)

Linear System

Output

f(x ) = mx + c

INPUT

OUTPUT 1 — [ f ( x) - f (xo ) ] m

( x - xo )

Recombine: changes in independent variable(input). Provide linear changes in the dependent variable(output). Slope m serves to adjust units and is called “sensitivity.”

Linear response about xo

f(x) f(b) f(xo )

output input

f(a)

O

a

x

xo

b

Exponential function eax — Taylor series about x = x0 in interval –∞ < x < ∞

e =e ax

ax0

+ ae

ax0

(x − x )

2

(x − x ) + a e

2 ax0

0

0

+L + a e

2 ax0

2!

(x − x )

n

0

n!

Factoring out the common exponential term:  a2 e ax = e ax0 1 + a x − x0 + x − x0 2 

(

)

(

)

2

+L +

(

an x − x0 n!

MacLaurin series about x0 = 0 in interval –∞ < x < ∞

)

n

 + L 

+L

SL3151Ch14Frame Page 647 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

(ax ) ± (ax ) = 1 ± ax + 2

e

± ax

3

2!

3!

(±ax ) +L +

n

∞

+L =

n!

∑

647

(±ax )

n

n!

n =0

Normal density-like function: ebx 2 — Taylor series about x = 0

e

bx 2

(2 b) x =1+ 0 +

2

2

= 1 + bx 2 +

+0+

( )

1 bx 2 2

2

( )

3 2b

2

24

x +0+

( )

+

2 bx 2 15

(

)

4

3

2Π

e− z

2

/2

=

1 2Π

Exact

[1 − z vs.

2

3

6!

x6 + L

+L 1

Standard Normal Distribution: N z ; 0 , 1 ≡ 1

( )

12 2 b

2Π

e− z

2

/2

]

/ 2 + z 4 / 8 − z 6 / 60 + L

Two

Three

Four Terms

z = ±0.5

0.3521

0.3490

0.3522

0.3519

z =±0.675 (Q1,3)

0.3177

0.3080

0.3184

0.3178

z = ±1.0

0.2420

0.1995

0.2494

-0.2427

99.74%

N(z; 0,1)

95.46 68.26 0.40 — 0.24

σ=1

0.05 0.004

2.5

-3

13.5 34.0 34.0 13.5

-2

-1

0

µ=0

1

2.5

2

Derivatives of exponential ebx2 about origin x = 0

3

Z

SL3151Ch14Frame Page 648 Thursday, September 12, 2002 5:59 PM

648

Six Sigma and Beyond

Zero: e bx

2

x =0

=1

First: 2

2 de bx de u de u du = = = 2bx e bx dx dx du dx

Second: d 2 e bx

2

=

dx 2

x =0

d dx

 d bx 2  d  bx 2   e =  2 bx e  dx dx  

( )

( )

2 2 2  = 2 b e bx + 2 bx e bx  = 2 b   x =0

Third: d 3 e bx dx 3

2

= x =0

2 2 2 d  d 2 bx 2  d  2 b e bx + 2 bx e bx   2 e =    dx  dx dx 

( )

( )

( )( )

( )

3 2 2 2  = 2 b 2 bx e bx + 2 2 bx 2 b e bx + 2 bx e bx  = 0   x =0

Fourth: d 4 e bx dx

2

=

4 x =0

d dx

( )

( )( )

( )

2 3  bx 2 bx 2 bx 2  2 b 2 bx e + 2 2 bx 2 b e + 2 bx e 

( )

( )

( )

2 2 2 2 2 2  =  2 b e bx + 2 b 2 bx e bx + 2 2 b e bx 

( )( )

2

( ) (2 b)e

2

+ 2 2 b 2 bx e bx + 3 2 bx Fifth: d 5e bx dx 5 Sixth:

2

=0 x =0

2

bx 2

( )

4 2 + 2 bx e bx   x=0

SL3151Ch14Frame Page 649 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

d 6 e bx dx 6

2

649

( )

= 12 2b

3

x =0

Sine function sin x — Taylor series about x = x0 in interval –∞ < x < ∞

(

(x − x )

(x − x )

2

)

0

sin x = sin x0 + cos x0 x − x0 − sin x0

3

0

− cos x0

2!

3!

MacLaurin series about x0 = 0 in interval –∞ < x < ∞

sin x = 0 + x − 0 −

x3 +L = 3!

∞

2 n +1

x ∑ (−1) (2n + 1)! n

n =0

Partial Derivatives Dependent variable has two or more independent variables f(x, y) Differentiate wrt to only one independent variable while holding the other variable constant e.g.,y = yo

( ) ≡ ∂ f ( x , y)

∂f x,y ∂x



∂x

f(x ,y)

=

(

∂ f x , yo

y = yo

∂x

) = f (x , y ) x

o

Line: f(x , y o)

Surface: f(x,y)

F(x o, y o)  Line: f(x o ,y)

Slope: fx (x o , y o)

●●

Height: f(x o,y o) xo (x o, y o)

(x o, y o)

Slope: fy (x o , y o)

● yo

Height: f(x ,y )

x

(x , y)

Taylor Series in Two-Dimensions Taylor series of f(x, y) about point (xo, yo):

y

(x, y)

SL3151Ch14Frame Page 650 Thursday, September 12, 2002 5:59 PM

650

Six Sigma and Beyond

( ) (

) (

f x , y = f xo , yo + x − xo

∂f x,y ) (∂ x )

(

o

+ y − yo x = xo

∂f x ,y ) (∂ x ) o

+L y = yo

Linear terms:

( ) (

) (

) (

) (

) (

f x , y = f xo , yo + x − xo f x x − xo + y − yo fy xo , yo

)

Taylor Series of Random Variable (RV) Functions Arbitrary function of two random variables X1 and X2 Y (X1, X2)

[(

)] (

Mean: µ Y ≡ E Y X1 , X2 = Y µ X 1 , µ X 2

)

Variance and Covariance Consider only linear terms of the Taylor series expansion about the mean of each random variable, µ Y = Y µ X 1 , µ X 2

(

(

)

) (

) (

Y X1 , X2 = Y µ X 1 , µ X 2 + X1 − µ X 1

) ∂∂XY

(

+ X2 − µ X 2

1 µ ,µ X1 X2

) ∂∂XY

2 µ ,µ X1 X2

Variation of function about its mean:

(

Y − µY = X1 − µ X 1

∂Y µ ) ∂(X ) + ( X X

2

− µX2

1

∂Y µ ) ∂ (X 2 ) X

Variance and covariance:

(

 σ 2Y = E  Y − µ Y 

)  2

  ∂Y = E  X1 − µ X 1  ∂ X1  

(

)

( ) 

 ∂Y µ X =σ   ∂X 1  2 X1

2

(

+ X2 − µ X 2 µX

( ) 

 ∂Y µ X + σ 2X 2    ∂X 2  

2

)

  ∂ X2 µ  X ∂Y

2

    

( )   ∂ Y (µ ) 

 ∂Y µ X + 2σ 2X 1 X 2    ∂X 1  

X

 ∂X  2  

SL3151Ch14Frame Page 651 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

Note: If X1 and X2 are independent RV, covariance σ X 1X 2 = 0 . Functions of Random Variables Sum or difference: Y = a1 X1 ± a2 X2 Mean:

µ Y = a1µ X 1 ± a2 µ X 2

Variance and covariance: 2

2

 ∂Y   ∂Y   ∂Y   ∂Y  σ =σ  + σ2X 2  + 2σ2X 1X 2       ∂ X1   ∂ X2   ∂ X1   ∂ X2  2 Y

2 X1

= a12 σ2X 1 + a22 σ2X 2 ± 2a1a2 σ2X 1X 2

Again, if X1 and X2 are independent RVs then the covariance is zero. Product: Y = ao X1 X2

( ) = (a X ) ∂X

∂Y µ X

o

2

1

µX

( ) = (a X )

∂Y µ X

= aoµ X 2 ;

∂ X2

o

1

µX

= aoµ X 1

Mean: µ Y = a oµ X 1µ X 2 Variance and covariance: 2

2

 ∂Y   ∂Y   ∂Y   ∂Y  σ 2Y = σ 2X 1  + σ 2X 2  + 2σ 2X 1 X 2       ∂ X1   ∂ X2   ∂ X1   ∂ X2 

(

)

2

(

)

2

(

)

= a oµ X 2 σ 2X 1 + a oµ X 1 σ 2X 2 ± 2 a 2o µ X 1µ X 2 σ 2X 1 X 2 Division of Random Variables

Y=

( ) = a

∂Y µ X ∂ X1

Mean:

µY =

ao X1 X2

( )

 −a X   ao ∂ Y µ X −a µ o 1 o = o2 X 1  X  = µ ; ∂X = 2  µX2  2  µX  X2  µX X2 2

a oµ X 1

µ X2 Variance and covariance:

651

SL3151Ch14Frame Page 652 Thursday, September 12, 2002 5:59 PM

652

Six Sigma and Beyond

2

2

 ∂Y   ∂Y   ∂Y   ∂Y  σY2 = σ2X 1 + σ2X 2  + 2σ2X 1X 2       ∂ X1   ∂ X2   ∂ X1   ∂ X2  2

2

 −a µ   a   a   −a µ  = σ2X 1 o  + σ2X 2  o2 X 1  + 2σ2X 1X 2  o   o2 X 1   µX2   µX2  µX2   µX2  or normalizing by the square of the mean of the quotient µY. σY2 µ2X 2

=

σ12 µ2X 1

+

σ2X 2 µ2X 2

−

2σ2X 1X 2 µ2X 1µ2X 2

Again, if X1 and X2 are independent RVs, then the covariance is zero. Powers of a Random Variable Single RV: X1 Y = ao X ± b

( ) = ±a b X

∂Y µ X ∂ X1 Mean:

± b−1 1

o

µX

=

± ao b X1± b X1

= µX

± bY X1

= µX

± b µY µ X1

µ Y = a oµ X 1µ X 2

( )

2

2  ∂Y µ   ± bY  X1 2 Variance: σ = σ  = σ X1    ∂ X   µ X1  1   or normalizing by the square of the means 2 Y

2 X1

σY2 µY2

= b2

σ2X 1 µ2X 1

Exponential of a Random Variable Single RV X1: Y = ± a o e ± b X1 where units of the RV X1 are those of 1/b, and units of the RV Y are the same as those of ao.

SL3151Ch14Frame Page 653 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

( ) = ±a b e

∂Y µ X ∂ X1

Mean:

o

± b X1 µX

= ± bY

µX

653

± b µY

µ Y = ± a o b e ± b X1

( )

2

 ∂Y µ  X1 = σ 2X 1 ± a o b e ± bµ X 1 Variance: σ = σ   ∂ X  1   or normalizing by the square of the means 2 Y

2 X1

(

σY2 µY2

)

2

= σ 2X 1 b 2 µ2Y

= b2 σ2X 1

Consider a constant raised to RV power: Y = c ± bX1

( )

then Y = c ± bX1 = e In c Variance:

σ 2Y µ

2 Y

± bX1

(

=e

)

(

)

± b In c X1

2

= b in c σ 2X 1

Constant Raised to RV Power Single RV X1: Y = c ± b X1 where units of the RV X1 are those of 1/b and units of the RV Y are the same as those of c

( )

then Y = c ± b X1 = e In c Mean: Variance:

± b X1

−e

µ Y = c ± bµ X 1 = e σ 2Y µ

2 Y

(

)

(

)

(

)

± b In c X1

± b In c µ X 1

2

= b In c σ 2X 1

Logarithm of Random Variable

( )

Single RV X1: Y = a o In bX1

where units of the RV X1 are those of 1/b and units of the RV Y are the same as those of ao then

( ) =±a

∂Y µ X ∂ X1

=

o

X1

µX

ao µ X1

SL3151Ch14Frame Page 654 Thursday, September 12, 2002 5:59 PM

654

Six Sigma and Beyond

Mean:

(

µ Y = a o In bµ X 1

)

2

 a  Variance: σ =  o  = σ 2X 1  µ X1  2 Y

Example: Horizontal Beam Deflection Deflection of the center of the beam of length L [m] under uniform loading W [N/m] is deterministically given by:

Y=

WL3 = aoWL3 48E I

where E = elastic modulus of the beam material [N/m] and I = moment of inertia of beam cross section about its center of area [m4]. Load and length can be considered r.v. with mean and ± one standard deviation is given as: W = µW ± 1σW = 4000 N ± 40 N L = µ L ± 1σ L = 20 m ± 0.2 m Find: The fractional standard deviation of the deflection Y Mean deflection: µ Y = ± a oµ W µ L Variance of deflection:

( ) 

 ∂Y µ W ,L = σ σ   ∂W  2 Y

2 W

(

= σW2 ± aoµ3L

)

2

3

2

( ) 

 ∂Y µ W ,L +σ    ∂L   2 L

(

+ σ2L 3aoµW µ2L

)

2

Fractional variance of deflection of beam: divide by µ 2Y 2

2

σ   σw    2 σl Y   =  +3    µw   µl   µY 

2

 

2

SL3151Ch14Frame Page 655 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

655

For the case given the fractional standard deviations of the two variables are equal: σW 40 = = 0.01 µW 4000 σL = 20 = 0.01 µL Numerical value for the fractional variance of the deflection: 2

 σY  2 2 2 2  µ  = 0.01 + 3 0.01 = 10 0.01 = 0.001  Y

( )

( )

( )

Numerical value for the fractional standard deviation:  σY   µ  = 0.032  Y Observations: 1. Although W and L have the same fractional standard deviation (0.01), the length — because it is a third power term in the deflection — is seen to have more significance on the standard deviation of the deflection. 2. The fractional standard deviation of the deflection Y is considerably larger than those of either the weight W or length L.

Example: Difference between Two Means Y = X1 − X2

Examples: 1. Clearance 2. Before and after comparison (e.g., treated vs. untreated) 3. Comparison of two suppliers

SL3151Ch14Frame Page 656 Thursday, September 12, 2002 5:59 PM

656

Six Sigma and Beyond

Mean: µ Y = µ X 1 − µ X 2 or Y = X1 − X2 Variance (assume independent so covariance is zero): σY2 = σ2X + σY2

sY2 = s12 + s22

or

Standardized form of sample difference X1 − X2

Z=

s / n + s22 / n2 2 1

t-distribution form of sample difference: Introduce “effective sample variance”

sY2 =

(n − 1)s + (n − 1)s ( n + n − 2) 2 1

1

1

2

2 2

2

then T=

X1 − X2 sY2 / n1 + sY2 / n2

MISCELLANEOUS In Chapter 11, we discussed axiomatic design and its four mapping domains (CAs, FRs, DPs, and PVs). Now, let us examine some of the mathematical relations for these domains. If, for example, we are interested in the functional requirements [FR(CTS)], then this can be expressed in the traditional six sigma notation of y = f(x) as FR = f(DP), where DP (design parameter) is an array of the mapped to DPs of size m. If we let each DP in the array be written as DPi = g(PVi), where PVi, i = 1,…,m, is an array of process variables that the mapped to DPi, soft changes may be implemented using sensitivities in physical (FR and DP) and process (DP and PV) mapping. Using the chain rule, we have: ∂FR  ∂FR   ∂DPi  = ∗  = f '[ g( PVi )]g'( PVij ) ∂PVij  ∂DPi   ∂PV j  where PVij is a process variable in the array PVj that can be adjusted to improve the problematic FR. The first term represents a design change while the second one represents a process change. An efficient DFSS strategy should utilize both terms in all potential improvements. After all, the ideal DFSS outcome is a design that a) exceeds customer wants, needs and expectations, b) exceeds competition market

SL3151Ch14Frame Page 657 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

657

TABLE 14.1 Possibilities of Selecting a DFSS Problem Zs

Zs does not exist

Zs exists

Xs does not exist

No problem; this type of design may not exist

Xs exists

Trivial problem; this type may be solved with design of experiments (DOE)

Need conceptual change; DFSS has potential while six sigma has no potential Both six sigma and DFSS have potential

performance as measured by reliability, robustness, and life cycle costs indices, and c) the rest of product features. This is very important because as we said earlier, DFSS is not for all designs and processes. We must be selective in how we use it. Table 14.1 may be of help. A final point about the axiomatic designs. The importance of the design or problem matrix has many perspectives. The main one is the revealing of coupling among the CTs. Knowledge of coupling is important because it gives the designer clues about where to find solutions, and make adjustments or changes and how to maintain them over the long term with minimal drift. So, for the uncoupled matrix we have  y1   A11    .  0  = .   . y   0  m 

0 A22

.

.

0

0  x1    .  .    0  .   Amm  xm 

For the coupled matrix we have  y1   A11     .   A21  = .   . y   A  m   m1

.

A12 A22

. .

Am ( p−1)

A1 p  x1    .  .    A(m−1) p  .    Amp  x p 

For the decoupled matrix we have  y1   A11     .   A21  = .   . y   A  m   m1

0 A22 . Am 2

. 0 . .

0  x1    .  .    0  .   Amm  xm 

SL3151Ch14Frame Page 658 Thursday, September 12, 2002 5:59 PM

658

Six Sigma and Beyond

These design matrices are obtained in a hierarchy when the zigzagging method is used (see Chapter 11). At lower levels of hierarchy, sensitivities can be obtained mathematically as the CTSs take the form of basic physical and engineering quantities. In some cases they are not available, and that means that the experimenter has to rely on some kind of simulation or modeling.

CLOSING REMARKS This chapter, especially, has focused on some mathematics that will allow the experimenter to pursue design for six sigma (DFSS). The rationale for this mathematical background (review) was to present a case for the integration of six sigma methodology with scientifically based design methods — in particular, reliability, axiomatic designs and the Define, Characterize, Optimize and Verify (DCOV) model in general. In Volume VII of this series we are going to use this background to show how important the mathematical base is and how one may apply this knowledge to optimize designs over two phases: 1. the conceptual design for capability phase, and 2. the tolerance optimization phase. Needless to say, all that may be done with understanding and application of “robustness” in our designs, products, processes, and so on.

SELECTED BIBLIOGRAPHY Chase, K.W. and Greenwood, W.H., Design issues in mechanical tolerance analysis, Manufacturing Review, 1, 50–59, 1988. El-Haik, B. and Yang, K., An Integer Programming Formulations for the Concept Selection Problem with an Axiomatic Perspective (Part I): Crisp Formulation, Proceedings of the First International Conference on Axiomatic Design, MIT, Cambridge, MA, Oct. 21–23, 2000. El-Haik, B. and Yang, K., An Integer Programming Formulations for the Concept Selection Problem with an Axiomatic Perspective (Part II): Fuzzy Formulation, Proceedings of the First international Conference on Axiomatic Design, MIT, Cambridge, MA, Oct. 21–23, 2000 Hubka, V., Principles of Engineering Design, Butterworth Scientific, London, 1980. Hughes-Hallett, D., et al., Calculus, 2nd ed., Wiley, New York, 1998. Kacker, R.N., Off-line quality control, parameter design, and the Taguchi method. Journal of Quality Technology, 17, 176–188, 1985. Kapur, K.C., An approach for the development for specifications for quality improvement, Quality Engineering, 1(1), 63–77, 1988. Kapur, K.C., Quality engineering and tolerance design, Concurrent Engineering: Automation, Tools and Techniques, Kusiak, A., Ed., John Wiley & Sons, NY, 287–306, 1992. McCormick, N.J., Reliability and Risk Analysis, Academic Press, New York, 1981. Stewart, J., Multivariable Calculus, 4th ed., Brooks/Cole Publishing Co., New York, 1999. Strang, G., Linear Algebra and Its Applications, 2nd ed., Academic Press, New York, 1980. Suh, N., Design and operation of large systems, Journal of Manufacturing Systems, 14(3), 1995.

SL3151Ch14Frame Page 659 Thursday, September 12, 2002 5:59 PM

Limited Mathematical Background for Design for Six Sigma (DFSS)

659

Suh, N.P., Development of the science base for the manufacturing field through the axiomatic approach, Robotics & Computer Integrated Manufacturing, Vol. 1 (3/4), pp. 397–415, 1984. Suh, N.P., The Principles of Design, Oxford University Press, New York, 1990.

SL3151Ch14Frame Page 660 Thursday, September 12, 2002 5:59 PM

SL3151Ch15Frame Page 661 Thursday, September 12, 2002 8:00 PM

15

Fundamentals of Finance and Accounting for Champions, Master Blacks, and Black Belts

This chapter is unique in the context of the six sigma and design for six sigma (DFSS) methodology. Our intent is not to present a complete course in financial management, but to introduce some key financial concepts for the Black Belt and Master Black Belt in dealing with projects, Champions, and management in general. As we have repeated many times, the intent of six sigma/DFSS is to satisfy the customer and make a profit (however defined) for the organization. Well, for Black Belts as well as Master Black Belts, that may be a goal, but the truth of the matter is that the majority of them have no clue about accounting or financial issues. In this chapter, we hope to sensitize all those individuals who are about to fix or improve or even contemplate a change in the system of operations with some understanding of the consequences of their recommendations to the organization as a whole. We do not pretend to have covered the topic exhaustively, but we believe that this is the minimum information that Champions, Shoguns (Master Black Belts), and Black Belts must have to be effective not only in selecting their projects but also in evaluating their outcome. We hope that the reader will understand that the discussion here is very broad and covers small and large organizations. As a consequence, not everyone pursuing six sigma/DFSS will encounter all the issues presented here. However, regardless of the organization, regardless of the project, somebody, somewhere, somehow in the organization will be asking or being asked the questions addressed in this chapter.

THE THEORY OF THE FIRM Ask a roomful of business people what the goal of their business is. “Maximize profit” will be the answer you hear most often, maybe exclusively. But is the individual manager really concerned with maximizing profit, or maximizing anything for that matter? Are you? The basic financial decisions of a company are concerned with (a) capital investments — for plants, equipment, working capital, etc., (b) pricing, (c) the level of production, and (d) the source of the money — either debt or equity — to do it all. How we make those decisions was the subject of Adam Smith’s book, Wealth of Nations, published in that distinguished year 1776 (also known for the issuance of Gibbon’s Decline and Fall and some local political events). 661

SL3151Ch15Frame Page 662 Thursday, September 12, 2002 8:00 PM

662

Six Sigma and Beyond

Smith told how the pernicious sins of covetousness, gluttony, sloth, and greed were somehow led by an “invisible hand” to benefit society. His remarkable work was the cornerstone of those studies now called microeconomics, and a good many business decisions can still be explained with Smith’s basic doctrine: Knowing their product’s demand, competition, and cost, business people will act to maximize profits. However, despite its simplicity and power, the theory suffers in real world application for two reasons. First, our information about product demand and competition is usually slight, at best, and even costs — though largely under our own control — occasionally veer away from expectations. Second, the wide separation of ownership and management in the modern corporation has brought additional motives to the mix; the invisible hand now guides by remote control, and the guided managers have ideas of their own. New theories have come forth since the 1950s to improve and update the original model. They attempt to include the impact of the manager’s motives, and because they concern you and me and what abides in our hearts, they are rather interesting. One theory has it that companies are concerned more with maximizing sales than profits. That might explain, for example, the current fascination with mergers and acquisitions that produce instant sales growth yet from a profit standpoint are often failures, about one third of them according to experts. You cannot be in management very long without seeing examples of profits sacrificed to sales: special discounts, loose credit terms, “prestige” products, low bids. Why? Because the size of a company, as measured by sales (a la the Fortune 500 list), is what brings managers the greatest satisfaction, salary, distinction, and seeming success. Moreover, we all identify with the company we keep and the one that keeps us. We take unto ourselves a bit of the power, reputation, and recognition associated with our employer. That may be only a small satisfaction, but it is considerably larger than the one we get from making profits for unknown shareholders. In fact profits, if they are too large, may be thought unseemly and become an embarrassment to us. When we are offered a bonus that is tied to net income, it is in part an attempt to overcome our natural qualms about “excess” profits.

BUDGETS Another modern theory suggests it is the size of his or her budget that gives a manager the most satisfaction. How often have you ascribed this motive to your governmental and not-for-profit colleagues? But it might apply to most corporate middle managers, as well. The number of employees you direct has a bearing on your salary; the amount of money that flows into and out of your control is a measure of your importance; the size of your department often dictates the size of your expense account, company car, office, etc. These things create far stronger urgings that a few extra pennies added to the EPS.

SL3151Ch15Frame Page 663 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

663

OUR ROMANCE WITH GROWTH A third model styles rate of growth as the principal objective of management. In annual reports, the obligatory sales graphs (that look like stairways to heaven), the inevitable “percentage change” figures, and the discussions of future products all bear witness to our fascination with growth. On the darker side of this enchantment are some questionable effects: the use of “creative accounting” to shift profits from one year to another (in order to smooth out the growth curve); the sacrifice of research and development (R&D) and other long term efforts so as to maintain current profit growth; and the indiscriminate purchase of earnings through acquisition and merger. Corporate growth, particularly fast growth, is a stage, not a characteristic. Its prime cause is customer demand for our products, something we can exploit but do little to create. Nothing we are aware of, including the universe itself, can expand forever, and a fast rate of growth (say 25+%) for businesses rarely lasts more than a decade.

THE NEW INDUSTRIAL STATE In his writings, John Kenneth Galbraith has argued that the executives of our larger corporations were moved more by a desire to remain secure and expand their influence than a longing to maximize the gain of a faceless, uncaring, avarice-driven, constantly changing body of shareholders. Security is a rare commodity in American business management, and, like atmosphere, it thins out the higher you go. Yet so potent is anxiety that when our security is threatened we may sacrifice our dignity, our better judgment, our friends, even our health to regain it. Now I am the first one to admit that everyone (me, too) needs a kick in the fanny now and then. But those firms that promote or even permit a rat race mentality can expect, and deservedly so, that their executives will make the securing of their own positions the first order of their business.

BEHAVIORAL THEORY Finally, there is the theory that suggests that maybe neither profits nor anything else is being maximized in the modern corporation. The firm is not one body under a single direction, but at least four bodies, each contributing a required input, and each seeking a different reward. The basic four are the shareholders, executives, employees, and government. They cluster together as does a cloverleaf, four distinct parts joined at the center. That center is a shifting axis representing the economic profit of the business. Each group demands its share in the form of taxes, dividends, security, better working conditions, and the like. Each has the power to close down or sabotage the business. If you accept this theory, then the task of management is not to maximize the shareholders’ immediate profits but by satisfying all groups, to forge a cooperative effort (optimize resources) that will yield a bigger reward for each. It is rare when

SL3151Ch15Frame Page 664 Thursday, September 12, 2002 8:00 PM

664

Six Sigma and Beyond

the various factions of a business pull together, but when they do the results are astonishing.

ACCOUNTING FUNDAMENTALS ACCOUNTING’S ROLE

IN

BUSINESS

To understand accounting’s role in business we might first look at the principal task of management. The manager’s job is to control and direct the business affairs under his or her command. To do so, the manager must understand the effects of past business transactions and thereby be able to estimate the effects of proposed future undertakings. Accounting has the dual role of (1) recording every occurrence that has a financial impact on the business, and (2) reporting these financial data in a form useful to management. Let us first look at the reports that accounting prepares for management, then later at the way transactions are recorded.

FINANCIAL REPORTS The balance sheet, income statement, and other reports summarize the results of a company’s activities. When all of the talking is done, it is to them you look to see how well the company is really doing. This is done through an evaluation of the assets, liabilities and owner’s equity or a balance sheet. Accountants are financial historians. Their task is first, to record every event in the life of a business that has a monetary impact; and second, to report those proceedings in forms that show management how far the company has come and in which direction it is heading. The Balance Sheet “Balance sheet” is the age-old name of a report that sets forth the assets, liabilities, and equity of a company. As accountants have become more educated and higher priced they have tried to substitute fancier names such as “statement of financial position” or “statement of condition,” but the old name lingers on. There are two important balancings or equalities in this report. The first is usually referred to as the “balance sheet equation”: Assets = Liabilities + Equity The counterpoise of these factors is the essence of the double-entry bookkeeping system, which says that for every action there is a reaction, for every benefit received a benefit bestowed, or an obligation to do so. Thus, for every dollar of assets owned by a company, someone among the creditors and shareholders holds a claim check. The other equality in the balance sheet is hidden, or rather, undisclosed. Although each item listed has a dollar value, there is another quality about it that is not revealed: The dollar amount is either a debit balance or a credit balance. The assets have debit balances while the liability and equity accounts have credit balances, so that on every balance sheet

SL3151Ch15Frame Page 665 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

665

Debits = Credits In our system of accounting, debits also represent expenses on the income statement, while revenues normally have a credit balance. The balance sheet represents the condition of the company on a particular day — in fact, the last working moment of that day. Every subsequent transaction changes it to some degree; an employee coming through the door to work the next morning, for example, starts the meter running on the liability called accrued salary expense. An undated balance sheet, therefore, is meaningless. Most condition reports actually give two balance sheets — the current one and one from the year before — so that a quick comparison can be made. Often, the changes in a balance sheet, from one year to the next, are more significant than the ending numbers themselves. Current Assets and Liabilities The balance sheet is normally divided by debits and credits; that is, the assets appear on the left side (or at the top), while the liabilities and equity accounts are on the right (or on the bottom half of the page). On each side (or in each section), the items are listed in the order of their exigency: their nearness to being converted to cash in the case of assets; their nearness to being paid off in the case of liabilities and equity. In the evolution of the various assets to and liabilities to maturity, a sharp line is drawn at one year beforehand. Those assets such as inventory, accounts receivable, and cash itself that are expected to convert to cash in the ensuing twelve months are called current assets. Likewise, those debts that will come due before the next annual financial statement are classified as current liabilities. The accuracy of these classifications is important in measuring a company’s liquidity — its ability to pay debts on time. Fixed Assets Items that are used in running the business, as distinguished from those things that are made or held for resale, are called fixed assets. Fixed assets are typically listed below the current assets, something like this: Property, plant, and equipment Less: accumulated depreciation Net fixed assets Accumulated depreciation shows how much of the cost of existing fixed assets has been expensed. It amortizes the cost over a period roughly akin to the useful life of the assets. Here is the accounting entry for the yearly write-off: Debit: depreciation expense (An income statement account) Credit: accumulated depreciation (The balance sheet account)

SL3151Ch15Frame Page 666 Thursday, September 12, 2002 8:00 PM

666

Six Sigma and Beyond

Accumulated depreciation has a credit balance. When it is listed, therefore, among the assets (which are debit balances), it is a negative amount. Other Slow Assets “Slow” refers to the fact that in the ordinary course of business these assets are not likely to be converted to cash in the coming year. Goodwill represents the premium over book value paid by one company when buying the assets of another. Down-to-earth accountants call goodwill, goodwill. Others label it something like, “Excess of cost over book value of acquired assets...” For example, Company A has assets with a book value of $1 million. Company B negotiates to buy those assets for $1.2 million. The accounting entry on B’s books would look like this: Debit: Assets (various kinds) Debit: Goodwill Credit: Cash

1,000,000 200,000 1,200,000

Because goodwill represents one buyer’s estimate of a worthwhile premium, bankers and other financial analyzers often eliminate it from consideration as an asset. (More on that later.) Current Liabilities Obligations that are due to be paid are called current liabilities. Current liabilities is an important classification to analysts because it represents money that must be paid from future receipts. Most current liabilities are renewable (or revolving), as long as creditors have confidence in the debtor. Working Capital Format Now and then you will see a balance sheet in a “working capital” format. (Working capital equals current assets minus current liabilities.) The layout might be something like this: Current assets Current liabilities Working capital

$1000 500 $500

Other assets

1200 $1700

Other liabilities Shareholders’ equity

$800 900 $1700

While the creators of this format probably had good intentions, it is confusing to read and even irritating because there is no figure for total assets. The working capital figure is of little use and may even be harmful if it is taken to be something it is not. You may subtract current liabilities from current assets on paper, but you cannot do it in real life; current liabilities are reduced only by cash.

SL3151Ch15Frame Page 667 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

667

Noncurrent Assets The noncurrent assets are those that take longer than a year to liquidity (e.g., longterm receivables), and those that the company has no intention of selling, such as property, plant, equipment, vehicles, and other so-called “fixed assets.” The fixed assets are listed at what they cost, less depreciation, and on the balance sheet itself no attempt is made to show their current market or replacement value. Intangible assets such as patents, organization expense, and goodwill (usually called something like “cost in excess of book value of acquired assets”) are also shown in the noncurrent assets section, although they may not be labeled as “intangible.” Noncurrent Liabilities Among the noncurrent liabilities are bonds payable and other “long-term debts,” deferred compensation, and maybe accrued pension liabilities. Any part of these obligations that falls due within the next 12 months is listed in the current liability section. Also frequently found here is the deferred income tax account, which is a liability in theory but seldom in practice; accountants (and everybody else) are so unsure about how to categorize this account that they usually skip giving a total liability figure on the balance sheet just to avoid having to classify it. On about one out of five balance sheets you will run into “minority interest.” It is usually found in between liabilities and equities because it is neither one nor the other. Minority interest represents the outside shareholders of not fully owned subsidiary corporations; the amount is not payable to them unless the subsidiary is closed down and liquidated. Shareholders’ Equity The remainder of the balance sheet is given over to the equities. Some accountants refer to them as a form of liability. They are … if you strain a little and reason that the company assets that are not owed to the creditors are owed to the stockholders. But in modern usage, equity is distinguished from liabilities, which are obligations to make payments on specified dates. Shareholders may be entitled to the equity share of the assets, but “cashing out” is a practical impossibility unless a majority of them act to liquidate the company. Of course, shareholders may sell their interest if the stock is publicly traded or they can find a buyer. Stockholders of today think of themselves more like depositors in an institution than owners of a company. The security, comfort, and convenience of modern investing has been purchased with the power and influence shareholders once had. As we stressed earlier, the balance sheet is constantly changing, and the changes year to year often give a clue as to where the business is heading. That information is given in the statement of changes, discussed below, but one item in the equity section — retained earnings — has a whole separate report to show how much and why it changed. That report is called the income statement. The Income Statement This statement is a report of a company’s sales, less the expense involved in getting those sales, and the resulting profits. It used to be called the “profit and loss”

SL3151Ch15Frame Page 668 Thursday, September 12, 2002 8:00 PM

668

Six Sigma and Beyond

statement — and was nicknamed “P&L” — but in the turbulent sixties corporations became sensitive to the word “profit,” and it has all but disappeared from their public utterances. (Nonprofit organizations are supersensitive about the word, as you might imagine, and refer to their profits as the “excess of revenues over expenses,” or some such dignified euphemism.) Income statements begin with the grandest number found in the business, revenues...the fount of all profits. In most firms the term “revenues” means sales, but there may be other forms of revenue, too — interest income, rents, royalties, and so on. The sales figure is usually “net” of returns and allowances. Gross Profit The rest of the income statement is a process of distilling the revenues by boiling off expenses at various stages until you are left with the essence of net profit. The figures you get along the way vary in importance. The first step is the deduction of the cost of sales (or cost of goods sold) — the largest expense in most companies. Sales – Cost of Sales = Gross Profit From these figures you can derive the gross profit margin (it is rarely given in the report) Gross Profit = Gross Profit margin Sales Gross profit (GP) and gross profit margin (GPM) are important because they reflect the basic climate of the business. In the typical firm, the gross profit margin will not vary more than two or three percentage points from year to year. If the figure is trending down, it may mean the company’s product line is getting old or the pressure from competitors is increasing — both of which are major problems. A Gaggle of Profits Like a gaggle of geese whose symmetrical formation in flight points gracefully toward their goal, profit calculations often taper gently inward as they descend to a point on the bottom line. Along the way you might find figures for: • • • • • •

Operating income EBIT (earnings before interest and taxes) Income before nonrecurring items Income before extraordinary items Income from continuing businesses Income before taxes

You might well ask whether all these numbers clarify the profit picture or deform it. Perhaps the biggest benefit of EBIT is that it gives management a bigger number to talk about. Most companies borrow money and pay interest with regularity; they

SL3151Ch15Frame Page 669 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

669

would not stop if they could, which they cannot, so there is not much point to deriving a profit without such a routine expense. The same could be said for profit before income taxes. It is like saying “look how much money we could make if we did not have to pay taxes.” So what? It would be as useful, and perhaps more interesting, to show us “income before president’s salary,” or “income before expense accounts.” On the other hand, the income before nonrecurring expense or rather, the nonrecurring expense itself can sometimes be revealing. Most often these charges are the bite-the-bullet kind; the company has a losing product or division or subsidiary that management decides to dump. There is some psychology at work here. The thought of profits being attrited year after year by some feeble division is depressing; the cost of getting rid of such a ball and chain is almost inconsequential, so long as it can be tagged nonrecurring. Management is saying “sure, there have been some problems or mistakes, but now they are behind us and we can look to a brighter future.” If you find such a write-off in some company’s glossy annual report, just turn to the front pages where the recent acquisitions and new products are described with unfettered optimism; see if you can guess which of them will be tomorrow’s nonrecurring expense. Earnings per Share The income statements of public corporations also give an earnings per share (EPS) figure. From the net income is deducted dividends, if any, on the preferred stock, and the remainder is divided by the number of common shares outstanding. The Statement of Changes The Statement of Changes in Financial Position is descended from a family of “funds statements” that include (a) the sources and applications of funds, (b) the sources and uses of cash, and (c) the where-got, where-gone statement. The purpose of the report is to describe whence money has come into the business, and how it has been used. There are, of course, thousands or millions of little pieces to that puzzle, so the statement of changes does some wholesale netting to get the report down to a manageable size. All of the transactions involving sales and expenses are combined in a net income or loss figure; to this are added back those deducted expenses that did not take any cash, such as depreciation, amortization, and deferred income taxes. The total of these items is often called the “sources of funds from operations.” The changes in the current accounts may be grouped together as a change in working capital (current assets minus current liabilities). Sources of Funds or Cash Besides the profits and noncash expenses, any increase in a company’s liabilities is considered a source of cash. Think of borrowing from a bank; you sign a note that increases your debts, and you walk out with a pocket full of cash. On the other hand, any decrease in assets is also a source of funds — as when you sell one of your trucks for cash.

SL3151Ch15Frame Page 670 Thursday, September 12, 2002 8:00 PM

670

Six Sigma and Beyond

Use of Funds Typically, the principal use of funds is for additions to property, plant, and equipment; also found here are increases in other slow assets, dividends paid, and net reductions in debts. A balancing figure — the change in working capital — is either included here or listed just below this section. Changes in Working Capital Items Some statements of changes have a section showing the changes in current assets and current liabilities. The net changes in the current assets — please pay attention, this is not easy — the net changes in the current assets minus the net changes in the current liabilities equals the net change in working capital. This figure will match the change in working capital calculated in the sections dealing with noncurrent assets and liabilities and equity. Say what? If after this simple explanation of the statement of changes you feel as if your brain is turning to mush, be assured it is not your brain that is the problem; it is the statement. Anyone can have a dud in his or her bag of tricks, and this is one the accountants have. The statement is hard to understand and has so many exchangeable opposites that the words increase and decrease tend to lose their meaning after a few minutes. Not many people, I have found, bother to read this report; but of the nonaccounting stalwarts who do, most fail to understand it, or worse, they misinterpret it. Nevertheless, the changes in the balance sheets, one year to the next, may be important. If that is the case, you can usually get just as good information — and sometimes better — by simply subtracting the side-by-side numbers in the two balance sheets listed, rather than from struggling with this unfortunate report. The Footnotes There is a cliche among analysts and accountants that the real lowdown on a firm will be found in the footnotes. There is usually plenty of information there, all right, maybe four times again as much as in the financial statements themselves. But the footnotes in a financial report are, like footnotes anywhere else, related information of lesser importance. Anything with a serious financial consequence will be expressed on the statements, and while additional details can often be found in the footnotes, they may or may not be of interest to you. A classic example on footnotes is the 1986 annual report of General Motors Corporation. It had a total stockholders’ equity figure on its balance sheet of $30.7 billion. In the footnotes, however, there was more than a full page of crammed data that reconciled changes in amounts for five different classes of stock, capital surplus, and retained earnings. While it may give some people comfort knowing that the extra information is there, for most readers it is not likely to add anything to the impression made by the single number. We suggest, therefore, not to bother with the footnotes unless you have a particular need for more details about an item. Another reason to go easy on the footnotes is that the language there is largely technical, and if you are unfamiliar with it you might be led to a wrong conclusion. Besides, all of us in business these days have more information available to us than we have the time to look at it. Excessive information is no friend to a good decision, and it is an enemy to action.

SL3151Ch15Frame Page 671 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

671

Accountants’ Report Financial reports that have been audited by “independent” CPA firms will contain a letter from them stating the scope of their involvement and giving their opinion about the financials. It is usually written in accounting boilerplate. If the letter is signed by an accounting firm, and it contains “in our opinion,” if it does not contain “except for” or “subject to,” and if it has no more than a few sentences, you are looking at a “clean” opinion and can feel very comfortable about the figures. For any exceptions to the above you had better wade through the whole letter — depending on how important it is to you. How to Look at an Annual Report Since I have looked at quite a few financial reports over the last 30 years, allow me to recommend a best way to go about it. The fact is, that there is no one best way for everybody. It is an individual thing, a little like the way you observe a member of the opposite sex walking toward you. You look first at one thing, then another, and if you are still interested you may turn around and look at a third. But each person develops the pattern that suits his or her own individual needs. Same with financials, so here is my pattern. Step 1. Look first to see if the statement has the independent clean opinion described earlier. Anything less means that you need to take a more careful look at the numbers (i.e., testing them against poor common sense and experience), and read every single word on the statements. Step 2: Turn to the income statement and look at: a. The latest net income figure. A loss is a red flag. b. The prior year’s net income to see the direction of profits. Two years’ losses back to back means standby the lifeboats. c. Total revenues or sales to see the direction they are heading. Sales are a proxy for demand for the product — the single most important requirement for success in business. Step 3: Now the balance sheet. Here is the sequence I follow: a. Right side, second to the last figure from the bottom — shareholders’ equity. Compare it with last year’s figure; if the company was profitable, equity should have gone up. Now compare it with the bottom number (total debt plus equity) just below; if the bottom figure is more than twice the equity, the firm may have too much debt. b. Run your eye up the page to the total current liabilities amount; remember the number (rough rounded). c. Now over to the current assets. Check the total of that section; if it is only slightly higher than the total current liabilities, that is bad; twice as much is good. d. Finally, look at cash (plus marketable securities). If it is less than ¹⁄₁₀ the current liabilities, it is not so hot; a good ratio would be 30% or more.

SL3151Ch15Frame Page 672 Thursday, September 12, 2002 8:00 PM

672

Six Sigma and Beyond

Step 4: The next step is to sit back and reflect for a moment. You have made mental tests of statement reliability, profitability, leverage, and liquidity; now form a preliminary opinion of the overall condition: excellent, good, fair, poor, or lousy. If you have a mixture of good news, bad news and/or you have to explain your judgment to others, go on to Step 5. Step 5 (Optional): For a second opinion, ask for professional advice. You will recall that we started this section by saying that accountants have a dual role in business: (1) to record every financial transaction, and (2) to report this financial data in a form useful to management. We have looked at the reports prepared by accountants. Now let us examine how transactions are recorded. The history of humankind, that is, the written record of human activities, goes back about ten thousand years. The earliest evidence of writing that we have discovered consists of some lumps of clay on which Sumerian farmers recorded their livestock — what we might fairly term “accounting records.” Today almost all accounting is done by the double-entry bookkeeping method, which was developed by the Roman Catholic Church. Thus the term “accounting clerk” derives from the word “cleric.” The first evidence of this system dates back to Genoa, Italy in the fourteenth century. I think we would all agree that the accounting profession has had plenty of time to settle on all the right procedures. But judging by the changes still going on in accounting, and the liveliness of the debates about them, you wonder if the development of accounting is even half complete. One reason for the continual changes may be that there is no unifying theory of accounting similar to, say, the supply-demand concept of economics or the ego-id theory of personality. Instead, accounting is based on conventions, that is, rules established by general consent, usage, and custom. These rules are called generally accepted accounting principles (GAAP), and they change from time to time. Accounting, then, is very much alive — if not completely well — and the challenges and opportunities it offers to good management are as fresh as ever.

RECORDING BUSINESS TRANSACTIONS Every business transaction involves both give and take. The double-entry bookkeeping system is an ingenious method of recording these activities in a complete, quick, and puzzling manner. The double-entry bookkeeping system (there are also single-entry systems — your checkbook is an example) dominates accounting in all the industrialized countries. The outstanding characteristic of this system is that it records both sides of every transaction, and every commercial transaction has two sides: there is the thing you give to the other party, and the thing you take in return. To account for a piece of business this give and take must be expressed in dollar values, and the doubleentry system always records an equal dollar amount. Profit or loss is frequently part of the transaction — the factor that makes the give and take balance.

SL3151Ch15Frame Page 673 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

673

TABLE 15.1 A Summary of Debits and Credits

Abbreviations Represent what is They often designate what is Or sometimes They are also the normal balances of

Debits

Credits

Dr Taken Owned Benefits received Assets Expenses

Cr Given Owed Money spent Liabilities Equity Revenues

Debits and Credits The terms debit and credit were devised to represent the take and the give; they are from the Latin language because our modern accounting system traces its origin to Catholic clerics of the fourteenth century. In general, debits represent what is taken from a transaction, and credit what is given up. Debits and credits have no more meaning than that, and we could just as easily have chosen other terms in their place, black and red, for example, or left and right. Of course the words debit and credit have other meanings in our language, but it will only confuse you to try to match them with the narrow usage in accounting. Table 15.1 summarizes the use of these terms in accounting. Sources and Uses of Cash In order not to make this come too easily to the uninitiated, financial people sometimes define debits as the uses of cash, and credits as the sources of cash. These are what I call definitions +; normal definitions plus one mental broad jump. Debits are said to be uses of cash because when cash is spent (that is the credit part of the transaction), something such as an asset is taken and recorded as a debit. They make a similar convolution to label credits as a source of cash. If you borrow money from a bank, the cash they give you is a debit — something received, an asset — but the source of that cash was the bank loan — a liability and a credit. How Debits and Credits Are Used Every item — asset, liability, or equity — on the balance sheet has a dollar value assigned to it; it is the company’s and their CPA’s best estimate of the worth of that item. But each account also has another quality about it, one that is hidden and not expressed: Every dollar amount on the balance sheet is also either a debit balance or a credit balance. If you look back at Table 15.1, you will see that assets are debit balances, while the liabilities and equity accounts are credits. The Balance Sheet Equations You will recall our earlier discussion of the balance sheet equation:

SL3151Ch15Frame Page 674 Thursday, September 12, 2002 8:00 PM

674

Six Sigma and Beyond

TABLE 15.2 Summary of Normal Debit/Credit Balances Normal Balance Type Asset Liability Equity Revenue Expense

Definition

Debit

What is owned What is owed The rest Sales, etc Costs

x

Credit

Balance Sheet

Income Stmt

x x x

x x x

x x

x

The Balance Sheet

The Income Statement

Assets Current Miscellaneous Fixed Intangible Liabilities Current Long term and Deferred Equity Common and Preferred Stock Retained Earnings

Sales Less: Cost of Sales Gross Profit Selling and Administrative Expense Interest Expense Other Income and Expense Income Taxes Net Income Less: Dividends Added to Retained Earnings

Assets = Liabilities + Equity And from the foregoing you can see also that Debit balances = Credit balances They will stay that way so long as every future transaction is recorded with equal amounts of debit and credit dollars. Forget + and – . In the use of debits and credits, they do not stand for plus and minus. Both may be either; it depends on the account they are applied to. If a debit is applied to an account that already has a debit balance, the two amounts are added together, and a larger debit balance results. If a credit is applied to an account with a debit balance, then the amounts are subtracted from one another. A similar rule holds for accounts with credit balances. Debits and credits, in other words, are added to their own kind but subtracted from their opposite number.

CLASSIFICATION

OF

ACCOUNTS

In accounting there are five basic types of accounts. On the balance sheet: assets, liabilities, and equity. On the income statement: revenues and expenses. In Table 15.2 is a summary of their normal debit/credit balances.

SL3151Ch15Frame Page 675 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

675

Recording Transactions Remember the basic rule in accounting that in the recording of a transaction debits must equal credit. We can readily see that every business dealing has both a give and a take to it. When a company buys merchandise it takes the goods and gives money in return. The opposite occurs when the goods are re-sold. When it hires a worker, a company takes the fruits of his or her labor and gives back cash in the form of wages. In a broad sense, debits represent the take in a business transaction, and credits the give. For example, your company buys a new computer, paying $900 in cash. Debit (Dr) Credit (Cr)

Office equip Cash

$900 $900

The give and take aspects of double-entry accounting are easily seen here (they are not always so readily apparent). The debit is what has been received; the credit is what has been given in exchange; the debit also represents a use of cash. In this example, both of the affected accounts are assets. The cost of the computer will be added to the cost of previously acquired office equipment, which is already shown on the balance sheet as an asset (debit). As we are combining a debit entry with a debit balance, the result will be a larger asset account on the next balance sheet. However, the cash used to buy the computer was also an asset (debit). Now when we combine the credit to cash with our beginning debit balance of cash, the result is a smaller debit balance of cash. Since only assets were affected by this transaction, the total of assets was unchanged. There were no effects at all on liabilities, equity, revenues, or expenses. Another example: Your company makes a sale amounting to $2150. As soon as an invoice is issued, the accountants will record the transaction. If the sale is for cash the entry will be Dr Cr

Cash Sales

$2150 $2150

In this situation, both entries are plus amounts. The debit is added to Cash and the credit is added to Sales, for Sales is a revenue account that normally has a credit balance. Note: In this case there are no effects on liabilities, equity, or expenses. The cost of the goods that were sold will be recorded in a separate transaction at the time we derive a new inventory figure. The Two Books of Account The transactions we just looked at and similar ones are recorded in a book called the General Journal. It sets out in chronological order all of the firm’s business dealings. It is like a diary of business transactions. Large firms have special journals, such as the Cash Receipts Journal, for recording certain classes of transactions, which are summarized at the end of the month or year in the general journal.

SL3151Ch15Frame Page 676 Thursday, September 12, 2002 8:00 PM

676

Six Sigma and Beyond

The second book of account is the General Ledger (GL), in which each and every account has its own page, on which all of the journal entries relating to that particular account are transcribed. With the GL you can look up an account such as Cash or Notes Payable or Salaries and see all of the transactions made during the year and the current balance of the account. The Trial Balance The process of closing out the books at the end of the year can be rather elaborate. There are often many adjusting entries to be made, and various accounts must be combined and fitted to form the final financial statements. The first step in that process is the preparation of the Trial Balance (TB). The TB is a listing of all the accounts in the General Ledger with the current balances shown in either a debit or a credit column. Since debits and credits are equal in every transaction, the two columns of the trial balance should also be equal. Finding the two columns equal is the “trial” part, for if they are not, the accountants must locate and correct the errors before they can proceed. The accounts listed in the trial balance are then divided among the balance sheet and the income statement to form those reports. The Mirror Image To the neophyte, the discussion about debits and credits may contribute to some confusion and inconsistencies to the understanding of these two concepts. For example, when we say that a “debit to cash” adds to our cash balance, it can create understandable confusion. And when we put money into our checking account the teller, if he or she speaks to us at all, may tell us that the bank is crediting our account. But is this not debiting instead? The confusion arises from the fact that our accounting entry in recording a transaction is often a mirror image of the other party’s. The cash that we take in (our debit) is the same cash that the other person has paid out (his credit). And the goods we delivered to him are deducted (by a credit) from our balance sheet, and added (by a debit) to his. When the bank tells you that your deposit is being credited to your account, they are speaking from their viewpoint, not yours. Their accounting entry is: Dr Cr

Cash Demand deposits

Since your demand deposit is a credit on their books, an entry crediting your account is one that increases the balance.

ACCRUAL BASIS

OF

ACCOUNTING

Accrual accounting has made us as dependent upon CPAs as we are on MDs and JDs. But if you want to know how much your business is really earning, it is the only way to go. First of all, accrual is hard to pronounce without sounding as if you had a mouth full of bubble gum. Even most accountants, who are steeped in reverence

SL3151Ch15Frame Page 677 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

677

for the accrual basis, give the quick two-syllable pronunciation, “a-krool” rather than the proper three-step version, “a-kroo-al.” Second, it rouses no sense of recognition or meaning. It is one of those words that requires you suspend all other thoughts while you struggle for its gist. And third, even when you remember that accrue means to accumulate or increase, it is still hard to make the connection with the accrual basis of accounting. Now that that is out of my system, accrual is the accounting principle that counts sales as income even though the cash has not yet been received, and records expenses in the period they produced sales although they may have been paid in some other period. If we look at the income statement in a company’s annual report, the first item is sales, for the whole year, though it is likely the invoices from the final month are still uncollected. By the same token some expenses, such as telephone and utilities, are counted although those incurred in the last month may not be paid until the following year. Sometimes there is the opposite effect, where the cash moves first, and the recording of income and expense comes later. For example, a customer sends in a check along with an order; the check may be deposited right away, but no sale is recorded until the goods are delivered. Or, say a company builds an elaborate display in December 2001 for a convention in January 2002. Assuming the company operates on a calendar year basis, the money spent in December would not be counted as an expense until 2002, the year in which the benefits of the expense are derived. Accrual Basis versus Cash Basis As individuals, most of us use a cash basis for tax purposes. We do not report money owed to us as income — only the cash we have received. Nor do we take a deduction for a medical expense before we have paid the bill. Some businesses — usually small — operate that way, too. There are a few advantages. It is a clean and simple way of accounting; the cash receipts and payments records serve for the income and expense statement as well and for tax purposes. The cash basis allows some maneuvering through the use of delayed billing or accelerated payments. The big disadvantage of the cash basis is that it is a crude measure of a fast-paced activity — so crude that a lot of damage can be done before a true assessment is made. Details, Details The accrual basis of accounting sometimes appears overly concerned with particulars. When the year ends one day after payday and an accountant spends hours calculating that one day’s accrued salaries so they can be charged to the old year, you may well wonder if the accounting profession is not feathering its bed. But if there is any one thing about business that is essential for management to know, it is an accurate picture of profit and loss. And given the large numbers we deal with and the slender profit margins that accompany them, getting accurate income figures is worth a lot of expense and bother.

SL3151Ch15Frame Page 678 Thursday, September 12, 2002 8:00 PM

678

Six Sigma and Beyond

Birth of the Balance Sheet Accrual accounting is father to the balance sheet. Look at the assets side of a balance sheet, and draw a line under accounts receivable. Just about everything beneath that line is a prepaid expense, an expenditure waiting to take its place in the expense section of some future income statement. On a cash basis, these assets would be counted as lost costs — a gross distortion of the truth. But with accrual basis accounting we assign them a value commensurate with their potential for producing future revenue. The accrual method is often a pain both to apply and to understand, but consider the investment and credit decisions (and if those do not move you, the management bonuses) that are dependent upon it. The more accurate the measure of our past activities, the better will be our future decisions. Profits versus Cash Businesses are in business to make a profit, but they run on cash. And if you think I am kidding, try getting on a bus with just your income statement. Because accrual accounting distinguishes between the profit effect and the cash effect of transactions, it is necessary for management to have a cash plan as well as a profit plan. Yes, it is possible to be profitable and still go broke, that is, run out of cash. The problem can be acute for highly seasonal businesses or those with a fluctuating cash/credit sales mix. Things Are Measured in Money Most annual reports begin with a hymn of praise by management for themselves, followed by colorful pictures of shiny products and smiling workers. It is the CPA’s job to express all this in terms of dollars in the income statement and balance sheet. From time to time, sentimentalists wonder why the human worth of the employees is not reflected on the financial statements. Most of us know workers who might qualify as assets, and others more properly described as liabilities, but no one has yet come up with an acceptable way of putting a number to these characteristics. Values Are Based on Historical Costs The value of an asset is continually changing as a result of wear and tear on the one hand, inflation on the other. Perhaps the true value of an asset is revealed only at those moments in time when it is sold. Only at those times are we certain of the asset’s hard cash value. For convenience, accountants have fastened on the moment of acquisition to value the fixed assets. The amount originally paid for an asset is the balance sheet value, and no attempt is normally made to adjust that value except for scheduled depreciation. For this convenience we pay a price, particularly in times of high inflation when assets often appreciate. One of the most often-heard criticisms of CPAs is their failure to value fixed assets at the current or replacement value. Leaving aside the controversy there are two things to keep in mind:

SL3151Ch15Frame Page 679 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

679

1. Most assets are difficult to appraise, and there is often a wide difference of opinion. 2. Business assets acquire value from the revenue stream they are able to produce. If the value of assets does indeed rise during inflation, the proof should be found in higher earnings.

UNDERSTANDING FINANCIAL STATEMENTS ASSETS Most assets are little more than deferred expenses. Their value lies in the future sales they can generate. And, as with baking a cake, that takes the right proportion of ingredients. Assets are the things a company owns. All assets have value, but not all things of value are assets under present accounting rules — only those that have a money value. This means that one will find no listing on a balance sheet for the trust customers might have in a company or the team spirit of the employees because those qualities cannot be measured in dollars. What gives an asset its true value is not the materials and workmanship that went into its creation, but rather its ability to help generate a future stream of income. It is not the bricks and mortar, glass, and metal that make a fast food restaurant on a busy street an asset, but the sales that will result from operating the place. The same facility lying in the middle of a wheat field would be of no value at all. The dollar amounts shown on the balance sheet are the original costs (accountants like to say the “historical costs”) of the assets. From the original costs is deducted the accumulated depreciation, if any, against those assets.

THE INFLATION EFFECT For the time being, balance sheets do not reflect the replacement costs of the assets, a fact that has inspired much derision of the accounting profession, especially in periods of high inflation. Beleaguered with various theories of value, accountants have at least picked one where the numbers can be verified. Every dollar on the balance sheet can be traced back to an invoice, contract, or some piece of paper in the files.

SUMMARY

OF

VALUATION METHODS

Historical Cost Historical cost is the present basis of balance sheet values. We can all agree that a thing is worth what someone will pay for it, so for at least one moment in time this was the unquestioned value. Liquidation Value Liquidation value is the under-the-hammer price, as in a bankruptcy sale. Most assets sold off this way would bring only a small fraction of their balance sheet value. (Nine cents on the dollar is a historical average in bankruptcy liquidations.)

SL3151Ch15Frame Page 680 Thursday, September 12, 2002 8:00 PM

680

Six Sigma and Beyond

Investment or Intrinsic Value This is the present value of all future income expected to be derived from the asset, discounted at a rate commensurate with business risk (typically 10% to 15%). While most financial experts would agree that this is an asset’s “truest” value, it is all based on the tricky task of estimating future income. When you apply mathematical precision to a “guess forecast,” you also get another version of a guess estimate with angular numbers instead of round ones. Psychic Value Often a factor in mergers and acquisitions, psychic value looks to the buyer’s state of mind rather than any characteristic of the asset. Unfortunately, trying to divine the hopes and dreams rattling around in the mind of a potential buyer is not any easier than estimating future income. Current Value or Replacement Cost These, as a result of double-digit inflation a while back, got a lot of attention from the Securities and Exchange Commission and the accounting profession, if not businesspeople themselves. The burden of their studies, however, was not that asset values are really much higher than stated, but that depreciation allowances based on historical costs understated the “true” expense, and thus led to overstated profits. The current value issue, like inflation itself, is about as predictable as the common cold, and as frustrating to cure. Bad as historical costing is, CPAs just have not found anything they like better. Assets versus Expenses Granted that it may sound like a contradiction, assets and expenses are very much alike. Except for financial assets (discussed below) and land, assets are little more than prepaid expenses. The reason we just do not call them expenses is that they still have some juice left in them — some power to generate future sales. All expenditures — except payments of debt — result in either an expense or an asset. The distinction rests on how long the item purchased will be of use. If it will be used up by the end of the year it is an expense; if its usefulness extends beyond the present accounting year it is an asset. Therefore, money spent for wages, electricity, or travel results in an expense, while money spent to acquire carpeting, a lathe, or a jet liner creates an asset. Some distinctions are not so easy. Money spent to incorporate a business may be listed as an asset (organization expense) on the theory that it will benefit the company throughout its life. On the other hand, it might be written off at once as just another legal expense. Most of the asset/expense decisions will be made by your CPA using established principles, but there are always some arguable cases. The key question is, do you want to bear the entire expense now or stretch it out? Since most managements exist at the sufferance of the bottom line, it is more than an academic issue. More often than not, if it is a borderline case, the course of action

SL3151Ch15Frame Page 681 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

681

is to expense off what you can and still keep your job. Remember that the issue will not affect your cash balances — the money has already been spent.

TYPES

OF

ASSETS

Assets may be classified according to their tangibility. This is not the usual way we distinguish them in financial reports, but it can add depth to our understanding of the nature of modern business. Financial Assets These include cash, marketable securities, accounts and notes receivable, and investments. Cash is the premier asset — it always gets the first position on the balance sheet. There is little need to explain why, for while other assets may interest us, cash generates something more akin to a fascination. I am reminded of something attributed to the Roman poet, Ovid, who is best known for writing “The Art of Love,” and its antidote, “The Remedies of Love.” He said: “How little you know this world if you fancy that honey is sweeter than cash in the hand.” Now if the poet Ovid sounds a little like an economist, the economist John Kenneth Galbraith sounds a little like a poet when he discusses money: “It ranks with love as the source of our greatest pleasure, and with death as the source of our greatest anxiety.” Accounts receivable are the monies due from customers. Nearly all firms that sell to other businesses sell on open account credit, so receivables usually represent one of the larger kinds of assets you need to run a company. Receivables are “claims on money,” and as such are maybe halfway to being cash. Some people are fond of reminding us that you still have to collect the account before you have something, but the average amount that ends up as bad debts is only about a third of a percent. Other cash claims such as receivables from and investments in affiliated companies may or may not be financial assets, depending on how readily redeemable they are. Like a loan to your brother-in-law, these may be more in the nature of gifts than financial assets. Financial assets make a shiny impression on those you deal with; they give you tactical flexibility; they invite opportunities to knock on your door; and they give you a sense of security and well-being. On the other hand, these financial assets are given to you (management) to do something with besides bathe in their glow, and by themselves they produce limited income. Later we will discuss the question of how much cash is too much. Unlike property and equipment, financial assets do not wear out or become obsolete. They do, however, suffer from inflation and, in the case of marketable securities, from fluctuation in market price. Physical Assets These include the inventory, land, buildings, equipment, and anything else you can paint. Most physical assets are subject to depreciation — the process of writing off (accountants say “expensing”) the assets over their useful life.

SL3151Ch15Frame Page 682 Thursday, September 12, 2002 8:00 PM

682

Six Sigma and Beyond

Exceptions are inventory, which is not kept long enough to depreciate (that is, it had better not be), and land, which is assumed not to depreciate (but do not try convincing the folks in the vicinity of Mt. St. Helen’s). Physical assets — other than these two — can be viewed as lost costs or prepaid expenses. Their value is manifest not in their cost, size, sturdiness, or beauty, but in their ability to create customers. Operating Leverage For the past 250 years, business has gradually increased its physical assets in proportion to the people it employs, through the process we call automation — the substitution of machines for people. This is referred to as operating leverage. In general, higher operating leverage (a higher assets to people ratio) results in higher profits in expansion and higher losses in a sales decline. That factor, so often neglected in capital budgeting decisions, can have a profound effect on a company’s long-term prospects. Determining the Value of Inventory Businesses use three principal methods to assign a value to their inventories: FIFO, LIFO, and the weighted average method. We may think of inventory as a reservoir of goods for sale. At the beginning of the year it stands at a certain level; during the year we add to it by purchasing or manufacturing more goods; from it we take the goods that we sell. And at the end of the year, we measure the level at which it stands. The value of our inventory is what it cost us to make (not what we think we can sell it for), but during the year costs may have fluctuated because of inflation or changes in the supply of and demand for the raw materials or goods we purchase. Moreover, in most companies businesspeople are not sure which goods — the higher costing or the lower costing — were sold and which remain in inventory at the end of the year. Therefore, the amount at which we value our ending inventory as well as the cost of goods sold during the year will depend on the valuation method we choose. FIFO

The First-in First-out (FIFO) method assumes that the oldest goods on hand are the first to be sold, and the inventory remaining consists of the latest goods to be purchased or made. This is a very reasonable assumption since most companies will sell their products in roughly the same chronological order they acquired them. LIFO

The Last-in First-out (LIFO) method assumes that the latest goods acquired were the first ones sold, and the year-end inventory consists of the oldest goods on hand. In most of the firms that use LIFO, this is clearly a fiction. The reason it is accepted is to defer the payment of income taxes. How that works can be explained in a threestep thought process: 1. The history of the world is inflation; we have always had it (except for brief periods), and there is no sign of it disappearing. 2. In inflation, the goods purchased or made earlier in the year are likely to cost less than those acquired near the end of the year. By assuming the

SL3151Ch15Frame Page 683 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

683

year-end inventory comprises the old lower-cost goods, we tend to understate the value of the inventory. Conversely we tend to overstate the cost of goods sold by assuming that the products delivered were the new higher-cost goods. 3. By overstating the cost of goods sold we will understate profits, and with lower reported profits we will have lower income tax payments. Weighted Average

With this method, the goods remaining in inventory will be valued at the same average cost as those that have been available for sale during the year. Depreciation One of our most useful financial concepts — for setting prices, providing funds to replace capital assets, and postponing taxes — is depreciation. Depreciation is the value a fixed asset loses through our use of it and the passage of time. In business, we recognize that depreciation, along with the cost of labor, materials, and taxes, is an expense of running the company. Accountants recognize (record) depreciation in an unimaginative, mechanical way that approximates real life in the long run but may vary widely from it in the short. We must recognize the expense of depreciation in order to correctly price our products and get a true picture of profits or losses. Suppose we bought an ice cream machine for $2000 and started selling cones for 50 cents, after determining that our out-of-pocket expense to make them was 30 cents. It is obvious we are not making a profit of 20 cents on each cone even though we have that much extra in our pocket, because the 2000-dollar machine is gradually wearing out and losing its value (especially when making my favorite, pecan-praline, because the little crunchies in there wear it out faster). It is possible that at a half a buck a cone we will wind up losing money. Useful Life Concept Under present accounting practices, a company that buys equipment or some other fixed asset must estimate the number of years it will use the item and what salvage or residual value it will have when the company is finished with it. The depreciable amount, that is, the cost minus the salvage value, is apportioned to expenses over the useful life of the asset in either equal or formulated amounts. Because depreciation is a legitimate expense, because it does not involve a cash payment to anyone, because it is based upon estimates of future wear and tear, and because a variety of depreciation methods are acceptable to tax authorities and CPAs, the depreciation process is almost an irresistible invitation to tax strategies and fiscal manipulation. The manipulators are themselves manipulated by the government, which frames depreciation rules so as to encourage businesses to buy new production equipment. (Not all organizations, however, bother with depreciation. A list of those that do not would include tiny companies with too little income to deduct depreciation from, as well as giant nonprofit institutions that neither charge for their services nor pay any taxes.)

SL3151Ch15Frame Page 684 Thursday, September 12, 2002 8:00 PM

684

Six Sigma and Beyond

Understanding depreciation can be tricky. In the typical business it has four different applications, each one giving a separate and sometimes opposite aspect, and it is easy to be exactly wrong about depreciation. Here are the four viewpoints: Depreciation as an Expense

Depreciation reduces profits. It always makes them lower. It never adds to or in any other way benefits the bottom line. Got that? In that way, depreciation is like rent, salaries, income tax, or any other expense: the more you have of it, the lower your net income. Yes, depreciation is a non-cash expense, but in the preparation of the income statement or the profit plan of the future, depreciation expense reduces profits. Depreciation as a Valuation Reserve

The word depreciation is also found in the phrases “accumulated depreciation” or the slightly old-fashioned “reserve for depreciation.” In this guise, it represents the total amount of depreciation expense recorded for an asset since it was acquired. Accountants have a peculiar way of recording depreciation. You might think that the accounting entry would be something like this: Debit Credit

Depreciation expense The asset

XXX XXX

Not so. For reasons best known to themselves, accountants like to preserve the original cost of the asset. And so they create a valuation reserve that accumulates the depreciation expensed each year; the accumulation is then deducted on the balance sheet from the original cost of the fixed assets to produce a book value. Here is the accounting entry: DR CR

Depreciation expense Accumulated depreciation

XXX XXX

As you can see, accumulated depreciation has a credit balance; it is located, however, in the fixed asset section of the balance sheet as a negative figure — a subtraction from the original costs of the assets. As a credit nestled in among the debits it is spoken of as a contra account or, being where it is, a contra asset. Do not think of accumulated depreciation as any sort of cash fund. It is simply a number, which when deducted from the original cost of the assets, gives their current book value, that is their undepreciated value. The depreciation of an asset continues until (a) the accumulation equals the depreciable amount, or (b) the asset is disposed of, in which case both the asset and the accumulated depreciation are written off the books. Depreciation as a Tax Strategy

Imagine your income tax for this year. Imagine an expense that you could legitimately deduct. Now imagine this expense could be varied up or down within a certain range, thereby giving you some flexibility in “setting” your income and, therefore, your income tax. If you have any imagination left, think of this expense

SL3151Ch15Frame Page 685 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

685

as existing merely on paper and not requiring any cash. Can you see how, by adjusting this non-cash expense upward, you can actually save yourself cash by lowering your income tax bill? This is the magic of depreciation. Increase any other expense and you have less cash; increase this one and you have more. Behind the enchantment, however, lie some essential truths — realities that are frequently overlooked — to the extent, that is, that real life permits such a thing: 1. Depreciation is not so much a saving of cash as a recovery of cash already spent. Money had to be laid out in the first instance to acquire the assets. In recognition of this some countries — and to a very minor extent our own — permit depreciating the entire cost of the asset in the year it is acquired. 2. When an organization accelerates its depreciation, it is merely borrowing tax deductions from future years, so that the cash saved out of reduced taxes is — in theory, at least —only a loan that will have to be repaid when the company has used up all of its deductible expense. Depreciation as Part of Cash Flow

A simplified definition of cash flow is profit + depreciation. The idea behind it is to measure the extra cash generated by a business — cash received from sales minus cash paid out for expenses. Depreciation expense reduces profits in the first place, but since it is a non-cash expense it is restored to profits when estimating cash flow. The positive role of depreciation in cash flow is so impressive that it often leads people to the mistaken notion that depreciation is a benefit to profits, also. As you can see, however, it is merely a restoration of money that was taken from profits in the first place. It is sort of like what they do when they make white bread — they mill out 65 nutrients and put back a half dozen and call the bread “enriched.” Investing in a business is not unlike making a loan to someone. When a payment is made to you on the loan, only part of it is interest income; the rest is principal, for the balance due you afterward is less. In a similar way, the surplus cash generated by a business comprises “interest” and “principal.” Profit is like the interest income, but the rest of the cash flow — represented by depreciation expense — is a partial return of the original investment, for the value of the assets is now less. The most common methods of depreciation are: Straight Line

This is the standard method used by most companies for financial (but not tax) purposes. Straight line depreciation is the easiest to compute and understand. You simply spread the amount to be written off equally over the years of useful life. The formula is: Original Cost – Salvage Value Years of Useful Life

SL3151Ch15Frame Page 686 Thursday, September 12, 2002 8:00 PM

686

Six Sigma and Beyond

Suppose, for example, you purchased an office copier for $5000, estimated its useful life at 4 years, and thought you could afterward trade it in for $1000. The depreciation would be: ($5000 – $1000)/4 yr = $1000 per year Sum-of-the-Years’ Digits (SYD)

This is a modest form of accelerated depreciation. That is, it makes the charges heavier in the early years, lighter in the later. There are two reasons we may want accelerated depreciation. First, it more nearly matches the way the market value of used equipment drops; just think of the drop in value of a new car the minute you drive it out of the showroom. Second, there may be a tax advantage to speeding up the deductions. The calculation of SYD is a cunning little arithmetic exercise that has nothing to do with real life except that it gets the job done. We start by adding the years’ digits in the estimated useful life. Sticking with our copier example, the calculation would be: 1 + 2 + 3 + 4 = 10 The 10 becomes the denominator in a fraction, the numerator of which for the first year is the last number in the sum: 4. The fraction is applied to the depreciable amount, thus: (4/10) × $4000 = $1600 The second year’s calculation is: (3/10) × $4000 = $1200 and so on through each digit until a total of 10/10, or 100 of the depreciable amount has been expensed. As you can see, the first year’s depreciation under SYD is significantly greater than under straight line. More expense means less income and income tax. Since the total deductible amount is $4000 in either case, however, SYD will have to compensate later on for the big numbers in the early years. Getting the sum of the years’ digits can be tedious if it is a big number, such as 15 years, so here is a formula you can use. Where N = the number of years’ useful life, SYD = [N(N + 1)]/2 For example, [15 × (15 + 1)]/2 = 240/2 = 120 Double Declining Balance (DDB)

This method, which accelerates depreciation even more than SYD, also has the blessing of the IRS. It depreciates at twice the rate of the straight line method, applied to the full cost. In our example, you can see that the rate of the straight line

SL3151Ch15Frame Page 687 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

687

depreciation was 25% per year, because the useful life is 4 years and 4 × 25% = 100%. If the useful life had been 5 years, the rate would be 20%, and so on. Under DDB this rate is doubled and applied to the beginning book value each year. For the first year in our example, the depreciation is: (25% × 2) × $5000 = $2500 As you can see, we have stopped kidding around; we are really talking depreciation now. At the end of Year 1 the book value of our copier is: $5000 – $2500 = $2500 and the second year’s depreciation is 50% × $2500 = $1250. The calculations continue in that manner until the book value is reduced to the salvage value. That usually means there is no depreciation at all in the last years. Unit of Production

This is a non-accelerated method based on usage — the number of units produced or the hours of use. Suppose you thought your copier would give you a half million copies before you traded it in, and the first year you got 100,000 copies. The depreciation under this method would be: 100,000 × $4000 = $800 500,000 Replacement Cost

Replacement cost is a theoretical method not used for either financial or tax purposes; the depreciation is based on future replacement rather than original cost. If you thought that at the end of five years you would have to pay $10,000 to replace your copier, you might consider the “true” depreciation cost to be: [$10,000 – $1000]/5 = $1800 per year Keep in mind, though, that this method is not authorized for financial reporting or income tax purposes. Advantages of Accelerated Depreciation

Depreciation expense is known as a non-cash expense because there is no out-ofpocket payment associated with it, as there is with almost every other expense. Of course, a payment is made at the time the asset is acquired. Afterward, however, the amount or rate of depreciation has no effect on a company’s cash except as it affects profits and profits affect the amount of income tax that is paid. By selecting an accelerated method of depreciation, a company can postpone the payment of some taxes that would be paid using the standard straight line method. This postponement is very like an interest-free loan from the government.

SL3151Ch15Frame Page 688 Thursday, September 12, 2002 8:00 PM

688

Six Sigma and Beyond

FINANCIAL STATEMENT ANALYSIS Financial statement analysis has long been considered an art. Though the numbers that analysts deal with are specific, the interpretation of those numbers is not, and even the question of which numbers relate in a meaningful way is largely unsettled. For example, in my own research, I have come across nearly 150 different ratios used to gauge financial condition. Each has its coterie of followers — bankers, financial managers, investment analysts, and CPAs — who daily seek some clue to the future by examining the financial statements of the past. Is it possible to predict the outcome of a business venture? Many people think so. Yet there is very little evidence that businesses evolve in a linear way. The random walk of the stock market, the whims of fashion, the variety of life’s experiences all testify that if the future held no surprises, that would be the biggest surprise of all. In recent years, however, a number of developments have acted to make financial statement analysis more scientific and less unpredictable. Included among these are 1. A broader use of mathematics and statistics in defining the major elements of a business and their relationships 2. The use of computers, with their enormous capacity for storing and classifying business data 3. The refinement of accounting practices, which has given us more reliable financial statements To be sure, the main difficulty with financial statements is what may be called the good news/bad news syndrome. Seldom do financial statements look completely good or completely bad; they nearly always exhibit both qualities. There are two principal ways of analyzing the financial strength of a company. One is through a ratio analysis of recent financial statements. The other involves a financial forecast of the near future. Ratio analysis is the easiest to learn and the fastest to use, and that is the method we will examine first. Financial forecasting is more difficult to learn and complex to apply, but it gives superior results. Forecasts often require us to make difficult estimates of unknowns, but they deal in specific goals and dates, such as earnings in the coming year or cash flow in the next 15 months. Ratios, on the other hand, are usually easy to calculate, but the results are often abstractions that may he hard to apply to real world problems. Does knowing that the current assets are 200 percent of the current liabilities tell you if you can pay your bills on time? As we discuss ratios, keep in mind that they are nothing but little numbers unless we have some standard by which to measure them. The 2:1 current ratio mentioned above does not help you much unless you know what number constitutes a good current ratio and whether it gets better as it gets higher, or vice versa.

RATIO ANALYSIS The dollar values of items on the income statement and balance sheet have little significance by themselves. Rather it is the proportion of accounts, or groups of

SL3151Ch15Frame Page 689 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

689

accounts, one to another that tells us whether a company is financially viable or not. For example: Suppose a businessperson tells you his company has $85,000 in its checking accounts. The figure means virtually nothing unless you can relate it to other aspects of the business. If the man runs a local shoe store, he may be well fixed, but if he turns out to be the president of the Eastman Kodak Company, he is talking about the amount of cash that flows in and out of the company approximately every minute of the business day. A major problem with this kind of analysis has been the proliferation of different ratios. Every financial statement lists several accounts; they may be compared to each other or to the same accounts in previous periods; combinations of accounts are related to individual items or to other combinations; and ratios themselves are often divided by other ratios to produce super ratios for determining trends. The possibilities and the confusion seem to be without limit. As an example let us look at a balance sheet and income sheet for a hypothetical Company X. BALANCE SHEET Analyst: Christin R. ---------------------------Statement Date: ASSETS Cash & Short Term Investments Accounts Receivable (Net) Inventories Deferred Taxes Prepaid Expense Current Assets Property, Plant & Equipment Less: (Accumulated Depreciation) Net Fixed Assets Investments Goodwill and Other Intangibles Other Assets Total Assets

Company: X Date May 15, 2002

$Millions

12–28–00

12–28–01

1585 1678 1703 230 50 5246

613 2563 2072 348 215 5811

6861 –3426 3435

12919 –6643 6276

5 68 8754

383 432 12902

1564

3440

482 201

209 142

2247

–3791

DEBT + EQUITY Notes Payable Current Maturities of LT Debt Accounts Payable Accrued Liabilities Taxes Payable Dividends Payable Current Liabilities

SL3151Ch15Frame Page 690 Thursday, September 12, 2002 8:00 PM

690

Six Sigma and Beyond

Long Term Debt Deferred Taxes Other

66 271 142

911 1209 603

Total Liabilities

2726

6514

Preferred Stock Common Stock Retained Earnings Less: (Treasury Stock)

674 5354

936 6533 –1081

Equity

6028

6388

Debt + Equity

8754

12902

Company: X Date: May 15, 2002

$Millions

12/28/00 Year

12/28/01 Year

Net Sales or Revenues Less: Cost of Sales

9734 6085

11550 7613

Gross Profit

3649

3937

Expenses: Selling, G&A

1753

2693

Operating Expenses

1753

2693

Operating Income [EBIT]

1896

1244

Interest Expense or (Income) Other Expense or (Income) Nonrecurring Expense or (Income)

–86 19

71 55 520

Income Taxes

809

224

Net Income

1154

374

Cash Dividends

517

551

INCOME STATEMENT Analyst: Christine R. Statement Date: Period:

Depreciation Research and Development

SL3151Ch15Frame Page 691 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

691

Ratio analysis encompasses scarcely a half dozen generations of analysts, so ratio names are not well settled or precisely defined. They are more so within particular groups such as CPAs, bankers, and stock market analysts. But between those groups the same ratio may have different names and the same name may be used for different ratios. Since this book is meant for a broad range of executives, I have used names and definitions as I found them in general business use, rather than in the specialty fields. Liquidity Ratios Liquidity refers to the ease with which an asset can be converted to cash. The liquid assets in a business are cash itself and those things that are near to being cash, such as accounts receivable, or that are readily convertible, such as marketable securities. The Securities and Exchange Commission and countless analysts have defined liquidity as the ability to pay debts when they come due. A gutsier definition might be simply “enough cash.” But enough for what? The answer to that is usually found in the denominator of liquidity ratios. Enough cash to pay the bills coming due; enough to pay recurring expenses such as payroll; and enough to cover unexpected needs and opportunities. In addition, that simple question often yields a perplexing answer. The elements of liquidity are in an active state of flux. Both the amount of cash a business has on hand and the amount it is obligated to pay changes with virtually every transaction that occurs. And even a modest-sized company may have 1000 employees spending its money — and 10,000 customers sending cash in. It is difficult if not time-wasting, therefore, to contemplate cash needs moment to moment. Most firms try to forecast cash flows in and out for a day, a week, or a month, and then add a cushion to cover normal variances. An even more serious problem in managing liquidity is that we are obliged to weigh an uncertainty against a certainty. There is little in life that is as fixed, certain, and unremitting as a debt owing. On the other hand, few things are as inconstant, fickle, and capricious as payments promised, loans pending, and sales forecasted. In using liquidity ratios it helps to identify the certain and uncertain elements, and how much of the latter it takes to balance the former. For example, in the popular current ratio (current assets/current liabilities), we see a blend of uncertainties in the numerator. We can count on the cash we already have, but the timing of receivable collections is somewhat uncertain, and the sale of inventory even more so. The amounts and due dates of the current liabilities, however, are known and fixed. In matching up the two elements, therefore, we know instinctively there should be more current assets than current liabilities in order to offset the uncertainty of the former. Can a firm be too liquid? Can it have too much cash for its own good? There are certain “conventional truths” that circulate among businesspeople that do not bear close scrutiny so well. One of them says that the company with abnormally high cash balances may be “missing investment opportunities that could bring growth and profits.” In my view, cash is the beginning asset in business and the final asset. And during the game it is like the queen on a chessboard. It travels in all directions, any number of squares. That is, it is the most powerful and flexible of assets, and

SL3151Ch15Frame Page 692 Thursday, September 12, 2002 8:00 PM

692

Six Sigma and Beyond

as long as you have plenty of it you are in a superior position for taking advantage of opportunities that float your way. Financial Leverage Financial leverage is the mix of debt and equity in a business. The perfect mix is one that exactly balances the entrepreneur’s love of leverage and the creditor’s fear of it. Leverage is the relationship between the amount of money creditors put in a business and the amount the owners contribute. Where there is plenty of debt and not much equity we speak of high leverage. Where there is little debt and lots of equity we talk of low leverage. Since leverage refers to the relationship of a firm’s debt and equity, it stands to reason that a ratio of debt to equity will measure it. And debt to equity is in fact the most popular ratio for gauging leverage. Other well-known leverage ratios include equity/debt, assets/equity, and debt/assets. All of these ratios have a direct mathematical link and tell exactly the same story. Only the scale is different. The problem is not in measuring leverage so much as it is in knowing when a company reaches a reasonable debt limit. Unfortunately, we cannot tell how much leverage is enough except by noting when there is too much. When a company goes bankrupt, we can say with a measure of confidence that the company should have had a little less leverage. At that point, however, the question itself is usually academic. Coverage Ratios Coverage ratios are intended to measure a company’s ability to pay the interest on its debt from its earnings. Some financial people consider these a form of leverage ratios, but in reality they are nothing more than earnings ratios, when they are useful at all. The most popular coverage ratio is the Times Interest Earned ratio. The formula is [Profit before interest and taxes]/Interest Earnings Of the three major financial characteristics — liquidity, leverage, and earnings — the last is the most complex but the easiest to understand. Simply stated, earnings are derived out of accounting’s most fundamental formula: Sales – Expenses = Earnings It is a formula that applies to the largest oil company, the smallest lemonade stand, and everything in between. The complex nature of earnings becomes apparent when you try to analyze them. Why are some companies profitable while others are not? And why do some firms, profitable for decades, suddenly turn stagnant? The key elements of profitability are:

SL3151Ch15Frame Page 693 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

• • • • • •

693

Demand for the firm’s products or services The severity of competition The effectiveness of cost control Employee motivation Management knowledge, experience, and judgment To a larger extent than we usually acknowledge — luck

Of these, demand for the company’s products or services is not only the greatest influence on earnings, but whatever is in second place (probably luck) is way behind it. Demand is the condition of being sought after, and it is made manifest in business by the willingness, coupled with the ability, of customers to buy what you are selling. This is the reason this section on accounting and finance is included in a discussion of six sigma/DFSS. Unless we internalize the concept of demand in relation to the functional requirements that the customer is ever seeking, we are not going to be profitable. Demand is a fickle friend; it comes often without warning and disappears the same way. It is not something we have a great deal of control over. Rather, it is a condition that arises within our customers and is difficult to predict — even by the customers themselves — unless we spend some time and investigate their needs, wants, and expectations. Businesses can stimulate demand a little with advertising and other marketing efforts, but by and large it is created by the customer in a way that we do not completely understand. All of this leads us to the basic business risk — the reason companies are deserving of making a profit. When you start a business, you have to create a product, gather the people and materials needed to make it, set up a distribution system, and advertise the product, all before you have the first evidence that people will buy the product from you. Earnings Ratios There are a number of earnings or profitability ratios in current use. Just about all of them use numbers from both the balance sheet and the income statement. Some are better than others, and we will touch on the most popular ones. Le ROI The king of the earnings ratios is often referred to as ROI — Return on Investment. That is the ratio of profit to equity. But in recent years, the interest in these measurements has multiplied so that there is now a whole family of ROI ratios, and ROI has become a generic term for several different kinds of measures. Most earnings ratios are called Return on Something, and the method of calculation is fairly standard. “Return on” indicates that some profit figure is in the numerator, and the “something” is the denominator of the fraction. The result usually falls into a range between 0 and .5 and is normally expressed as a percentage figure. Many of the return ratios come in two colors, profit before tax and profit after tax (PAT). Both types are commonplace, but the latter is about twice the size of the former, so you have to pay attention to what you are looking at. I will always be

SL3151Ch15Frame Page 694 Thursday, September 12, 2002 8:00 PM

694

Six Sigma and Beyond

referring to PAT unless I say otherwise. Here is a brief description of the three most popular Return ratios, all three of which are calculated by most companies. ROE: Return on Equity

Profit Equity ROE is the last word in profitability ratios. When the smoke and mirrors of this “special factor” and that “extra adjustment” are put aside, this is the measurement that tells you whether you really have a business or not. ROA: Return on Assets

Profit Assets You will remember that Assets = Debt + Equity, so ROA is like ROE except that the denominator is bigger and the percentage return is therefore smaller than ROE. This ratio is popular among larger companies for measuring the performance of subsidiaries and divisions. ROE would be a better measure, but when companies cannot determine a true equity figure for a division, this works pretty well. ROS: Return on Sales

Profit Sales ROS tells how many cents out of each sales dollar go into the owners’ pockets. Nearly every company calculates this ratio, but it is not really very useful because there is no standard to gauge it by, as there is with ROE. For example, Company A may cater to the carriage trade and have limited sales but a high ROS. Company B may be a mass merchandiser with a low ROS, yet B could have a higher ROE than A because profit is the product of ROS and the volume of sales. The average ROS for all companies in the US is between 5 and 6%, but the number varies widely from company to company, even in the same industry. Other Return Ratios

There are dozens of other return ratios in active use, but their definitions are not well settled. Here are a few of the more common 3+ letter jobs, but even with these, definitions vary among users. RONCE: Return on Net Capital Employed — The denominator, net capital employed, usually refers to total debt plus equity minus non-interest-bearing debt such as accounts payable. But this is not always what it means, so if it is important for you to know the precise meaning, ask the user to define the ratio.

SL3151Ch15Frame Page 695 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

695

ROAM: Return on Assets Managed — Often used in management bonus plans; similar to RONA, which follows. RONA: Return on Net Assets — Here the denominator starts out with total assets and then certain ones are deducted. Often the assets taken out are those not directly related to the running of the business — for example, investments. ROGA: Return on Gross Assets — This is likely to be the same as ROA, Return on Assets.

FINANCIAL RATING SYSTEMS We first encounter rating systems on our grade school report cards — and spend the rest of our lives complaining about their inadequacies. Financial rating systems, too, have their limitations, but they can be an effective means of getting a quick measure of financial strength. Financial rating companies have been around for most of this century. Even now, in this age of rigid accounting standards and computer assisted analyses, businesspeople continue to rely heavily on rating services rather than doing the rating themselves. Apparently, the virtues of the rating companies — thoroughness, consistency, and conservatism — outweigh their draw-backs of old information, obscure criteria, and cost in the minds of many users. Rating services are best employed where the user is dealing with a large number of companies, and the investment or credit risk the user is taking with any one of them is small. But if the risks are concentrated or a lot of money is at stake with particular companies, the user should learn to make the analysis personally or at least be able to confirm the judgment of the rating companies. Research studies have shown that the rating firms are often slow to react when a company’s financial strength is on a downslide.

BOND RATING COMPANIES The generally wealthy institutions and investors who buy most bonds make extensive use of rating services; over 4000 issues are rated on a regular basis. Stripping it to its essentials, the analytical process is one of comparing: 1. Total debts against expected future profits, which are the primary source of interest and principal payments 2. Total debts against total assets, the liquidation of which is a secondary, albeit dire, source of repayment Moody’s et al. Moody’s and Standard and Poor’s are the best known of the bond rating companies. Neither of these firms reveals exactly how they arrive at their ratings, but the following criteria figure prominently in their classifications. 1. Financial leverage; the lower the leverage, the better the rating. 2. Profitability or rather, the avoidance of losses.

SL3151Ch15Frame Page 696 Thursday, September 12, 2002 8:00 PM

696

Six Sigma and Beyond

3. Steadiness of profits, the importance of which has led many firms to attempt “managing” or smoothing out their year-to-year earnings. 4. Total revenues or extent of market share; being an industry leader makes you a stronger competitor or is it the other way around? Both Moody’s and Standard and Poor’s use letter ratings beginning with triple A. Here are samples of the definitions they give their classifications. Moody’s Aaa Bonds carry the smallest degree of investment risk and are generally referred to as “gilt edge.” Interest payments are protected by a large or an exceptionally stable margin and principal is secure. Caa Bonds are of poor standing. Such issues may be in default or there may be present elements of danger with respect to principal or interest. Standard and Poor’s AAA is the highest rating and indicates an extremely strong capacity to pay principal and interest. Bonds rated BB, B, CCC, and CC are regarded, on balance, as predominantly speculative with respect to the issuer’s capacity to pay interest and repay principal in accordance with the terms of the obligation. As one can see, neither company gets very specific about the meaning of the ratings, and no attempt is made to predict the future of the subject firm. Rather, the ratings convey a “feeling” about it. That feeling represents the risk side of the investment equation: Risk = Return The return, on the other hand, is represented by a specific number — the bond’s yield. Now, if the bond’s rating could also be expressed as a specific number — the percentage probability of loss — investment decisions would be greatly simplified. The rating agencies maintain a rigid independence from the companies they analyze. Rigidity is necessary because of the millions of dollars of higher or lower interest costs that often ride on the change of a rating (that is, in subsequent bond issues, not the ones outstanding). Now and then you will see a company emit an outraged howl at being downgraded.

RATINGS

ON

COMMON STOCKS

In addition to rating bonds, these firms rate preferred stock and commercial paper on similar scales. As to common stocks, there are hundreds of companies that dispense investment advice. While a few confine themselves to issuing data, most have some subjective or objective method of selecting stocks for purchase or sale. None give full descriptions of the selection process, of course, because their advice, and the mystique that surrounds it, are all that they have to sell. Among the major services there are two that have established rating systems for use in selection of common stocks: Standard and Poor’s and The Value Line Investment Survey. Standard and Poor’s publishes a monthly stock guide that is crammed with

SL3151Ch15Frame Page 697 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

697

financial data on about 5000 common and preferred stocks. Most of the stock issues are rated on an eight-level scale running from A+ (highest) to D (in reorganization). The S&P Rating Method The rating formula is based on a computerized scoring system that traces the trends of earnings and dividends over the previous ten years. The basic scores are then adjusted for growth, stability, and cycles; final scores are measured against a matrix of a large sample of stocks. The Standard and Poor’s (S&P) rating serves well enough as a measure of a company’s past performance; but as it ignores the condition of the balance sheet and future earnings estimates, it is only a starter in an investment analysis. Considering the price of the analyses, however, about ten cents a gross, there hardly seems to be grounds for complaint. The Value Line Method The Value Line Investment Survey tracks over 1700 companies on a regular basis. Using unpublished equations, it rates each company for “safety” and investment “timeliness.” Both factors are ranked on a scale of 1 (highest) to 5 (lowest), the rankings being relative to all 1700 stocks, not to some absolute standard. The Value Line safety ranking is based on such factors as leverage, fixed charge coverage (the number of times over that profits could pay the annual interest expense), liquidity, and the riskiness of that type of business. The timeliness factor is a comparison of a stock’s price trend against its expected earnings. A company may have a terrific near-term profit outlook, but if the market price of its stock is hovering somewhere in the stratosphere, it may not be a “timely” buy. Good Ole Ben Graham Among the published formulas for investing in stocks, one of the most famous and enduring is based on the “intrinsic value” theory of the late Benjamin Graham. Professor Graham, a pioneer in “fundamental” analysis, did most of his research in pre-computer days when computations were made on those clunky mechanical calculators and laboriously recorded by hand. Graham’s notion was that stocks at any given time are likely to be undervalued or overvalued; that is, many investors are buying and selling shares for reasons other than their fundamental value. The smart investor will appraise that value (relative to the price) and snap up the undervalued bargains. Among the criteria Graham proposed were the following: 1. For financial soundness, a ratio of total liabilities to current assets of no more than 60% 2. A ratio of equity to total debt of at least 1007; Mr Graham often said, “a company should not owe more than it is worth.” 3. One buy signal: a price less than 2/3 of the “net-current-asset” value, defined as Current assets – (Total liabilities + Preferred stock)/Number of shares outstanding.

SL3151Ch15Frame Page 698 Thursday, September 12, 2002 8:00 PM

698

Six Sigma and Beyond

4. Another price criterion: the earnings to price ratio (the reciprocal of the P/E ratio) should be at least double the average Triple-A bond yield for industrials. So if the bonds were averaging 10%, the E/P ratio should be at least 207, which means a P/E ratio of 5 or less. Special warning: Before you rush out and sell the farm to try one of these stock market systems, be aware that no system has ever had a consistent, long-term run of higher than average profits. Even Ben Graham’s common-sense method suffers from the perverse nature of the market: No matter how adept you become at finding “undervalued” stocks, the only way you can make money on them is if the rest of the investors come to the same realization soon after you have bought some shares. Sometimes they never do, and you could be left sitting on top of an undiscovered gold mine until it is, well, too late.

COMMERCIAL CREDIT RATINGS Dun & Bradstreet Dun & Bradstreet is the granddaddy of commercial credit rating firms. While its main business is furnishing information about a company’s finances and paying habits, it also assigns two ratings to those firms about which it has sufficient information. The first is a 15-level scale of a firm’s “estimated financial strength” or equity. The highest classification is 5A for companies with a net worth of $50 million or more; midway down the list is BB, $200,000 to $299,999; the lowest rank is HH, covering an equity less than $5,000. D&B will only issue a rating when it can obtain an equity figure, usually from the subject company, and has no reason to think it is inaccurate. The second rating is a “composite credit appraisal,” which is derived by an unpublished formula, presumably taking into account a company’s liquidity, leverage, and profitability, as well as paying habits and any adverse events such as tax liens. This appraisal has four grades: Grade 1 2 3 4

Meaning High Good Fair Limited

Each credit appraisal is done in conjunction with the financial strength rating, so that the “1” in a rating of EE1 does not reflect the same standards as the “1” in 4A1. While there are smaller credit agencies that also issue ratings, D&B’s system stands virtually alone. So extensively is it used by vendors granting trade credit that for thousands of firms a good D&B rating is all that is necessary to establish an open account.

SL3151Ch15Frame Page 699 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

699

TABLE 15.3 The Z Score Ratio Working Capital/Total Assets Retained Earnings/Total Assets EBIT/Total Assets Shares Market Value/Total Debt Sales/Total Assets

Factor .012 .014 .033 .006 .999

Other Systems Sometimes, a do-it-yourself approach is used in evaluation. That means the analyst uses multiple statistical tools to create a scoring system. Two major approaches are known for this endeavor. The first type consists of systems that base their ratings on a statistical analysis. The methodology is to gather a bunch of financial statements, several years’ worth, from companies that have gone bankrupt, and a bunch from otherwise similar companies that have remained afloat. Various ratios are then calculated in an effort to find those indicators that best distinguish the one group from the other. Perhaps the best-known work in this field has been done by Professor Edward I. Altman of New York University, who in 1968 developed a statistical model for the prediction of corporate bankruptcy. Altman’s formula, known as the “Z score,” combined five ratios selected by an advanced statistical method called discriminant analysis, which searches out the best combination of ratios rather than the single best ratio for predicting the future. The five ratios, shown in Table 15.3, are multiplied by conforming factors, and the products are added together to get the score. The critical Z Score was 1.81, according to the study, which found that all of the firms in its group with a score that low had gone bankrupt. Moreover, the model correctly classified 959 of the total sample one year prior to bankruptcy. (A note for this analysis: This test did not hold its effectiveness for the long term as an increasing proportion of its predictions became false. In 1977, Professor Altman revised the Z score but again it has not caught on.)

COMPANY AND PRODUCT LIFE CYCLE Businesses, like governments, like all institutions created by human beings, are mortal. In the course of time they are created, enjoy their season, and then vanish. Their evolution can be traced through four stages — tryout, growth, maturity, and decline. Products, too, have a limited life, and if a demand for them develops, they can be expected to obey a somewhat similar pattern, that depicted in Figure 15.1. Although few companies or products will imitate this design exactly, the concept is useful in forecasting sales and cash flow.

SL3151Ch15Frame Page 700 Thursday, September 12, 2002 8:00 PM

700

Six Sigma and Beyond

Tryout

Growth

Maturity

Decline Time

FIGURE 15.1 Life cycle of a typical company or product.

The tryout period is one of experimentation, finding the product, the price, the method of distribution, the niche that will create customers. If and when the growth stage develops, a heavy investment in promotion and production is needed. During this period, which may last a decade or two, it is usual for more cash to be spent than received, even though the operation is highly profitable. With maturity the cash flow turns positive as sales level off. The last stage is the least predictable, some companies going out with a bang, some with a whimper, others merging themselves quietly into the operations of a more viable firm.

CASH FLOW Cash flow has been defined as: Profit + Depreciation In recent years a new definition has been taking shape: Profit + Depreciation + Deferred Taxes Cash flow is intended to represent “discretionary funds” that are over and above what is needed to continue running the business, and may, therefore, be used to expand the company, pay off loans, pay extra dividends, and so on. When it was first conceived, the idea of adding profit and depreciation to get cash flow found overnight acceptance among business executives. If profit was vanilla ice cream, cash flow was a chocolate sundae. But it also produced an unwanted side effect — “the non-cash illusion.” Cash flow is a popular term with business managers. It is a phrase that is vague enough to make you sound like you know what you are talking about even when you do not. It is also useful in cases where you do know what you are talking about but do not want to talk about it. As when a supplier calls you about an overdue bill. Which would you rather say?

SL3151Ch15Frame Page 701 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

701

“We do not have the money right now.”

or “We are experiencing a temporary cash flow problem.”

The latter statement conveys your understanding of mysterious economic forces and implies that the solution to the problem is just around the corner. Unfortunately, such a fine phrase as “cash flow” has been used in so many different ways you have to verify its meaning each time you hear it. Often, as in the example above, it means the same as “cash”; but sometimes it is the amount of cash flowing into the business each month or year; at other times it is the difference between the inflow of cash from sales and the outflow for expenses; and to those who enjoy elevating obscurity to its highest plane it is “the funds available as working capital and for expansion.” Most of the time when professionals, especially financial professionals, speak of cash flow they are talking about the specific dollar amount derived by adding depreciation back to profit. When we analyze cash flow we are asking what activities brought money into the business and what activities caused it to flow out. In the simplest terms, where did the cash come from, where did it go? Most of the cash coming into a business is the proceeds of sales; most of it going out is to pay expenses. So we can start our cash flow analysis with the difference between the two, which is profit. Profit = Sales – Expenses Included in the expenses is depreciation and maybe amortization, but these are non-cash expenses; that is, there is no money paid out for this expense because it was all paid out at the time the asset was bought. The concept of cash flow is lame in one respect. It fails to recognize the need to replenish fixed assets. Plants and equipment must be replenished, just as inventory is. To say that funds from depreciation do not have to be spent on new fixed assets is as deceptive as saying that cash received from a sale does not have to be spent to buy new merchandise. (If you need any further convincing, look at some published annual reports — the statistical section where you often find figures for depreciation and new capital investment going back five or ten years. Count the number of years that the value of new equipment exceeded the depreciation charges; chances are it will be at least nine times out of ten.) In other words, companies are not only using all of the depreciation money to buy new fixed assets, they need a good deal more besides. A better formula for calculating cash flow would be: Profit + Depreciation – New Fixed Assets Even if a company is not growing, chances are that inflation will push replacement costs higher than depreciation rates.

A FINAL THOUGHT

ABOUT

CASH FLOW

Because of the non-cash illusion, the concept of cash flow has little relevance to day-to-day management. It is useful in calculating the return on proposed capital

SL3151Ch15Frame Page 702 Thursday, September 12, 2002 8:00 PM

702

Six Sigma and Beyond

investments, as we will see later, but for the management of cash and cash planning it is not. Those activities are best managed with detailed budgets and forecasts. Moreover, the mystique of cash flow has been known to replace common sense, as in the airline industry where enormous depreciation charges often mask treacherous losses; likewise in some tax shelter schemes where non-cash charges are used to reduce taxes and thus appear to actually generate money. In the last analysis all firms, all tax schemes, must be profitable to be successful. Profits are the true test of any investment, and to the extent that cash flow confuses this ultimate reckoning it does us a disservice. It is apparent that a company can lose money and still have a positive cash flow. What is not so clear is that a firm can have a positive cash flow and still go broke — a common hazard for rapidly growing companies. The term “working capital,” like the term “cash flow,” is frequently heard in the daily chatter about business finance. It, too, suffers from liberties taken with its definition and usage. Most often, and especially when financial people are talking, working capital means the specific dollar amount derived from the formula Current assets – Current liabilities However, it has also been used to mean “cash,” or “cash + receivables,” or “current assets,” or “funds.” When it comes to applying the idea of working capital in some useful business way, we encounter two nearly fatal flaws. 1. Working capital is a concept that has no existence in the real world. You cannot hold working capital in your hand or put it in your pocket. Nor can you actually offset current liabilities with current assets. Nearly all current liabilities can only be satisfied with cash. 2. While businesspeople are fond of calculating working capital, no one has yet come up with a rule stating how much of it a company should have. It seems reasonable enough that the more sales a business has, the more working capital there should be also. But we cannot seem to pin it down to an actual standard. Working capital, therefore, is a measurement without much meaning. In business there is only one excuse for an expense: it will help to produce revenue. Some expenses, however, have nothing to do with producing revenue — the entire accounting department, for example — but they are necessary nevertheless. Others, such as income tax and vacation pay, have only a roundabout effect on your sales but are also unavoidable. Some pay our debts to society, such as the expense of unemployment insurance, or make us good neighbors, such as a little landscaping, while the sole purpose of some expenses is to reduce overall expense, that is, increase productivity. Unlike revenues, which are the result of a customer taking action, expenses result when you take action. They are largely controllable and therefore a direct reflection of your management ability. It may be hard to measure the value of what you do, but the cost of your doing it is right there in the printout for all to see.

SL3151Ch15Frame Page 703 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

A HANDY GUIDE

TO

703

COST TERMS

Actual — Actual costs are distinguished from standard costs; the latter are estimates used for convenience, and an adjustment must be made to the actual costs at least yearly. Alternative — The costs of optional solutions; used in “what-if” analyses. Controllable — Costs for which some manager can be held responsible. Cost of Sales — Also called “cost of goods sold”; the cost of making or buying the products a business sells. In a manufacturing firm it comprises direct labor, materials, and manufacturing overhead. Differential — The difference in the costs of two or more optional activities. Direct — Costs that can be laid solely to a particular activity. In manufacturing, the wages of the workers who make the products and the cost of the materials used are direct costs; they are often referred to as “direct labor” and “direct materials.” Discretionary — More or less unnecessary but desirable outlays, such as the office Christmas party or management seminars. Estimated — Predetermined by an informed guess. Extraordinary — Expenses due to abnormal events, such as an earthquake. Costs in this category should be not only unexpected, but rare. Fixed — Costs that remain the same despite changes in sales or some other output. Examples are lease payments on property and depreciation on equipment. Compare to variable costs. As used here the fixedness is a matter of degree; almost every cost is affected somewhat by the volume of sales. Historical — The original cost of an asset. Imputed — The imagined or estimated cost of a sacrifice; not a cash outlay but the giving up of something you could have had; a cost often recognized in the decision process but not recorded on the books. When a company has accounts receivable, for example, there is an imputed cost of the interest it could be earning on the funds tied up in receivables. Incremental — A cost that will be added or eliminated if some change is made. Similar to “differential cost.” Indirect — A general or overhead cost that is allocated to a product or department on the theory that the receiver shares in the benefit of the thing, and, besides, somebody has to pay for it. Joint — Also called “common cost”; A cost shared by two or more products or departments, as for example the expense of a company lunchroom. Noncontrollable — Costs that are prerequisites to doing business, such as a city license or smog control equipment. These costs are often allocated to a department, but there is no point in holding the manager accountable for them. Opportunity — A theoretical cost of not using an asset in one way because you are using it in another. For example, the opportunity cost of a companyowned headquarters building is the money the company could get by renting it to others.

SL3151Ch15Frame Page 704 Thursday, September 12, 2002 8:00 PM

704

Six Sigma and Beyond

Out of Pocket — Expenses requiring a cash outlay, as opposed to the expense of using facilities you already own. When you use your car for company business, the cost of gas, tolls, and parking are out of pocket expenses, but not the depreciation on the car. Period — Costs related to a time period rather than an amount of output or activity. Examples are rent and the controller’s salary. Prepaid — Expenses paid before, rather than as, they are used. A prepaid expense, such as a year’s insurance premium, is really a current asset that gradually converts to an expense with passing time. Prime — In manufacturing, direct labor and material costs. Product — Costs related to output or the amount of activity, as opposed to period costs. Production — A term used by the oil and gas industry in referring to the cost of operating a well. Replacement — An estimate of the current cost of an asset as contrasted with its historical cost. It is often used in estimating the “true” cost of the current year’s depreciation. Standard — The estimated average or budgeted cost of making a product. When a product is finished, it is often more convenient to record its value at a standard rather than its actual cost. At least once a year the total actual costs are compared with the total standard costs recorded, and the latter is adjusted to the real figure. Standard costs must be changed from time to time if labor and material costs change. Sunk — A cost already incurred that cannot be undone or readily put to some other use. Variable — A cost closely tied to the level of output or activity. Most of these costs vary directly with sales. Classifying costs between variable and fixed is necessary in order to calculate a breakeven point.

USEFUL CONCEPTS FOR FINANCIAL DECISIONS THE MODIFIED

DUPONT

FORMULA

The duPont system of financial analysis combines profit margin and asset turnover to produce the return on assets. The modified version brings financial leverage into the equation to produce return on equity as well. The formulas are: 1. Asset Turnover × Return on Sales = Return on Assets Sales/Assets × Profit/Sales = Profit/Assets 2. Return on Assets × Financial Leverage = Return on Equity Profit/Assets × Assets/(Equity ×) = Profit/Equity A visual approach to duPont’s concept is shown in Figure 15.2.

SL3151Ch15Frame Page 705 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

705

Return on total assets Return on sales

Asset turnover

Multiply by

Divided into

Divided into Net income

Sales

Total assets Plus

Subtracted from Total costs

Sales

Sales

Cost of goods sold

Operating expense

Depreciation

Interest

Taxes

Less other income

Fixed assets

Current assets

Inventories

Cash

Accounts receivable

Marketable securities

FIGURE 15.2 A pictorial approach of DuPont’s formula.

BREAKEVEN ANALYSIS The breakeven point for a business is that volume of sales at which the revenues equal the expenses. Above that point lie glory and profit; below lie infamy and loss. At least that is the theory. In real life, it is very difficult to calculate a breakeven point because the expenses of most businesses do not fit comfortably into just a fixed or variable category. Breakeven analysis can be done visually using a graph like the one in Figure 15.3, or mathematically. Profit = Sales – Fixed Costs – Variable Costs If Fixed Costs = $12,000 and Variable Costs = 40% of Sales, then Profit = Sales – $12,000 – .4*Sales or Profit = .6*Sales – $12,000 At breakeven, profit will be 0, therefore 0 = .6*Sales – 12,000 and Sales = $12,000/.6 = $20,000 at that point.

SL3151Ch15Frame Page 706 Thursday, September 12, 2002 8:00 PM

706

Six Sigma and Beyond

Revenue

$000s

Expense Profit 20 Fixed costs 10 Loss 5 0

10

20

30

40

-------------------------------------------------------------------------------→ Units FIGURE 15.3 Breakeven analysis.

CONTRIBUTION MARGIN ANALYSIS We have seen that a firm will break even when its total sales exactly equal the sum of its variable and fixed costs. Beyond breakeven only the variable costs need be paid, the fixed costs having been taken care of for the year. The difference between the sales price of the company’s products and the variable costs is called the contribution margin. In our example where variable costs equaled 50% of sales, the contribution margin was the other 50%. Knowing the contribution margin gives you another way of calculating the breakeven point:

Break - even sales =

Fixed Costs Contribution Margin

Using the figures in our illustrative example

Break - even sales =

$10,000 = $20, 000 .50

Most firms have a variety of products or services that contribute to profits at different rates. In comparing margins, you must also take into account the proportionate fixed costs associated with each. It is possible that a product with a smaller contribution margin would be the more profitable because the other product bears enormous fixed costs.

SL3151Ch15Frame Page 707 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

707

PRICE–VOLUME VARIANCE ANALYSIS A price–volume analysis of profit plan variances often helps management zero in on problems or quickly exploit market advantages. Here is a simple example of such an analysis. Suppose a firm planned to sell 1000 units of Product A at $20 each but in fact 1100 units were sold at a price of $21 each. The planned revenue was 1000 × $20 = $20,000 Actual revenue was 1100 × $21 = $23,100 and the revenue variance = $3100. That variance can be broken down as follows: Effect of price change only: $1 × 1000 = $1000 Effect of quantity change only: $20 × 100 = $2000 Effect of both price and quantity changes: $1 × 100 = $100 Total effect = $3,100

INVENTORY’S EOQ MODEL The Economic Order Quantity model is designed to minimize the total cost of ordering and carrying inventory items. Here is the formula: EOQ = (2*Q*P/C).5 where Q = quantity needed for the period; P = the cost of placing one order; and C = the cost of carrying one unit for one period. EXAMPLE Standard Office Furniture sells 1800 “B” desks more or less evenly over 12 months. The cost of placing and receiving an order from the manufacturer is $45. Standard’s annual carrying costs are 20% of the inventory value. The “B” wholesales for $75, so the annual carrying cost per desk is

.20 × $75 = $15 The economic order quantity can then be calculated using the model:

EOQ = (2*45*1800/15).5 = 104 desks We can also calculate Standard’s optimal inventory cycle for these desks:

[365 × 104]/1800 = 21 days

SL3151Ch15Frame Page 708 Thursday, September 12, 2002 8:00 PM

708

Six Sigma and Beyond

RETURN

ON INVESTMENT

ANALYSIS

EXAMPLE: ROI CONCEPTS This is an example where a company has to decide between two different manufacturing machines it wants to purchase. The costs and benefits of each are set out below. MACHINE A End of Year → Revenues Direct cost, mtl, labor, etc Operating exp, Selling, G&A Depn, Straight Ln Profit Before Tax Income Tax (50%) Net Income5,000 Cash Flow Investment Machine B $50,000 End of Year → Revenues Direct cost, mtl, labor, etc Operating exp, Selling, G&A Depn, Straight Ln Profit Before Tax Income Tax (50%) Net Income5,000 Cash Flow Investment

$50,000 1 30,000 5,000 5,000 10,000 10,000 5,000 5,000 15,000 40,000

2 30,000 5,000 5,000 10,000 10,000 5,000 5,000 15,000 30,000

3 30,000 5,000 5,000 10,000 10,000 5,000 5,000 15,000 20,000

4 30,000 5,000 5,000 10,000 10,000 5,000 5,000 15,000 10,000

5 30,000 5,000 5,000 10,000 10,000 5,000 5,000 15,000 0

1 45,000 7,500 5,000 12,500 20,000 10,000 10,000 22,500 37,500

2 40,000 7,500 5,000 12,500 15,000 7,500 7,500 20,000 25,000

3 32,000 5,000 5,000 12,500 10,000 5,000 5,000 17,500 12,500

4 25,000 2,500 5,000 12,500 5,000 2,500 2,500 15,000 0

5

Payback Method: Payback answers the question, how long will it take us to recover our original investment? Year → MACHINE A Balance to Recover Cash Flow Cumulative Years MACHINE B Balance to Recover Cash Flow Cumulative Years

1

2

3

4

5

50,000 15,000 1.00

35,000 15,000 2.00

20,000 15,000 3.00

5,000 15,000 3.33

0 15,000 3.33

50,000 22,500 1.00

27,500 20,000 2.00

7,500 17,500 2.43

0 15,000 2.43

Average Rate of Return: Average rate of return is our old friend ROE, Profit/Equity (or in this case, Profit/Investment), except we call for the average return over the period covered.

SL3151Ch15Frame Page 709 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting End of Year → MACHINE A Profit Average Profit:

709

1 5,000 5000

2

3

4

5

5,000

5,000

5,000

5,000

Beginning Investment = $50,000 Ending Investment = 0 Average Investment = $25,000 Average Rate of Return = $5000/$25,000 = 20% MACHINE B Profit 10,000 7,500 5,000 2,500 Average Profit: 6250 Beginning Investment = $50,000 Ending Investment = 0 Average Investment = $25,000 Average Rate of Return = $6250/$25,000 = 25%

Net Present Value (NPV) NPV equals the cash receipts from an investment minus the cash outlays, all discounted at an acceptable rate, sometimes called the hurdle rate. The formula is n

NPV =

∑ (1 + r ) CFt

t

t =0

where n = the number of periods; t = the time period; r = the per period cost of capital; and CFt = the cash flow in time period t. Internal Rate of Return (IRR) IRR is at present the “truest” rate of return we know how to calculate. Technically, it is the “hurdle” or discount rate that produces an NPV equal to zero. The formula is n

NPV = 0 =

∑ (1 + r ) CFt

t

t =0

A special caution is needed here. The IRR calculation can turn awkward when there is more than one sign change in the cash flow stream. You may get more than one answer for the same series of payments. One way around the problem is to do a modified IRR in which you calculate the present value of all the outflows (negatives) using, say, the company’s average interest rate on loans; then compute the IRR using the single outflow figure (CFCM). The Financial Management Rate of Return (FMRR) developed by Findlay and Messner in 1973 goes one step further. It starts by calculating the present value of

SL3151Ch15Frame Page 710 Thursday, September 12, 2002 8:00 PM

710

Six Sigma and Beyond

all cash outlays, as does the modified IRR, and then calculates a future value for the positive cash flows (inflows). The rate for this future value calculation is the expected rate at which the inflows will be employed.

PROFIT PLANNING The most critical element of all in financial planning is the revenue, or sales, forecast. It is the basis of the cost and profit forecasts, and the key element in planning a firm’s people, money, and material needs. At the same time, the revenue or sales forecast is the most difficult to make of all forecasts. Anyone who can project sales a year ahead and come in less than 10% off the mark is doing pretty well.

THE NATURE

OF

SALES FORECASTING

Sales are mostly a function of demand for the product, and demand is largely out of the control of the seller. Numerous factors influence a company’s sales; most of them cannot be controlled, and many of them are not even known. But the overall effect of these factors usually changes rather slowly over time. For that reason a naive projection of sales always merits consideration in the forecasting process. A naive forecast is one that simply extrapolates past figures and trends into the future. However, most companies use a goals down–plans up approach to sales forecasting. Top management defines the ballpark by specifying what is in their opinion an achievable sales goal; it is then up to the sales staff to submit detailed plans on how that goal will be met. Another form of forecasting is the sales goal form. This form or template helps top management set a sales goal. It starts with the latest 12 months’ actual sales and a tentative goal for the coming year. The difference between the two will have several causes. Here are some causes for a change in sales volume and a typical annual impact of each cause. Inflation — Up 2 to 12 percent Demand for the product — Up or down 0–10% State of the economy — Up or down 0–10% New products — Up 0–10% The sales goal form also aids in the forecasting of gross profit, operating profit, net income, and earnings per share. In addition, it permits what-if analyses, showing the effect on net income of a change in the sales or cost inputs. The Plans Up Form This format is meant as an aid to individual salespeople who forecast the revenues in their own territories. The cumulative totals of these estimates must eventually be reconciled to the sales goal figures issued by top management.

SL3151Ch15Frame Page 711 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

711

Statistical Analysis A few things in business lend themselves to statistical projections. Chief among them is revenues or sales forecasting. Statistics are tricky. They can describe with precision the behavior of two or more business components, but it is left to you to decide if the activities are related, unrelated, or coincident, and in the case of the former, which is the cause and which the effect. A statistical forecast of sales, on the other hand, is useful mainly as an anchor for the subjective guesses of salespeople and top management, or as a starting point in developing judicious estimates. While the end purpose of a sales forecast is a projection of profits, direct statistical analyses of past profits are normally useless, if not worse. Compound Growth Rates Sales patterns usually trace a gentle curve rather than a straight line. All products, and for that matter, all companies and the people who run them, move through life cycles that can often be represented by a lazy S curve. Its components are tryout, growth, maturity and decline — see Figure 15.1. The compound growth rates over short segments of the life cycle can be useful in forecasting 12 months or so ahead. The rates will vary depending on how far back you go, and it is up to you to decide which rate is best to project. While you are working that out, remember that the sales trend is a curve that changes direction now and then. It is also up to you to figure out if such a change is about to occur. Regression Analysis Regression analysis is a mathematical procedure that figuratively plots past sales on a graph and then draws a line through the middle of the points; extending the line into the future gives the forecast. The results will vary depending on how much past data you use, and the whole process is meaningful only if the past plot points form a linear pattern. Even then, we must be mindful that unusual past events can skew the pattern, and that nothing in life really travels in a straight line. Revenues and Costs Once revenues have been forecast, the task of projecting costs is largely routine. Most costs, even those we consider fixed, relate to sales or revenues. When sales rise, costs will soon follow. When sales go down, however, costs tend to follow more slowly due to a natural human reluctance to do without something we once had and a natural hope that conditions will soon be better again. The largest expense in most businesses is the cost of sales, sometimes called the cost of goods sold, and it can usually be determined with fair precision. The task of manufacturing a product is well defined, as are the materials and labor that lend it value. Departmental Budgets Beyond the estimates for the direct costs of manufacturing or buying products for resale lie the budgets for administrative expense, overhead, marketing, R&D, and so on. We speak of these as budgets rather than forecasts because we have some control over them.

SL3151Ch15Frame Page 712 Thursday, September 12, 2002 8:00 PM

712

Six Sigma and Beyond

The computer is a helpful tool in the process because such budgets are negotiated rather than merely extrapolated from revenues, planned rather than merely projected. There are often several versions prepared before the final plan is hammered out. A good deal of posturing and gamesmanship goes into the process, and a not untypical scenario finds the department manager padding his or her budget secure in the knowledge that top management will cut it, and top management cutting it because they know it is always padded a little. How to Budget The easiest way to budget is to increase last year’s budget by some percentage that will allow a comfortable salary raise for everyone, plus a few more bodies to ease the workload, and some extra gadgets and trips to make it more fun. A more businesslike method is to adjust last year’s budget to the expected level of next year’s principal activities of the department. Every department has its tasks, and if the output or the activities can be quantified, this will furnish a standard for setting the new budget. The best lever for controlling department expense is the number of people employed. Employees not only cost salaries, fringe benefits, and taxes, but desks, computers, food service, and parking spaces, too. Zero-Growth Budgeting This is not to be confused with zero-based budgeting, which deservedly died a quiet death a few years back. Zero-growth budgeting is a term for a plan that seeks to hold expenses at the current level while revenues grow. If it works it will obviously mean more profit. Of the two ways to get rich in life, the more excusable is to increase your productivity — work harder or faster or smarter so that the value of your output goes up. To the department manager, zero-growth budgeting says, “Look, we expect sales to rise 10% this year but we would like to handle the increased business with the same budget as last year. What is more, we think the better people should get nice raises, but out of the same pot as last year. That means you have to handle the work more efficiently, and perhaps not replace people so fast when they leave.” The idea is to challenge people to work smarter but not threaten them with the annihilation that was implicit in the zero-based budgeting concept. Most productivity gains are made inch by inch, and if too much is asked of people, they tend to give up.

SELECTED BIBLIOGRAPHY Ainworth, P. et al., Introduction to Accounting: An Integrating Approach, Irwin, Homewood, IL, 2001. Albright, S.C. et al., Managerial Statistics. Duxbury, Pacific Grove, CA, 2000. Baker, R.E., Lembke, V.C., and King, T.E., Advanced Financial Accounting, 5th ed., Irwin, Homewood, IL, 2002. Block, E., Chen, K. and Lin, T., Cost Management, Irwin, Homewood, IL, 2001. Brealey, R.A., Fundamentals of Corporate Finance, 3rd. ed., Irwin, Homewood, IL, 2001. Brigham, E.F. and Ehrhardt, M.E., Financial Management: Theory and Practice, 10th ed., The Dryden Press, New Rochelle, NY, 2002.

SL3151Ch15Frame Page 713 Thursday, September 12, 2002 8:00 PM

Fundamentals of Finance and Accounting

713

Brigham, E.F. and Houston, J.F., Fundamentals of Financial Management, Concise, 3rd. ed., The Dryden Press, New Rochelle, NY, 2002. Dauten, C.A. and Welshans, M.T., Principles of Finance, 3rd. ed., South-Western Publishing Co., New Rochelle, NY, 1970. Edmonds, T.P. et al., Fundamental Financial Accounting Concepts, Irwin, Homewood, IL, 2000. Moore, F.G., Manufacturing Management, 5th ed., Irwin, Homewood, IL, 1969. Pyle, W.W. and White, J.A., Fundamental Accounting Principles, 6th ed., Irwin, Homewood, IL, 1972. Weston, J.F. and Brigham, E.F., Essentials of Managerial Finance, 3rd ed., The Dryden Press, Hinsdale, IL, 1974.

SL3151Ch15Frame Page 714 Thursday, September 12, 2002 8:00 PM

SL3151Ch16Frame Page 715 Thursday, September 12, 2002 5:57 PM

16

Closing Thoughts about Design for Six Sigma (DFSS)

Design for six sigma (DFSS) is really a breakthrough strategy to improvement, as well as customer satisfaction. In the new millennium, it is the most advantageous way as well as an economical way to plan. Fundamentally, the process of DFSS is really a four-step approach. It recognizes the customer and progressively builds on the system concept for robustness in product or service development and finally testing as well as verifying the results against the design. Some of the essential tools used in DFSS are: Define • Customer understanding • Market research • Kano model • Organizational knowledge • Target setting Characterize • Concept selection • Pugh selection • Value analysis • System diagram • Structure matrix • Functional flow • Interface • QFD • TRIZ • Conjoint analysis • Robustness • Reliability checklist • Signal process flow diagrams • Axiomatic designs • P-diagram • Validation • Verification • Specifications Optimize • Parameter and tolerance design • Simulation 715

SL3151Ch16Frame Page 716 Thursday, September 12, 2002 5:57 PM

716

Six Sigma and Beyond

Define

Characterize

Optimize

Verify

Consumer understanding Quality history Understanding of customer requirements (functionality – CTS(Y)) Kano model

CTS’s to metrics ➠ (Y toy) Technical metrics ➠ (y tox) Concept selection Function structure CAE modeling P- diagram Noise management Reliability and robustness

Process capability Process flow diagram Gauge R&R control plans Parameter design Tolerance design Robustness assessment Statistical

Design verification process Key life testing Long term process capability Reliability and robustness Product performance over time

FIGURE 16.1 The DFSS model.

• • • • • • • • • Verify • • • • • •

Taguchi Statistical tolerancing QLF Design and process failure mode and effects analysis (FMEA) Robustness Reliability checklist Process capability Gauge R & R Control plan Assessment (validation and verification score cards) Design verification plan and report Robustness reliability Process capability Gauge R & R Control plan

The concept of DFSS may be translated into a model shown in Figure 16.1. This model not only identifies the components DCOV (define, characterize, optimize, verify), but it also identifies the key characteristics of each one of the stages. To understand and appreciate how and why this model works, one must understand the purpose and the deliverables of each stage in the model. So, let us give a summary of what the DFSS is all about.

SL3151Ch16Frame Page 717 Thursday, September 12, 2002 5:57 PM

Closing Thoughts about Design for Six Sigma (DFSS)

717

In the Define (D) stage, it is imperative to make sure that the customer is understood. The “spoken” and the “unspoken” requirements must be accounted for and then the definition of the CTS drivers takes place. It is very tempting to jump right away into the Yi without really knowing what the critical characteristics (or functionalities) are for the customer. Unless they are understood, establishing operating window(s) for these Ys will be fruitless. So the question then is, “How do we get to this point?” And the answer in general terms (the specific answer depends on the organization and its product or service) is that the inputs must be developed from a variety of sources including but not limited to the following — the order does not matter: • • • • • • • • • • • • • • •

Consumer understanding Kano model application Regulatory requirements Corporate requirements Quality/customer satisfaction history Functional, serviceability, expectations Understanding of integration targets process Brand profiler/DNA Benchmarking Quality Function Deployment (QFD) Product Design Specifications (PDS) Business strategy Competitive environment Market segmentation Technology assessment

Once these inputs have been identified, developed, and understood by the DFSS team, then the translation of these “functionalities” may be articulated into the Ys and thus the iteration process begins. How is this done? By making sure all the active individuals participate and have ownership of the project as well as technical knowledge. Specifically, in this stage the owners of the DFSS project will be looking to make correlated connections of what they expect and what they have found in their research. Thus, the search for a “transformation function” begins and the journey to improvement begins in a formal way. Some of the steps to identify the Ys are: • • • •

Define customer and product needs/requirements. Relate needs/requirements to customer satisfaction; benchmark. Prioritize needs/requirements to determine CTS Ys. Review and develop consensus.

Once the technical team has finished its review and come up with a consensus for “action,” the following deliverables are expected:

SL3151Ch16Frame Page 718 Thursday, September 12, 2002 5:57 PM

718

Six Sigma and Beyond

• • • • • •

Kano diagram Targets and ranges for CTS Y’s Y relationship to customer satisfaction Y relationship to customer satisfaction Benchmarked CTSs CTS scorecard

At this point, one of the most important steps must be completed before the DFSS team must go officially into the next step — characterize. This step is the evaluation process. A thorough question and answer session takes place with focus on what has transpired in this stage. It is important to ask questions such as: Are we sure that our CTS Ys are really associated with customer satisfaction? Did we review all attributes and functionalities? And so on. Typical tools for the basis of the evaluation are: • • • • • • •

Consumer insight Market research Quality history Kano analysis QFD Regression modeling Conjoint analysis

When everyone is satisfied and consensus has been reached, then the team officially moves into the characterize (C) stage. In this stage, all participants must make sure that the system is understood. As a result of this understanding, the team begins to formalize the concepts. The process for this development proceeds as follows: • Flow CTS Ys down to lower level y’s, e.g., Y = f(y1, y2,… yn). • Relate y’s to CTQ parameters (x’s and n’s), y = f(x1,…, xk, n1,…, nj) (x is the characteristic and n is the noise). • Characterize robustness opportunities (optimize characteristics in the presence of noise). Specifically, the inputs for this discussion are the: • • • •

Kano diagram CTS Ys, with targets and ranges Customer satisfaction scorecard Functional boundaries and interfaces from system design specification(s) and/or verification analysis • Existing hardware FMEA data Once these inputs have been identified, developed, and understood, then the formal decomposition of Y to y to y1 as well as the relationship of X to x to x1 and

SL3151Ch16Frame Page 719 Thursday, September 12, 2002 5:57 PM

Closing Thoughts about Design for Six Sigma (DFSS)

719

n’s to the Ys begins. How is this done? By making sure all the active individuals participate and all have ownership of the project as well as technical knowledge. Specifically, in this stage the owners of the DFSS project will be looking to make correlated connections of what they expect and what they have found in their research. Thus, the formal search for the “transformation function,” preferably the “ideal” function, gets underway. Some of the steps to identify both the decomposition of the Ys and its relationship to x are (order does not matter, since in most cases these items will be worked on simultaneously): • • • • • •

Identify functions associated with CTSs Identify control and noise factors Create function structure or other model for identified functions Select Ys that measure the intended function Create general or explicit transfer function Peer review

The deliverables of this activity are: • Function diagram(s) • Mapping of Y → functions → critical functions → y’s • P-diagram, including critical • Control factors, x’s, • Technical metrics, y’s, • Noise factors, n’s • Transfer function • Scorecard with target and range for y’s and x’s • Plan for optimization and verification (R&R checklist) At this point, one of the most important steps must be completed before the DFSS team must go officially into the next step — optimization. This step is the evaluation process. A thorough question and answer session takes place with focus on what has transpired in this stage. It is important to ask questions such as: Have all the y’s technical metrices been accounted for? Are all the CTQ x’s measurable and correlated to the Ys of the customer? Are all functionalities accounted for? And so on. Typical tools for the basis of the evaluation are: • • • • •

Function structures P-diagram Robustness/reliability checklist Modeling using design of experiments (DOE) TRIZ

When everyone is satisfied, then the team officially moves into the optimization (O) stage. In this stage, we make sure that the system is designed with robustness in mind, which means the focus is on

SL3151Ch16Frame Page 720 Thursday, September 12, 2002 5:57 PM

720

Six Sigma and Beyond

1. Minimizing product sensitivity to manufacturing and usage conditions 2. Minimizing process sensitivity to product and manufacturing variations In essence here, we design for producibility. The process for this development follows the following steps: • • • •

Understand capability and stability of present processes. Understand the high time-in-service robustness of the present product. Minimize product sensitivity to noise, as required. Minimize process sensitivity to product and manufacturing variations, as required.

The inputs for this process are based on the following processes and information: • • • • • • • •

Present process capability (µ target and σ) P-diagram, with critical y’s, x’s, n’s Transfer function (as developed to date) Manufacturing and assembly process flow diagrams, maps Gage R&R capability studies PFMEA & DFMEA data Verification plans: robustness and reliability checklist Noise management strategy

Once these inputs have been identified, developed, and understood, then the formal optimization begins. Remember, there is a big difference between maximization and optimization. We are interested in optimization because we want to equalize our input in such a way that when we do the trade-off analysis we are still ahead. That means we want to decrease variability and satisfy the customer without adding extra cost. How is this done? By making sure all the active individuals participate and all have ownership of the project as well as technical knowledge. Specifically, in this stage, the owners of the DFSS project will be looking to make adjustments in both variability and sensitivity using optimization and modeling equations and calculations to optimize both product and process. The central formula is DMAIC 12

2   ∂y 2 2  ∂y  2   σ y = ( ) σ x1 +  σ ... + + x2   ∂x1   ∂x2   

DCOV Whereas the focus of the DMAIC model is to reduce σ2xx (variability), the focus of the DCOV is to reduce the ∂y ∂x (sensitivity). This is very important, and it

(

)

SL3151Ch16Frame Page 721 Thursday, September 12, 2002 5:57 PM

Closing Thoughts about Design for Six Sigma (DFSS)

721

is why we use the partial derivatives of the x’s to define the Ys. Of course, if the transformation function is a linear one, then the only thing we can do is to control variability. Needless to say, in most cases we deal with polynomials, and that is why DOE and especially parameter design are very important in any DFSS endeavor. Some of the steps to identify this optimizing process are (order does not matter, since in most cases these items will be worked on simultaneously): • • • • • •

Minimize variability in y by selecting optimal nominals for x’s. Optimize process to achieve appropriate σx. Ensure ease of assembly and manufacturability (in both steps above). Eliminate specific failure modes. Update control plan. Review and develop consensus.

The deliverables of this stage are: • • • • • • • • •

Transfer function Scorecard with estimate of σy Target nominal values identified for x’s Variability metric for CTS Y or related function, e.g., range, standard deviation, S/N ratio improvement Tolerances specified for important characteristics Short-term capability, “z” score Long-term capability Updated verification plans: robustness and reliability checklist Updated control plan

At this point, one of the most important steps must be completed before the DFSS team must go officially into the next step — testing and verification. This step is the evaluation process. A thorough question and answer session takes place with focus on what has transpired in this stage. It is important to ask questions such as: Have all the z scores for the CTQs been identified? How about their targets and ranges? Is there a clear definition of the product variability over time metric? And so on. Typical tools for the basis of the evaluation are: • • • • • • • • • • •

Experimental plan with two-step optimization and confirmation run Design FMEA with robustness linkages Process FMEA (including noise factor analysis) Parameter design Robustness assessment Simulation software Statistical tolerancing Tolerance design Error prevention: compensation, eliminate noise, Poka-Yoke Gage R&R studies Control plan

SL3151Ch16Frame Page 722 Thursday, September 12, 2002 5:57 PM

722

Six Sigma and Beyond

After the team is satisfied with the progress thus far, it is ready to address the issues and concerns of the last leg of the model — verification of results (V). In this stage, the team focuses on assessing the performance, the reliability, and the manufacturability of what has been designed. The process for developing the verification begins by emphasizing two items: 1. Assessing the actual performance, reliability and manufacturing capability 2. Demonstrating customer-correlated performance over time. The inputs to generate this information are based on but not limited to the following: • • • •

Updated verification plans: robustness and reliability checklist Scorecard with predicted values of y, σy, based upon µ x and σx Historical design verification plan(s) and reliability — if available Control plan

Once these inputs have been identified, developed, and understood, then the team is entering perhaps one of the most critical phases in the DFSS process. This is where the experience and knowledge of the team members through synergy will indeed shine. This is where the team members will be expected to come up with physical and analytical performance test(s) as well as key life testing to verify the correlation of what has been designed and the functionality that the customer is seeking. In other words, the team is actually testing the “ideal function” and the model generating the characteristics that will delight the customer. Awesome responsibility indeed, but doable. The approach of generating some of the tests is: • • • •

Enhance tests with key noise factors. Improve ability of tests to discriminate good/bad commodities. Apply test strategy to maximize resource efficiency. Review and develop consensus.

The deliverables are test results, such as: • • • • • •

Product performance over time Weibull, hazard plot, etc. Long-term process capabilities Completed robustness and reliability checklist with demonstration matrix Scorecard with actual values of y, σy. Lessons learned captured in system design specifications, component design specifications, and verification design system.

To say that we have such and such a test that will do this and this and will conform to such and such condition or circumstance is not a big issue or important. What is important and essential is to be able to assess the performance of what you have designed against the customer’s functionalities. In other words: Are all your

SL3151Ch16Frame Page 723 Thursday, September 12, 2002 5:57 PM

Closing Thoughts about Design for Six Sigma (DFSS)

723

x’s correlated (and if so, to what degree) to Xs which in turn correlate to y which in turn correlate to the Y (the real functional definition of the customer)? Have the phases D, C, and O of the model been appropriately assessed in every stage? How reliable is the testing? And so on. Some of the approaches and methodologies used are (order does not matter, since in most cases these items will be worked on simultaneously): • • • •

Reliability/robustness plan Design verification plan with key noises Correlation: tests to customer usage Reliability/robustness demonstration matrix

SL3151Ch16Frame Page 724 Thursday, September 12, 2002 5:57 PM

SL3151ZAppAFrame Page 725 Thursday, September 12, 2002 5:56 PM

Appendix: The Four Stages of Quality Function Deployment STAGE 1: ESTABLISH TARGETS Develop a planning matrix (multiple functions). • Recognize the voice of the customer. • Analyze major product features. • Perform market and technical evaluation of competitive products. • Establish targets for major features. Evaluate strengths and weaknesses of your product offering (engineering leadership role). • Design, technology, reliability, cost • Major selling features, critical targets, necessary breakthroughs Note: First QFD meeting held: Vice president of engineering chairs meeting for the purpose of critiquing the design concept and target setting.

STAGE 2: FINALIZE DESIGN TIMETABLES AND PROTOTYPE PLANS Discuss/evaluate all possible means of achieving important characteristics. • Decision on technology to be used • Targets and tolerances for critical components identified • Fault tree analysis/design FMEA/Taguchi optimization methods used • Final characteristic deployment matrix generated Note: Second QFD management meeting held: Discuss process of targeting and mass production planning.

STAGE 3: ESTABLISH CONDITIONS OF PRODUCTION Primary emphasis is process design. • Critical component targets and tolerances related to prototype processing conditions • Optimization experiments performed as needed 725

SL3151ZAppAFrame Page 726 Thursday, September 12, 2002 5:56 PM

726

Six Sigma and Beyond

• • • •

Significant control items and means of control established Process FMEA/FTA/mistake proofing methods established Further need for breakthroughs defined Trial runs used to verify forecasted process stability, capability, adequacy of control points, and product quality • Output should be QA method providing excellent results with minimum factory effort • Process quality planning matrices generated • Leadership of the QFD process transferred from engineering to manufacturing when the decision to go ahead with mass production is made

Note: Third QFD management meeting held: • Evaluate probability of success of pending mass production program • Discuss details of quality assurance system

STAGE 4: BEGIN MASS PRODUCTION STARTUP • Develop database on mass production capabilities versus plan. • Identify problems and areas for further improvement. • Integrate operator-suggested efficiency or effectiveness improvements into plan. • Identify additional customer inputs. Note: Fourth and final quality management meeting held three months after start of mass production. Manufacturing leads; engineering participates. • Actual performance data integrated into QFD package • Additional study needs detailed and commitments made The importance and focus on the “voice of the customer” results in: 1. Reinforcing the quality linkage between departments 2. Setting the priorities for achieving marketplace advantage 3. Reducing the time from concept to product delivery

TANGIBLE BENEFITS • • • • •

Major reduction in development time Virtual elimination of late engineering changes Lower cost designs at outset Enhanced design reliability Economical factory controls

SL3151ZAppAFrame Page 727 Thursday, September 12, 2002 5:56 PM

Appendix: The Four Stages of Quality Function Deployment

INTANGIBLE BENEFITS • • • • • •

Increased customer satisfaction Stable duality assurance planning activity QFD documentation package Often applies to generic family Transferable storehouse of engineering know-how Basis for improvement planning

SUMMARY VALUE Strengthens current development process • Clear targets defined early based on market business demands • Simultaneous focus on product and process technologies • Key issues remain visible for prioritizing resource allocation • Communication and teamwork enhanced Desired output efficiently achieved • Products meet customers’ needs. • Products provide a competitive edge.

THE QFD PROCESS 1. The project a. Selection • Broad appeal • Simple but not trivial • Opportunity to improve • Management support • Available expertise • Available market information b. Scope/targets • Project limitations, operating constraints, product constraints • Market segment • Regularity requirements • Cost • Mass c. Objectives • Reason for doing • Expected results/outcome d. Timing • Spans full product cycle • Months work • Concentrated effort • Hours meetings/members • Significant time commitment

727

SL3151ZAppAFrame Page 728 Thursday, September 12, 2002 5:56 PM

728

Six Sigma and Beyond

2. Team a. Selection • Cross-functional • Members • Membership • Product planning • Styling • Marketing • Product/manufacturing engineering • Operations • Key supplier • Service • Product assurance/testing • Expertise (not position) • Keep ranks about equal • Open-minded members b. Operation • Facilitator/leader • Recorder • Regular meetings • Meeting to organize • Team consensus • Agreement • Not voting • No one dominates • No factions c. Agree to support group decisions d. Team training • At least one person knowledgeable of QFD • QFD overview • Other training as needed • Team building • Creative thinking • Problem solving • Meeting skills • Facilitator skills • Interview/survey methods • Employee involvement team skills

MANAGING THE PROCESS 1. Timing • Process spans a major portion of the product development process. • Identify intermediate measures of progress. • Major projects will require 50–60 hours of meetings.

SL3151ZAppAFrame Page 729 Thursday, September 12, 2002 5:56 PM

Appendix: The Four Stages of Quality Function Deployment

2.

3.

4.

5.

729

• Meetings are used to coordinate activities and update charts. • Most of the work happens outside the meetings. Supporting the team • Provide the time. • Demonstrate your commitment. • Push for progress, but not too hard. • Be realistic. • Review charts — make sure you understand. • Set priorities if needed. • Help team through the rough spots. • Keep asking the right questions. What to look for • Blank rows — unfulfilled customer wants • Blank columns • Unnecessary requirements • Incomplete customer wants • Rows of columns with only weak relationships — banking a lot on “maybes” • Unmeasurable “hows” • Too many relationships (Less than 50% “white space” makes it hard to prioritize.) • Opportunities to excel • Negative correlations (Try to eliminate, trade off if needed.) • Conflicting competitive assessments Common pitfalls • QFD on everything • Inadequate priorities • Lack of teamwork • Wrong participants • Turf issues • Lack of team skills • Lack of support • Too much “chart focus” • Handling trade-offs • Internal focus • “Stuck on tradition” • “Hurry up and get done” • Failure to integrate QFD Some right questions • How was the voice of the customer determined? • How were the design requirements (etc.) determined? (Challenge the usual in-house standards.) • How do we compare to our competition? • What opportunities can we identify to gain a competitive edge? • What further information do we need? How can we get it? • How can we proceed with what we have?

SL3151ZAppAFrame Page 730 Thursday, September 12, 2002 5:56 PM

730

Six Sigma and Beyond

• What trade-off decisions are needed? • What can I do to help? 6. Points to remember • The process may look easy but requires effort. • Many of the entries look obvious — after they are written down. • If there are not “tough spots,” it probably is not being done right. • Focus on the end-user customer. • Charts are not the objective. • Charts are the means of achieving the objective. • Find reasons to succeed, not excuses for failure.

SL3151ZAppAFrame Page 731 Thursday, September 12, 2002 5:56 PM

Selected Bibliography Adams, L., Measuring by Comparison, Quality, May 2001, pp. 32–34. Allen, M.J. and Yen, M., Introduction to Measurement Theory, Wadsworth, Belmont, CA, 1979. Anon., Statistics Roundtable: Statistical Gymnastics Revisited, Quality Progress, Feb. 1999, pp. 84–94. Anon., Statistical thinking and its contribution to total quality, The American Statistician, 44, 116–121, 1990. Atkinson, H., Hamburg, J., and Ittner, C., Linking Quality to Profits: Quality Based Cost Management, Quality Press, Milwaukee, 1994. Aubrey, C., Six Sigma Creates Success in Service Sector, QFD and Design for Six Sigma, Proceedings 45th EOQ Congress 2001, Instanbul, Sept. 20, 2001. Balasubramanian, R., Concurrent Engineering — A Powerful Enabler of Supply Chain Management, Quality Progress, June 2001, pp. 47–54. Barnes, E.B. and Mohanty, G.P., Tolerance allocation methodology for manufacturing, 1996 SAE Reliability, Maintainability, Supportability and Logistics Conferences and Workshop, Proceedings of the 8th Annual SAE RMS Workshop, 1996, pp. 49–54. Bongiorno, J., Improving FMEAs, Quality Digest, Oct. 2000, pp. 37–40. Boyce, W.E. and DiPrima, R.C., Elementary Differential Equations and Boundary Value Problems, 7th ed, Wiley, New York, 2000. Brauer, J.R., Finite Element Analysis, Marcel Dekker, New York, 1998. Breyfogle, F.W., Implementing Six Sigma: Smarter Solutions Using Statistical Methods, Wiley-Interscience, New York, 1999. Breyfogle, F.W., Statistical Methods for Testing, Development and Manufacturing, WileyInterscience, New York, 1992. Britz, G. et al., Statistical Thinking, Special Publication, ASQC Statistics Division, Spring 1996, pp. 2–23. Burke, R.J. and Maier, N.R.F., Attempts to predict success on an insight problem, Psychological Reports, 17, 303–310, 1965. Campanella, J., Ed., Principles of Quality Costs: Principles, Implementation, and Use, 3rd ed., Quality Press, Milwaukee, 1999. Chen, I.J. et al., Quality Managers and the Successful Management of Quality: An Insight, Quality Management Journal, 2000, pp. 40–54. Cone, G., Six Sigma: Black Belt Training for Quality, Automotive Excellence, Fall 1998, pp. 10–14. Daniels, S.E., Product Safety and Reliability: The Failures, SUV Rollovers Put Quality on Trial, Quality Progress, Dec. 2000, pp. 30–48. Daniels, S.E. and Hagen, M.R., Making the Pitch in the Executive Suite, Quality Progress, Apr. 1999, pp. 30–48. Davenport, T.H., Process Innovation: Reengineering Work Through Information Technology, Harvard Business School Press, Boston, 1993. De Pablos, L.A., Six Sigma Quality Metric vs. Taguchi Loss Function, QFD and Design for Six Sigma, Proceedings 45th EOQ Congress 2001, Instanbul, Sept. 20, 2001.

731

SL3151ZAppAFrame Page 732 Thursday, September 12, 2002 5:56 PM

732

Six Sigma and Beyond

Deleryd, M., Deltin, J., and Klefsjo, B., Critical factors for successful implementation of process capability studies, Quality Management Journal, 6(1), 40–59, 1999. Denton, K., The Tool Box for the Mind: Finding and Implementing Creative Solutions in the Workplace, Quality Press, Milwaukee, 1999. Dimitriades, Z.S., Empowerment in total quality: designing and implementing effective employee decision-making strategies, Quality Management Journal, 8(2), 19–28, 2001. Dodson, B., Weibull Analysis, Quality Press, Milwaukee, 1995. Dovich, R.A., Reliability Statistics, Quality Press, Milwaukee, 1990. Draman, R.H. and Chakravorty, S.S., An Evaluation of Quality Improvement Project Selection Alternatives, Quality Management Journal, 2000, pp. 58–73. Dusharme, D., Gage Use and Abuse: A Guide to Common Gage Misuse, Quality Digest, Feb. 1999, pp. 30–33. Field, J.M., Beyond design: implementing effective production work teams, Quality Management Journal, 8(2), 29–43, 2001. Fleischer, M. and Liker, J.K., Concurrent Engineering Effectiveness, Hanser Gardner, Cincinnati, 1997. Finn, G., Building Quality into Design Engineering, Quality Digest, Feb. 2000, pp. 35–38. Fraenkel, J., Wallen, N., and Sawin, E.I., Visual Statistics. A Conceptual Primer, Allyn & Bacon, Needham Heights, MA, 1999. Fuller, W.A., Measurement Error Models, Wiley, New York, 1987. Gebrael, M.G., Markov Modeling as a Reliability Tool, 1996 SAE Reliability, Maintainability, Supportability and Logistics Conferences and Workshop, Proceedings of the 8th Annual SAE RMS Workshop, 1996, pp. 37–44. Genest, D.H., Improving Measurement System Compatibility, Quality Digest, Apr. 2001, pp. 35–40. Ghiselin, B., Ed., The Creative Process, University of California Press, Berkeley, 1952. Goldratt, E., Critical Chain, North River Press, Great Barrington, MA, 1998. Goldratt, E., Essays on the Theory of Constraints, North River Press, Great Barrington, MA, 1998. Goldratt, E., Necessary But Not Sufficient, North River Press, Great Barrington, MA, 2000. Goodden, R., How a Good Quality Management System Can Limit Lawsuits, Quality Progress, June 2001, pp. 55–59. Goodenow, W.H., How to Become a Master Black Belt Organization Without Six Sigma, Quality in Manufacturing, Jan./Feb. 2001, pp. 16–18. Gorsuch, R.L., Factor Analysis, W.B. Saunders, Philadelphia, 1974. Griffith, G.K., Statistical Process Control Methods for Long and Short Run, 2nd ed., Quality Press, Milwaukee, 1996. Hammer, M., Beyond Reengineering — How the Process Centered Organization is Changing Our Work and Our Lives, Harper Business, New York, 1996. Hammer, M. and Champy, J., Reengineering the Corporation. A Manifesto for Business Revolution, Harper Business, New York, 1993. Harrington, H.J., Project Management: It’s a More Important Tool than Six Sigma, Quality Digest, June 2000, p. 20. Harrington, H.J., Business Process Improvement: The Breakthrough Strategy for Total Quality, Productivity, and Competitiveness, McGraw-Hill, New York, 1991. Heil, G., Parker, T., and Stephens, D.C., One Size Fits One. Building Relationships One Customer and One Employee at a Time, Wiley, New York, 1999. Heiser, D.R. and Schikora, P., Flowcharting with Excel, Quality Management Journal, 2001, pp. 26–35.

SL3151ZAppAFrame Page 733 Thursday, September 12, 2002 5:56 PM

Selected Bibliography

733

Hoerl, R.W., Six Sigma and the Future of the Quality Profession, Quality Progress, June 1998, pp. 35–44. Holmes, D.S. and Mergen, A.E., Building an Acceptance Chart., Quality Digest, June 2000, pp. 35–36. Holpp, L., Managing Teams, McGraw-Hill, New York, 1999. Hoyer, R.W. and Hoyer, B.B.Y., What Is Quality? Quality Progress, July 2001, pp. 52–62. Hunter, J.S., Metrics for Uncertainty: A Look at Probability, Evidence and a Seldom Used Additive Metric, Quality Progress, Dec. 2000, pp. 72–73. Imparato, N. and Harari, O., Jumping the Curve. Innovation And Strategic Choice in an Age of Transition, Jossey-Bass, San Francisco, 1994. Ireson, W.G., Coombs, C.F., and Moss, R.Y., Handbook of Reliability Engineering and Management, 2nd ed., Quality Press, Milwaukee, 1996. Isaacson, J. and Chambers, W., An Introduction to Optical Measurement, Quality Digest, Oct. 2000, pp. 28–32. Janov, J., The Inventive Organization. Hope and Daring at Work, Jossey-Bass, San Francisco, 1994. Kales, P., Reliability: For Technology, Engineering, and Management, Quality Press, Milwaukee, 1998. Kalfut, M., Riding the Benchmark, Technology Century, Dec. 1997/Jan. 1998, pp. 30–31. Kall, J., Manufacturing Execution Systems: Leveraging Data for Competitive Advantage (Part I), Quality Digest, Aug. 1999, pp. 31–34. Kall, J., Manufacturing Execution Systems: Leveraging Data for Competitive Advantage (Part II), Quality Digest, Sept. 1999, pp. 31–33. Kanyamibwa, F., Christy, D.P., and Fong, D.K.H., Variable selection in product design, Quality Management Journal, 8(1), 62–79, 2001. Kaplan, R.S. and Norton, D.P., The Balanced Scorecard, Harvard Business School Press, Boston, 1996. Kapur, K.C. and Lamberson., L.R., Reliability in Engineering Design, Wiley, New York, 1977. Kay, M., Applying Six Sigma in a Public Service Organization, QFD and Design for Six Sigma, Proceedings 45th EOQ Congress 2001, Instanbul, Sept. 20, 2001. Kelada, J.N., Intergrading Reengineering with Total Quality, Quality Press, Milwaukee, 1996. Kelly, C. and Kachatorian. L., Robust Design for Six Sigma Manufacturability, 1996 SAE Reliability, Maintainability, Supportability and Logistics Conferences and Workshop, Proceedings of the 8th Annual SAE RMS Workshop, 1996, pp. 25–28. Kelly, L. and Morath, P., How Do You Know the Change Worked? Quality Progress, July 2001, pp. 68–74. Kish, L., Some statistical problems in research design, American Sociological Review, 24, 328–338, 1959. Kish, F.J., Utilizing value engineering as a problem solving management tool, SAE National Combined Farm Construction and Industrial Machinery, Powerplant, and Transportation Meetings, Society of Automotive Engineers, Milwaukee, Sept. 9–12, 1968, paper 680567. Knouse, S.B. and Strutton, H.D., Getting Employee Buy-In to Quality Management, Quality Progress, Apr. 1999, pp. 61–64. Krishnamoorthi, K.S., Reliability Methods for Engineers, Quality Press, Milwaukee, 1992. Kume, H., Statistical Methods for Quality Improvement, The Association for Overseas Technical Scholarship, Tokyo, 1985. Lathin, D. and Mitchell, R., Learning from Mistakes, Quality Progress, June 2001, pp. 39–46. Lehmann, E.L., Testing Statistical Hypotheses, Wiley, New York, 1986. Lepi, S.M., Practical Guide to Finite Elements, Marcel Dekker. New York, 1998.

SL3151ZAppAFrame Page 734 Thursday, September 12, 2002 5:56 PM

734

Six Sigma and Beyond

Levinson, W.A., How to Design Attribute Sample Plans on a Computer, Quality Digest, July 1999, pp. 45–47. Liberatore, R., Teaching the Role of SPC in Industrial Statistics, Quality Progress, July 2001, pp. 89–94. Livingston, S., Creating the Right Atmosphere: Setting the Stage for Innovative Thinking in Ideation Sessions, Quirk’s Marketing Research Review, May 2001, pp. 32–39. Maier, N.R.F., Problem Solving and Creativity in Individuals and Groups, Brooks/Cole Publishing Co., Wadsworth Publishing Co., Belmont, CA, 1970. Marash, S.A., A New Look at Six Sigma, Quality Digest, Mar. 1999, p. 18. Mazur, G., QFD and Design for Six Sigma, Proceedings 45th EOQ Congress 2001, Instanbul, Sept. 20, 2001. McLean, H.W., HALT, HASS and HASA Explained: Accelerated Reliability Techniques, Quality Press, Milwaukee, 2000. Meeker, W.Q., Doganaksy, N., and Hahn, G.J., Using Degradation Data for Product Reliability Analysis, Quality Progress, June 2001, pp. 60–65. Mitchell, R.H., Process Capability Indices, ASQ Statistics Division Newsletter, Winter 1999, pp. 16–20. Mitchell, E., Web-Based APQP Keeps Everyone Connected, Quality, July 2001, pp. 40–44. Modares, M., Kaminski, M., and Krivtsov, V., Reliability Engineering and Risk Analysis: A Practical Guide, Marcel Dekker, New York, 1999. Myers, R.E. and Torrance, E.P., Invitations to Thinking and Doing, Ginn, Boston, 1964. O’Connell, V., Advertising, Wall Street Journal, Nov. 27, 2000, p. B21. O’Conor, P.D.T., Practical Reliability Engineering, 3rd ed., Quality Press, Milwaukee, 1995. Orme, B., Assessing the Monetary Value of Attribute Levels with Conjoint Analysis: Warnings and Suggestions, Quirk’s Marketing Research Review, May 2001, pp. 16, 44–47 Osborn, A.F., Applied Imagination, 3rd ed., Scribener, New York, 1963. Paul, L.G., Outsourcing and Analyzing the Value Proposition, CFO, Aug. 2001, pp. 60–61. Peterman, M., Lean Manufacturing Techniques Support the Quest for Quality, Quality in Manufacturing, Jan./Feb. 2001, pp. 24–25. Peterman, M., Simulation Nation: Process Simulation Is Key in a Lean Manufacturing Company Hungering for Big Results, Quality Digest, May 2000, pp. 39–42. Porter, A. and Adams, L., Quality Begins with Good Data, Quality, May 2001, pp. 32–34. Porter, M.E., Competitive Advantage: Creating and Sustaining Superior Performance, Free Press, New York, 1985. Pylipow, P.E., Can It Be This Easy? You Can Alter Drawing Practices to Achieve Six Sigma, But Only if You Understand All the Implications, Quality Progress, July 2001, pp. 139–140. Pyzdek, T., Considering Constraints, Quality Digest, June 2000, p. 22. Pyzdek, T., The 1.5 Sigma Shift, Quality Digest, May 2001, p. 22. Quesenberry, C., Statistical Gymnastics, Quality Progress, Sept. 1998, pp. 75–78. Rosen, R. and Digh, P., Developing globally literate leaders, T+D, 55(5), 70–83, 2001. Salzman, R.H. and Liddy, R.G., Product Life Predictions from Warranty Data, 1996 SAE Reliability, Maintainability, Supportability and Logistics Conferences and Workshop, Proceedings of the 8th Annual SAE RMS Workshop, 1996, pp. 45–48. Schuetz, G., Gaged and Confused, Quality Digest, May 2001, pp. 44–47. Schwarz, F.C., Managing Progress Through Value Engineering, SAE National Combined Farm Construction and Industrial Machinery, Powerplant, and Transportation Meetings, Society of Automotive Engineers. Milwaukee, Sept. 9–12, 1968, paper 680566. Slater, R., Jack Welch and the GE Way: Management Insights and Leadership Secrets of the Legendary CEO, McGraw-Hill, New York, 1999.

SL3151ZAppAFrame Page 735 Thursday, September 12, 2002 5:56 PM

Selected Bibliography

735

Smith, D., How Good Are Your Data? Quality Digest, June 2000, pp. 50–51. Stalk, G., Jr. and Hout, T.M., Competing Against Time: How Time-Based Competition is Reshaping Global Markets, Free Press, New York, 1990. Stamatis, D.H., The Nuts and Bolts of Reengineering, Paton Press, Red Bluff, CA, 1998. Stamatis, D.H., TQM Engineering Handbook, Marcel Dekker, New York, 1997. Stasiowski, F.A. and Burstein, D., Total Quality Management for the Design Firm: How to Improve Quality, Increase Sales, and Reduce Costs, Wiley, New York, 1993. Steel, J., Truth, Lies and Advertising, Wiley, New York, 1998. Steele, J. M., Applied Finite Element Modeling: Practical Problem Solving for Engineers, Marcel Dekker, New York, 1998. Stein, P., All You Ever Wanted to Know About Resolution, Quality Progress, July 2001, pp. 141–142. Stevens, D.P., A stochastic approach for analyzing for analyzing product tolerances, Quality Engineering, 6(3), 439–449, 1994. Sun, H., Comparing quality management practices in the manufacturing and service industries: learning opportunities, Quality Management Journal, 8(2), 53–71, 2001. Subramanian, K., The System Approach: A Strategy to Survive and Succeed in the Global Economy, Hanser Gardner, Cincinnati, 2000. Taraschi, R., Cutting the Ties that Bind, Training and Development, Nov. 1998, pp. 12–14. Tichy, N.M. and Sherman, S., Control Your Destiny or Someone Else Will: Lessons in Mastering Change — From the Principles Jack Welch Is Using to Revolutionize GE, Harper Business, New York, 1993. Umble, E.J. and Umble, M.M., Developing Control Charts and Illustrating Type I and Type II Errors, Quality Management Journal, 2000, pp. 23–31. Valance, N., Prices Without Borders? CFO, Aug. 2001, pp. 71–73. Van Mieghem, T., Lessons Learned From Alexander the Great, Quality Progress, Jan. 1998, pp. 41–48. Vasilash, G.S., For Robust Products, Automotive Design and Production, Aug. 2001, p. 8. Ward, S., How Much Data is Needed? Quality, July 2001, pp. 26–29. Wearing, C. and Karl, D.P., The Importance of Following GD&T Specifications, Quality Progress, Feb. 1995, pp. 95–98. Wetmore, D., The Juggling Act, Training and Development, Sept. 2000, pp. 67–68. White, D.A. and Kall, J., Coherent Laser Radar: True Noncontact Three-dimensional Measurement Has Arrived, Quality Digest, Aug. 1999, pp. 35–38. Whitfield, K., The Current State of Quality at Honda and Toyota, Automotive Design and Production, Aug. 2001, pp. 50–52. Yilmaz, M.R. and Chatterjee, S., Six sigma beyond manufacturing — a concept for robust management, Quality Management Journal, 7(3), 67–78, 2000.

SL3151ZAppAFrame Page 736 Thursday, September 12, 2002 5:56 PM

SL3151ZIndex Page 737 Thursday, September 26, 2002 8:56 PM

Index* A Abstracting and indexing services, 145 Accelerated degradation testing (ADT), 336 Accelerated depreciation, 687 Accelerated life testing (ALT), 336, 362 Accelerated stress test (AST), 310–311 Accelerated testing, 305 ADT (accelerated degradation testing), 336 ALT (accelerated life testing), 336, 362 AST (accelerated stress test), 310–311 constant-stress testing, 305–306 definition of, 362 HALT (highly accelerated life test), 310 HASS (highly accelerated stress screens), 310 methods, 305–306 models, 306–309 PASS (production accelerated stress screen), 311–312 progressive-stress testing, 306 step-stress testing, 306 Acceleration factor (A), 308–309 Acclaro (software), 545–547 Accountants, 663 clean opinions of, 671 reports of, 671–672 Accounting accrual basis of, 676–677 books of account in, 675–676 in business assessments, 138 cash basis of, 677 and depreciation, 684 earliest evidence of, 672 entries in, 675–676 financial reports in, 664 and financial statement analysis, 688 as measure of quality cost, 492 recording business transactions in, 672–675 roles in business, 664 valuation methods in, 679–681 Accounts books of, 675–676 contra, 684 types of, 674 Accounts receivables, 681, 691

Accrual accounting, 676–677 Accrued pension liabilities, 667 Accumulated depreciation, 665–666, 684 Achieved availability, 292 Acquisitions, in product design, 196 Action plans, 161–162 based on facts and data, 107 creative planning process in, 162 documenting, 162 in FMEA (failure modes and effects analysis), 253–258 monitoring and controlling, 162–163 prioritizing, 162 Action standards. see standards Activation energy type constant (Ea), 308 Active repair time, 293 Activities in benchmarking after visits to partners, 156 defining, 150 drivers of, 151 flowcharting, 153–154 modeling, 152–153 output, 151 performance measure, 151–152 resource requirements, 151 triggering events, 150 during visits to partners, 155 Activity analysis, 150–152 Activity benchmarking, 123 Activity drivers, 151 Activity performance measure, 151–152 Actual costs, 478–480, 568, 703 Actual operating hours, 525 Actual size, 522 Actual usage, amount of, 525 Administrative process cost of, 570 improving, 490–492 as measure of quality cost, 493 ADT (accelerated degradation testing), 336 Advanced product quality planning. see APQP Advanced quality planning. see AQP Advanced Systems and Designs Inc., 405 Aerospace industry, 226 Aesthetics, 113

* Note: Italicized numbers refer to illustrations and tables

737

SL3151ZIndex Page 738 Thursday, September 26, 2002 8:56 PM

738 Aggressors, in team systems, 25 Aircraft, 196 Airline industry, 702 Allowance, 568 Almanacs, 145 Alpha tests, 266 ALT (accelerated life testing), 336, 362 Alternative costs, 703 Alternative lists, in trade-off studies, 472 Alternative rank, 476 Altman, Edward I., 699 Altshuller, Genrich, 549 Aluminum, 531 Amateur errors, 211 American National Standards Institute (ANSI), 61 Amortization, 669 Analysis of variance. see ANOVA Angular dimensions, 520 Angular measurements, 527 Annual reports, 146, 671–672, 678 ANOVA (analysis of variance), 396, 407–410 for cumulative frequency, 423 for NTB signal-to-noise ratios, 431–432 for raw data, 430, 434, 436 signal-to-noise (S/N) ratio as raw data, 434, 437 for transformed data, 438 typical table setup, 432 ANOVA-TM computer program, 405 ANSI (American National Standards Institute), 61 ANSYS program, 183–185 Antifreeze, 289 Apollo program, 226 Apple Computer, 195 Appraisal costs, 101, 482, 489 APQP (advanced product quality planning), 266 vs. AQP (advanced quality planning), 43 in DFSS (design for six sigma), 45–47 and product reliability, 298 AQP (advanced quality planning), 40–42 vs. APQP (advanced product quality planning), 43 basic requirements for, 42 demonstrating, 42 pitfalls in, 43–44 qualitative methodology in, 44–45 reasons for using, 42 workable plans for, 43 Archiving, 40 Area sensors, 218 ARIZ algorithm, 549 Arrhenius model, 308–309 Assembly lines, 207 simulation of, 170–175 two-station, 173–174

Six Sigma and Beyond Assembly mistakes, 213 Assembly omissions, 214 Assembly process, 206 Assessment items, 476 Assets, 679 in balance sheet equation, 664, 674 buying, 666 contra, 684 current, 665 current value of, 680 vs. expenses, 680–681 financial, 681 in financial statements, 679 fixed, 665–666 historical cost of, 679 inflation effect on, 679 intrinsic value of, 680 as investments, 680 liquidation value of, 679 noncurrent, 667 physical, 681–682 psychic value of, 680 replacement cost, 680 return on assets (ROA), 694 return on assets managed (ROAM), 695 return on gross assets (ROGA), 695 return on net assets (RONA), 695 selling, 669 slow, 670 types of, 681–682 valuation methods, 679–681 values based on historical costs, 678–679 Assets/equity ratio, 692 Asset turnover, 704–705 AST (accelerated stress test), 310–311 AT&T, benchmarking in, 122–123 Attributes, 116 Attributes tests, 313–314, 423 Auditing, 597 Automation, 682 Automobile industry, 54–5 AQP (advanced quality planning) in, 41 commonly used elements in, 176 product reliability in, 296–297 six sigma philosophy in, 2 Automobile parts industry, 54 Availability, 292, 356 Axiomatic design applying to cars, 543–544 axioms in, 542 benefits of, 545–547, 545–547 and change management, 545 changing existing designs with, 544–545 creating new designs with, 544 diagnosing existing designs with, 544

SL3151ZIndex Page 739 Thursday, September 26, 2002 8:56 PM

Index and project workflow, 545 vs. robust design, 543 Axiomatic designs, 541, 715 Axioms, 542

B Balance sheets, 664–665 accrual accounting in, 678 in annual reports, 671 changes in working capital items in, 670 current assets and liabilities, 665 current liabilities, 665–666 earning per share, 669 equation, 664, 673–675 fixed assets, 665–666 footnotes in, 670 gross profit, 668 income statements, 667–668 noncurrent assets, 667 noncurrent liabilities, 667 ratio analysis of, 689–690 shareholder's equity, 667 slow assets, 666 sources of funds, 669 statement of changes in, 669 in summary of normal debit/credit balances, 674 use of funds, 670 working capital format in, 666 Bank filings, 147 Bankruptcy, 679 Barriers to market, 129 Barter, 53 Basic functions, 557, 574–575 Basic manufacturing process, 206 Basic needs, 229 Basic quality, 68–69, 70 Bathtub curve, 293 Beams, 176 Beam sensors, 218 Behavioral theory, 663–664 Beliefs, and change management, 127 Benchmarking, 97 alternatives, financial analysis of, 163–164 alternatives, identifying, 129–132 alternatives, prioritizing, 139–142 areas of application of, 97–99 and business strategy development, 99–102 and change management, 126–129 classical approach to, 102–103 common mistakes in, 166 continuum process, 98 and Deming management method, 110–111

739 and Deming wheel, 111–112 in design FMEA, 267–268 in DFSS (design for six sigma), 717 financial, 157 in FMEA (failure modes and effects analysis), 230 gaining cooperation of partners in, 148 generic, 122 history of, 97 identifying candidates for, 129–134 identifying cause of problems with, 134 in least cost strategy, 100–101 making contacts for, 149 as a management tool, 119–120 managing for performance, 164–166 and national quality award winners, 107–110 operations process, 123 and organizational change, 126–129 organizations for, 123–124 performance and process analysis, 149–158 preparing proposals for, 149 activities before visiting partners, 149 understanding own operations, 149 activity analysis, 150–152 activity modeling, 152–153 flowcharting process, 152–153 activities during visit, 155 understanding partners' activities, 155 identifying success factors, 155–156 activities after visit, 156 activities of partners, 155–156 in process FMEA, 275–276 project evaluations, 165 resistance to, 127 scopes of, 120–121 and Shewart cycle, 111–112 and six sigma, 105–107 sources, 142–149 and SQM (strategic quality management), 102–105 success factors in, 124–126, 164–166 technical competitive, 78 ten-step process in, 121–122 types of, 122–123 Benefit-cost analysis, 610 Beta tests, 266 Bibliographies, 145 Bilateral tolerance, 523 Binomial distribution in fixed-sample tests, 315–316 in sequential tests, 317–318 Biographical sources, 145 Black Belts, in dealing with projects, 661 Bladed wheel hopper feeders, 207

SL3151ZIndex Page 740 Thursday, September 26, 2002 8:56 PM

740 Blast-create-define method, 582–584 Block diagrams, 234, 323–325 Blockers, in team systems, 25 Boeing Co., 169, 196 Bolted joints, 181 Boltzman's constant (Kb), 308 Bond rating companies, 695–696 Bonds, 695–696 Bonds payable, 667 Bookkeeping, 672 Bookshelf data, 349 Books of account, 675–676 Book value, 666, 687 Boothroyd, Geoffrey, 202–203 Boundaries, in teams, 29 Boundary diagrams, 258–260 Box, George, 404, 429 Brainstorming, 230, 267–268 and concept of functives, 57 in creative phase of job plans, 582 in design FMEA, 267–268 in determining causes of failures, 247 in developing alternatives to functions, 558–559 in planning DOE (design of experiments), 372–373 in process FMEA, 275–276 in value control approach, 582 Branch transmissions, 535 Brand names, 89–94 Breakdowns, 278 Breakeven analysis, 705–706 Breakthrough strategies, 160 Buckling, 176, 179 Budgets, 711–712 calculating, 604 departmental, 711–712 managing, 712 and satisfaction of management, 662 zero-based, 712 zero-growth, 712 Burden, 568–569 Business assessment forms, 135–139 Business assessments, 133 Business assets. see assets BusinessLine, 146 Business meetings, 19 Business reviews, 145 Business strategy, and benchmarking, 99–100 Business transactions, recording, 672, 675 BusinessWire, 146 Buyer/supplier relationship. see customer/supplier relationship Buying groups, 148

Six Sigma and Beyond

C Cadillac, 107–108 Calendar elapsed time, 525 Calibration, 526 Calipers, 527 Capacitive tests, 530 Capital investments, 661 Capital surplus, 670 Carlzon, Jan, 125–126 Case studies, 148 Cash, 681 in annual reports, 671 in business transactions, 675 vs. profits, 678 ratio analysis of, 691–692 recording, 675 sources of, 669, 673 uses of, 673 Cash basis, of accounting, 677 Cash flow, 700 calculating, 701 and change management, 127 and current assets and liabilities, 702 definition of, 700 depreciation as part of, 685 forecasting, 691 as measure in TOC (theory of constraints), 462 in NPV (net present value) analysis, 610 present value of, 162–163, 165 and tax shelter schemes, 702 and working capital, 702 Cashing out, 667 Cash receipts journals, 675 Casting, 204 Catalogs, 582 Categories, 477 Category lists, in trade-off studies, 472 Catholic clerics, 672, 673 Causality, 33 Cause and effect relationships, 33, 134, 376 CDI (customer desirability index), 77 Cendata, 146 Census, 146 Centerboard hopper feeders, 207 Center for Advanced Purchasing Studies, 157 Centralized benchmarking, 124 Centrifugal hopper feeders, 207 Chain rule, 656 Chambers of commerce, 147 Change, psychology of, 126–127 Channel value, 54 Characteristic matrix, 63–64 Characteristics, 254–255 Charting, 133

SL3151ZIndex Page 741 Thursday, September 26, 2002 8:56 PM

Index Check sheets, 484 Chemical measurements, 527 Chi-square test, 334, 626 Classification, 254–255 Classified attribute analysis, 422–426 Classified data, analysis of, 421–430 Classified responses, 422 Classified variable analysis, 426–428 Clean opinions (accounting), 671 Clerical process, as measure of quality cost, 493 Closed systems, 35 CNC lathe, 61–63 Coefficient of expansion, 531 COGS (cost of goods sold), 711 in financial benchmarking, 157 in product design cycle, 194 reducing, 545 Color marking sensors, 218 Column interaction tables, 384, 386 Combination design, 415–418 Combinex method, 589–591 Commerce Business Daily, 147 Commercial cost, 570 Commercial credit ratings, 698 Commodity management organizations, 15 Common stocks, 696–697 Company life cycle, 133 Comparators, 527 Compensation costs, 36 Competition and DFSS (design for six sigma), 717 and earnings, 693 and product demand, 662 Competitive assessments, 82, 83–84 Competitive best performers, 143 Competitive bidding, 10 Competitive evaluations, 118–119, 131 Competitive quality ratings, 487 Competitive Strategy (book), 99 Competitors, 118 Complaints handling, 61–62 indices for, 484 processing and resolution of, 484 Complex reliability block diagrams, 323–325 Components, 205 costs of, 571–572 levels of, 442 tolerance levels of, 447–454 Component testing, 266 Component view, 238 Composite credit appraisal, 698 Compound growth rates, 711 Comptrollers, 503 Computer databases, 146

741 Computer formats, 339 Concept FMEA, 224, 262 Concept phase, 295 Concurrent engineering, 199, 468 Condition, statement of. see balance sheets Conduction, 179 Conference method, 513–515 Confidence level around estimation, 409 of demonstration tests, 312 Configuration, probability of, 637–638 Conformance, 29–30, 112 Conjoint analysis, 88 in DFSS (design for six sigma), 715, 718 empirical example of, 90–94 hypothetical example of, 89–90 managerial uses of, 95–95 Constant dollars, 484 Constant rate failure, 619 Constant-stress testing, 305–306 Constraints, 180, 457–458, 463–465 Construction contractors, 147 Consultants, 148 Consumer's risk, 313 Consumer groups, 147 Contamination, 289 Continuous production flow manufacturing, 207–208 Continuous time waveform, 621 Continuous transfer manufacturing, 206 Contra account, 684 Contra asset, 684 Contractors, 358 Contribution margin analysis, 706 Control charts, 484, 621–624 Control factors, 393, 411 in DFSS (design for six sigma), 719 in monitoring team performance, 33 and noise interactions, 337 Controlled radius, 522 Convection, 179 Conventional dimensioning, 518 Conventional tolerancing, 518 Conveyors, 206 Cooperation. see partnering Coordinate measuring machines, 527 Copper, 531 Copper plating, six sigma in, 6 Corporate culture, 127–128 Corporate general interest buyers, 117 Corporate growth, 663 Correlation matrix, 83 Corrosive materials, 289, 294 Cost analysis, 156, 485 Cost benchmarking, 123

SL3151ZIndex Page 742 Thursday, September 26, 2002 8:56 PM

742 Cost-function worksheets, 566 Cost of goods sold. see COGS Cost of non-quality, 101 Cost of sales, 703, 711 Costs, 101 actual, 478–480, 568, 703 alternative, 703 analyzing, 156 appraisal, 101, 482, 489 benefits of DFM/DFA (design for manufacturability/assembly) on, 189 comparison reports, 480 of components, 571–572 definition of, 569 design, 569 differential, 703 direct, 36, 703 and earnings, 693 elements of, 571 of engineering changes, 297–298 estimated, 703 external failure, 101, 483, 491 extraordinary, 703 fixed, 569, 703 freight, 570 functional area, 573 and functions, 580 of goods, 683–684 historical, 679, 703 imputed, 703 incremental, 569, 703 indirect, 36, 703 internal failure, 101, 483, 490 joint, 703 manufacturing, 569, 571 monitoring system, 478 noncontrollable, 703 opportunity, 703 out of pocket, 704 per dimension, 572 per functional property, 572–573 period, 704 per period of time, 572 per pound, 572 prepaid, 704 presentation formats for, 485 prevention, 101, 482, 488 prime, 704 of processes, 571–572 product, 704 production, 704 of product unreliability, 294 quantitative, 572–573 reducing, 480 replacement, 680, 704

Six Sigma and Beyond and revenues, 711 of sales, 711 sources of information, 570 standard, 478, 570, 704 sunk, 704 in theory of firm, 662 vs. throughput, 461 tolerance limit, 480 total, 570 and value, 558 and value control, 556 variable, 704 variance of, 480 visibility of, 564–565, 568 Costs of quality. see quality costs Cost/time analysis, 134 Cost visibility, 568 in cost-function worksheets, 564, 566 techniques, 571–573 Counterpart characteristics, 73 Counters, 219 County courthouses, 147 Coupled matrix, 657 Court records, 147 Covariance, 650–651 Coverage ratios, 692 Crashes, 177 Crash programs, 194–196 Creative phase (job plans), 582–584 Credit appraisal, 698 Credit balance, 664 Credits, 664–665 in business assessments, 136 in business transactions, 672 recording, 675 revolving, 666 using, 673 Critical condition indicators, 219 Critical design review, 466 Critical success factors, 129 Crosby, P., 481 Cross-functional teams, 472, 604 CTP (process characteristics), 510 CTQ (quality characteristics), 510, 719 Cumulative density function, 618 Cumulative distributions, 170–171, 640 Cumulative frequency, 422 Cumulative rate of occurrence, 425–426 Current assets, 665 in annual reports, 671 net changes in, 670 ratio of total liabilities to, 697 Current controls, 282 Current liabilities, 665–666 analysis of, 691

SL3151ZIndex Page 743 Thursday, September 26, 2002 8:56 PM

Index in annual reports, 671 and cash flow, 702 net changes in, 670 Current ratio, 691 Current value, 680 Customer attributes, 201–202 Customer axis, 77–78 Customer desirability index (CDI), 77 Customer duty cycles, 291 Customer requirement planning matrix, 72 Customer requirements, 85–88 Customers customer axis, 77–78 in evaluation of competitive products, 203 fast response to, 107 overarching, 54 perception of performance vs. importance, 131 perception of quality, 117–119 in process FMEA, 269 processing and resolution of complaints, 484 roles in customer/supplier relationship, 13 satisfaction of. see customer satisfaction service hot lines for, 118 surveying, 118–119 types of, 229 view on quality, 114 voice of, 73, 83 wants and needs of, 53–54, 228–230 Customer satisfaction, 74 and benchmarking, 104 vs. customer service, 49–51 in expanded partnering, 25 levels of, 49 vs. loyalty, 49 in partnering, 23, 25 and product design, 113–114 and product performance, 288 and product reliability, 292–293 scorecard, 718 Y relationship to, 718 Customer service. see services Customer/supplier relationship, 11–14 checklists of, 21–23 improving, 20–21 interface meetings in, 16 major issues with, 19–20 Customs and traditions, 582

D DAA (dimensional assembly analysis), 199 DaimlerChrysler, 203, 296–297 Dana Corp., 54 Databases, 146 Data failure distribution, 633

743 Data processing, 6, 493 Data recording, 526 DDB (double declining balance) method, 686–687 Death spiral symptom, 461 Debit balance, 664 Debits, 664–665 in business transactions, 672 recording, 675 using, 673 Debt in annual reports, 671 and equity, 692, 697 long-term, 667 net reduction in, 670 in theory of firm, 661 Debt/assets ratio, 692 Decay time, 618 Decentralized benchmarking, 124 Decimal dimensions, 520 Decision analysis, 610 Decline and Fall (book), 661 Decline period, of product life cycle, 699–700 Decoupled designs, 542 Decoupled matrix, 657 Defective parts, 278 Defect matrices, 485 Defects, 209–210 detecting, 216 examples of, 213 matrices for, 484 as measure in TOC (theory of constraints), 462 mistakes as sources of, 212–213 preventing, 216 quality defects, 291 reliability defects, 291 zero, 483 Defense Technical Information Center, 340 Deferred compensation, 667 Deferred income taxes, 669 Deflection, 654–655 Deformations, 177 Degrees of freedom, 383, 407, 428 Dell Corp., 169 Delphi Automotive Systems, 169 Demand and earnings, 693 factors affecting, 129 and sales forecasting, 710 Deming, W.E., 110, 480–481 Deming management method, 110–111 Deming wheel, 111–112 Demographic data, 146 Density, 180 Density function, 633 Departmental budgets, 711–712

SL3151ZIndex Page 744 Thursday, September 26, 2002 8:56 PM

744 Department benchmarking, 123 Department of Defense, 555 Depreciation, 684 accelerated, 687 accumulated, 665–666, 684 in cash flow analysis, 701 as expenses, 665, 669, 684 as part of cash flow, 685 of physical assets, 681 replacement cost, 687 straight line, 685–686 sum-of-the-years' digits (SYD), 686 as tax strategy, 684–-685 as valuation reserve, 684 Derating, 359, 362 Descriptive feedback, 31, 33 Design controls, 249, 265–266 Design cost, 569 Design customers, 229 Design engineering, as measure of quality cost, 493–494 Design engineers, 269 Design FMEA, 224–225, 262; see also FMEA (failure modes and effects analysis) calculating RPN (risk priority number) in, 267 describing anticipated failure modes in, 264 describing causes of failure in, 264–265 describing effect of failure in, 264 describing functions of design/product in, 264 detection table, 252 in DFSS (design for six sigma), 721 estimating failure detection in, 266–267 estimating frequency of occurrence of failure in, 265 estimating severity of failures in, 265 failure modes, 240–244 forming teams for, 263 functions of, 264 identifying system and design controls in, 265 linkages to process FMEA and control plans, 258–260 objectives of, 263 occurrence rating, 249 purpose of, 265 in QFD (quality function deployment), 725 recommending corrective actions in, 267–268 requirements for, 263 severity rating, 246 special characteristics for, 257 timing, 263 Design for manufacturability/assembly. see DFM/DA Design for six sigma. see DFSS Design margins, 359–360

Six Sigma and Beyond Design of experiments. see DOE Design optimization, 178, 182–185 Design parameters, 336–337, 542, 656 Design phase, 5 Design reliability, 313 Design requirements, 87–88 Design reviews, 464 checklists of, 468 definition of, 362 FMEA (failure modes and effects analysis) in, 467 objectives of, 466 in R&M (reliability and maintainability), 352 sequential phases of, 465–467 in system/component level testing, 266 Designs, 183 axiomatic, 541–544 existing, 544 extensions and changes to, 544 new, 544 Design sets, 184 Design synthesis, 38–39 Design variables, 183–185 Destructive testing, 529 Detect delivery chutes, 219 Detection ratings, 250 and lowering risks, 267–268, 276 vs. occurrence ratings, 253 in surrogate machinery FMEAs, 282 Detectors, 216–219 Development risk, 195 Deviations, 91–93 Dewhurst. Peter, 202–203 DFM/DFA (design for manufacturability/assembly), 187–189 approach alternatives to, 198–199 business expectations from, 189–190 charters, 193 and cost of quality, 509–510 effects on product design, 204 elements of success in, 192–194 fundamental design guidance for, 204–205 instruction manuals for, 199, 200–203 mechanics, 199 objectives of, 187–189 in product design, 195, 204 product design in, 194 product plans in, 194 and product reliability, 298 sequential approach to, 191 simultaneous approach to, 191 tools and methods for, 198–199 use of human body in, 199–200 DFSS (design for six sigma), 9

SL3151ZIndex Page 745 Thursday, September 26, 2002 8:56 PM

Index and APQP (advanced product quality planning), 45–47 and AQP (advanced quality planning), 40–47 and cost of quality, 510 essential tools in, 715–716 implementing with project management, 605–609 model, 716 partnering in, 9–25 physical and performance tests, 722–723 and project management, 608–609 quality engineering approach in, 25–33 and R&M (reliability and maintainability), 364–365 and reengineering, 516–517 and simulation, 185 stages in, 717–723 systems engineering in, 34–40 and TOC (theory of constraints), 463 transfer function in, 52 transformation functions in, 717–718 Diameter, 522 Difference between two means, 655–656 Differential costs, 703 Differentiation strategy, 101–102, 110 Digital Equipment Corp., 203 Digital signal processing, 622 Digitizing, 178 Dimensional assembly analysis (DAA), 199 Dimensional mistakes, 214 Dimensioning, 518–522 Dimensions, 522 Direct costs, 36, 703 Direct labor, 569 Direct magnitude evaluation (DME), 586 Direct materials, 569 Directories, 145–146 Direct product competitor benchmarks, 122 Discrete time, 621 Discretionary funds, 700 Discriminant analysis, 699 Discriminators, 477 Displacements, 179–180 Displacement sensors, 218 Disposable razor, failure modes in, 66–67 Distribution, 54 Diversity, in team systems, 29–30 Dividends, 670, 700 DMAIC model, 7 DME (direct magnitude evaluation), 586 Documentation, 40, 474 DOE (design of experiments), 367–370; see also experiments analysis, 405–420 ANOVA (analysis of variance), 407–410

745 classified data, 421–430 combination design, 415–418 graphical analysis, 405–407 signal-to-noise (S/N) ratio, 411–415 comparisons using, 369 confirmatory tests in, 418–421 definition of, 362 in DFM/DFA (design for manufacturability/assembly), 199 dynamic situations in, 430–441 group runs using, 369 loss function in, 397–398 parameter design, 441–447 planning, 372–380 in reliability applications, 335–336 setting up experiments, 380–395 signal-to-noise (S/N) ratio in, 403–404 Taguchi approach, 370–371 tolerance design, 447–454 Dominance factors, 273 Donneley demographics, 146 Door intrusion beams, 177 Double declining balance (DDB) method, 686–687 Double-entry bookkeeping, 672, 675 Double feed sensors, 218 Dow Jones, 146 Downtimes, 238, 278 Dry run, 362 Duane model, 361 Dun and Bradstreet, 146, 698 Dupont Connector Systems, 6 DuPont system, modified, 157, 704–705 Durability, 112 Durability life, 289 Dust, 289 Dynamic analysis, 176 Dynamic process, vs. static process, 1–2 Dynamic situations, 430–441

E Earning ratios, 693–695 Earnings, 692–693 and change management, 127 and luck, 693 retained, 670 Earnings before interest and taxes (EBIT), 668 Earnings per share (EPS), 662, 669 EBIT (earnings before interest and taxes), 668 Economic buyers, 117 Economic order quantity (EOQ) model, 707 Economies of scale, 155 Effort goals, 159 Eigenvalues, 177 Eigenvectors, 177

SL3151ZIndex Page 746 Thursday, September 26, 2002 8:56 PM

746 Eight-level factors, 389 Elastic buckling, 176 Elasticity, 177, 179 Elastic of modulus, 654 Electrical design margins, 359 Electrical discharges, 289 Electrical measurements, 527 Electroforming, 204 Electronics industry, 2, 5 Element connectivities, 180 Element data recovery, 181 Element properties, 180 Elevating hopper feeders, 207 Emission standards, 54 Employees, 663 and benchmarking, 104, 107 motivation and earnings of, 693 Enclosures, 358 Encyclopedia of Business Information Services, 145 Encyclopedias, 145 Engineering in business assessments, 136–137 conformance elements in, 502 manufacturing, 494 nonconformance elements in, 503 plant, 494–495 Engineering analysis, 266 Engineering changes, costs of, 297–298 Enhancing functions, 58–59, 264 Environmental controls, 526 Environmental FMEA, 225 Environmental laws, 54 Environmental Protection Agency (EPA), 147 EOQ (economic order quantity) model, 707 EPS (earnings per share), 662, 669 Equal bilateral tolerance, 523 Equipment, 362, 675 Equipment errors, 526 Equity in balance sheet equation, 664, 674 and debt, 692 ratio to total debt, 697 of shareholders, 667 in theory of firm, 661 Equity/debt ratios, 692 Equity earnings, 484 Erlicher, Harry, 555 Errors, 526–527 eliminating, 212 inevitability of, 212 proofing, 208, 274 variables, 336–337 Essential functions, 58–59 Esteem value, 558

Six Sigma and Beyond Estimated costs, 703 Euclid, 542 Euler buckling analysis, 176, 177 Evaluation phase (job plans), 585–591 Evaluation summary, 587 Evidence books, 475, 477 Excel (software), 182 Exchange value, 558 Excitement needs, 229 Excitement quality, 69–71 Executives, 663 Expanded partnering, 12–14 Expansion, coefficient of, 531 Expected customer life, 289 Expenditures, 680 Expenses, 701 vs. assets, 680–681 depreciation as, 684 and productivity, 702 Experiments, 249; see also DOE (design of experiments) analysis in, 405–410, 415–418 column interaction tables in, 384 confirmatory tests in, 418–421 degrees of freedom in, 383 dynamic situations in, 430–441 factor levels in, 380–382 factors with large numbers of levels in, 392 factors with three levels in two-level arrays in, 391 factors with two levels in three-level arrays in, 390–391 hardware test setups in, 385–386 inner and outer arrays in, 393 linear graphs in, 382–384 nesting of factors in, 392 orthogonal arrays in, 383–384 parameter design, 441–447 planning, 372–380 randomization of tests in, 394 test arrays in, 387–389 tolerance design, 447–454 Exponential distribution, 617, 641 in fixed-sample tests, 318–320 in reliability problems, 618 in sequential tests, 321–323 Exponential function, 619 Extended interior penalty functions, 185 External failure costs, 101, 483, 491 External gate hopper feeders, 207 External manufacturing, 6–7 External variations, 28 Extraordinary costs, 703 Extrusion, 204

SL3151ZIndex Page 747 Thursday, September 26, 2002 8:56 PM

Index

F F.W. Dodge reports, 145 Fabrication process, 206 Factors, 380 ANOVA (analysis of variance) decomposition of, 410–411 choosing number of levels, 380–382 decomposition of, 410 effects of, 424–425 eight-level, 389 four-level, 389 nesting of, 392 nine-level, 390 test matrix for, 377 three-level, 385, 391, 392 in three-level arrays, 390–391 two-level, 390 types of, 393 Facts, and change management, 127 Fail safe design, 208 Failure, 362 causes of, 246–247, 264–265, 272–273 constant rate, 619 costs of, 483, 490–491 cumulative function, 635 detecting, 267, 274–275 effects of, 264, 271–272 free time, 643 logs of, 282 methods of determining, 247–249 occurrence rating, 249, 265, 273 probability of, 618–620 severity of, 265, 273 user costs, 484 Failure Definition and Scoring Criterion (book), 288 Failure modes, 240 cause and occurrence, 246–247 describing, 264, 270–271 design controls for, 249–250 detection rating, 250 determining, 66, 247–249 effects of, 243–245, 278 examples of, 242–243 in FMEA (failure modes and effects analysis), 242–244 in function diagrams, 66 in machinery FMEA, 277–278 process controls for, 249–250 severity rating, 244–245 Failure modes and effects analysis. see FMEA Failure rate, 633 conversion to MTBF (mean time between failures), 361

747 in failure-truncated tests, 318 as measure of product reliability, 290 and product life, 293–295 in R&M (reliability and maintainability), 355 and system failure, 629 Failure reporting, analysis, and corrective action system (FRACAS), 341, 352, 362 Failure-truncated tests, 318–319 FAST (functional analysis system technique), 567, 577–580, 712 Fatigue, 294 Fault tree analysis. see FTA Fax machines, in customer/supplier communications, 13 FEA (finite element analysis), 175 analysis procedure in, 178–179 common problems in, 182 definition of, 362 input to models in, 180 outputs from, 180–181 procedures in, 178 solution procedure in, 179–180 techniques in, 182 types of, 176–177 Feasibility, 362 Feasible designs, 183–184 Feature of size, 522 Features, 112, 522 Federal Database Finder, 147 Federal depository libraries, 147 Federal Reserve banks, 147 Feedback, 30–31 descriptive, 31, 33 loops, 30–31 negative, 35 positive, 35 systems, 31, 35 Feeders, 207 Feelings, and change management, 127 Fiber optical tests, 530 Fiber sensors, 218 Field performance data, 487 Field service reports, 279 FIFO (first-in first-out) method), 683 Finance, as measure of quality cost, 495 Financial analysis, 704–709 breakeven analysis, 704–705 contribution margin analysis, 706 EOQ (Economic Order Quantity) model, 707 IRR (internal rate of return) method, 709–710 modified duPont system, 704–705 NPV (net present value) method, 709 price–volume variance analysis, 707 return on investment analysis, 708–709 ROI (return on investment), 708–-709

SL3151ZIndex Page 748 Thursday, September 26, 2002 8:56 PM

748 Financial assets; see also assets Financial benchmarking, 157 Financial forecasting, 688 Financial leverage in annual reports, 672 and earnings, 692 in financial comparison, 131 in modified duPont formula, 704–705 in rating bonds, 695 in rating stocks, 697 ratios, 692 Financial management rate of return (FMRR), 709–710 Financial planning, 710–712 Financial position, statement of. see balance sheets Financial rating companies, 695 Financial rating systems, 695 bond rating companies, 695–696 commercial credit ratings, 698 Financial ratios, 145 Financial reports, 146, 664 accountants' report, 671 annual reports, 671 audited, 671 balance sheets, 664–665 Financial statement analysis, 688 Finished product inspection, 528 Finite element analysis. see FEA Finite elements, 175–176 Firm, theory of, 661–662 First-in first-out (FIFO) method, 683 Fishbone diagram, 134, 348, 373 Fixed assets, 665–666 accumulated depreciation of, 684 in cash flow analysis, 701 in financial comparison, 131 as noncurrent assets, 667 Fixed burden costs, 569 Fixed costs, 569 in breakeven analysis, 704–705 in contribution margin analysis, 706 definition of, 703 Fixed-sample tests, 314 using binomial distribution, 315–316 using exponential distribution, 318–320 using hypergeometric distribution, 315 using Poisson distribution, 316 using Weibull and normal distributions, 320 Florida Power and Light, 143 Flow charts, 61–63, 153–154 Fluid mechanics, 177 FMEA (failure modes and effects analysis), 223–224 action plans in, 253–258 in benchmarking, 134

Six Sigma and Beyond benefits of, 226 common problems in, 260–262 in core engineering process, 299 definition of, 224, 362 design FMEA. see design FMEA design/process controls in, 250 in design reviews, 467 detection rating in, 250 in DFM/DFA (design for manufacturability/assembly), 199 in DFSS (design for six sigma), 716, 721 failure mode analysis in, 240–242 forms, 235–238 vs. FTA (fault tree analysis), 469 function concepts in, 53, 64–68 getting started with, 228–235 history of, 226–227 initiating, 227–228 learning stages in, 262 machinery. see machinery FMEA in problem solving, 225–226 process. see process FMEA in quality lever, 227 scopes of, 236 steps of, 469 transferring causes and occurrences to forms, 250 transferring current controls and detection to forms, 254 transferring RPN to forms, 256 transferring severity and classification to forms, 248 types of, 224–225 typical body, 237 typical header, 236 FMRR (financial management rate of return), 709–710 Focus groups, 131, 147 Footnotes, in balance sheets, 670 Force, 527 Force field analysis, 128–129 Ford Motor Co., 41, 169, 203, 296–297 Forecasts in expanded partnering, 17 financial, 688 as measure of quality cost, 495 sales forecasting, 710 technology forecasting, 156–157 Forgetfulness, 210 Forging, 204 Fork lifts, 207 Formal qualification review, 466 Four-level factors, 389 FRACAS (failure reporting, analysis, and corrective action system), 341, 352, 362

SL3151ZIndex Page 749 Thursday, September 26, 2002 8:56 PM

Index F ratio statistical test, 408 Freedom of Information Act, 147 Free-state conditions, 522 Freight costs, 570 Frequency distributions, 170 Friction, 179, 294 FTA (fault tree analysis), 299 definition of, 362 in design FMEA, 267 in determining causes of failures, 248 vs. FMEA (failure modes and effects analysis), 469 in QFD (quality function deployment), 725 in R&M (reliability and maintainability), 348, 355 seven-step approach to, 469–470 Fuji-Xerox, 143 Functional analysis, 156 Functional analysis system technique (FAST), 567, 577–580, 712 Functional area costs, 573 Functional benchmarking, 122, 123 Functional requirements, 541–543 Function journals, 145 Functions, 52–53, 574 alternatives to, 558–559 analysis and evaluation, 567–568, 575–580 basic functions, 574–575 and costs, 580 definition of, 52, 230 of designs, 264 determining, 567, 573–574 developing, 238 diagrams, 55–56 as dimension of product quality, 113 enhancing, 58–59 essential, 58–59 evaluating, 580–582 failure modes, 66–67 in FMEA (failure modes and effects analysis), 64–68, 230 objective, 183 organizing, 239–240 penalty, 185 in product flow diagrams, 56–61 of products, 264 in QFD (quality function deployment), 64–68 secondary, 575 task, 58–59 terminus, 66–67 tree structure, 239–240 types of, 264 in VA (Value Analysis), 64–68 in value analysis, 557 in value control, 567–568, 573–574

749 values, 581–582 Function tree process, 239–240 Functives, 55–56 Funds in annual reports, 671 in balance sheets, 669 discretionary, 700 sources of, 661, 669 use of, 670

G G. Heilman Brewing, 99 GAAP (generally accepted accounting principles), 672 Gages, 527 accuracy, 533 blocks, 531–533 in hierarchy of standards, 525 linearity, 533 repeatability, 533 reproducibility, 533 stability, 533 Galbraith, John Kenneth, 663, 681 Gale Research, 145 Gamma distribution, 625–631 Gamma functions, 626–631 Gamma ray tests, 530 Gap analysis, 158–159 Gasoline fumes, 289 Gates, 219 GD&T (geometric dimensioning and tolerancing), 199, 518–523 General design standards, 338 General Electric Co., 555 General journals, 675 General ledgers, 676 Generally accepted accounting principles (GAAP), 672 General Motors Corp., 296–297 annual report (1986) of, 670 AQP (advanced quality planning) in, 41 ROE (return on equity), 99 General services, as measure of quality cost, 501 Generic benchmarking, 122 Generic products, 89–94 Geometric analysis, 176 Geometric dimensioning and tolerancing (GD&T), 199, 518–523 Geometry, 180 Goals, 159 characteristics of, 159 customer-oriented, 131 guiding principles, 160 interdepartmental, 161

SL3151ZIndex Page 750 Thursday, September 26, 2002 8:56 PM

750 philosophy in setting, 159–160 in project management, 608 results vs. effort, 159 service/quality, 131 strategic, 19 structures of, 160–161 tactical, 19 Goals down–plans up (forecasting), 710 Goodwill, 491, 666 Gorton Fish Co., 169 Government, 663 Government Printing Office Index, 147 Government regulations, 129 Graham, Benjamin, 697–698 Graphical analysis, 405–407 Graph transmissions, 535–537 Grid point data recovery, 181 Gross assets, 695 Gross profit, 668 Gross profit margin, 130, 157, 668 Groups. see teams Group technology (GT), 199 Growth period, of product life cycle, 699–700 Growth rates, 663, 711 GT (group technology), 199 Guide rods/pins, 219 Guide to the Project Management Body of Knowledge (book), 599

H Habits, and productivity, 582 HALT (highly accelerated life test), 310 Handbooks, 145 Hardened tool steel, 531 Hardness testers, 527 HASS (highly accelerated stress screens), 310 Hazard rate, 294, 634–635, 643 Heat transfer, 179, 357 Heavy equipment industry, 200 Help-seekers, in team systems, 25 Hidden costs, 36 Hidden factories, 36 Highly accelerated life test (HALT), 310 Highly accelerated stress screens (HASS), 310 Histograms, 484 Historical costs, 679, 703 Historic data, 133, 266 Homeostasis, 31 Hood buckling, 177 Hopper feeders, 207 Horizontal beam deflection, 654–655 House of Quality matrix, 73, 140–141 Human body, in DFM/DFA (design for manufacturability/assembly), 199–200

Six Sigma and Beyond Human mistakes, 210–212 Human resources conformance elements in, 506 in customer/supplier relationship, 22 in expanded partnering, 24 nonconformance elements in, 506 in partnering, 24 Humidity, 289, 526 Hypergeometric distribution, 315

I IBM, 109–110, 195 Identification mistakes, 210–211 Image, in quality, 113 Implementation phase (job plans), 591–592 attitude, 596 audit results, 597 goals, 591–592 organization, 594–595 plans, 592–593 principles, 593–594 system evaluation, 593 value council, 596–597 Importance/feasibility matrix, 141 Importance/performance analysis, 131 Importance rating, 84 Improvement potential, 141 Imputed costs, 703 Inadvertent mistakes, 211 Incentives, in benchmarking, 104–105 Inch dimensions, 520–521 Income after taxes, 484 Income before extraordinary items, 668–669 Income before nonrecurring items, 668–669 Income before taxes, 668 Income from continuing business, 668 Income statements, 664, 667–668 in annual reports, 671 ratio analysis of, 690 in summary of normal debit/credit balances, 674 Income taxes, 683–685 Incoming material inspection, 528 Incremental costs, 569, 703 Independent quality rating, 487 Indexing mechanisms, 206 Indicators, 21, 527 Indirect costs, 36, 703 Indirect labor, 569 Indirect materials, 569 Industrial cleanser, 89–94 Industrial engineering, 508 Industrial state, 663 Industry analysis, 129–130

SL3151ZIndex Page 751 Thursday, September 26, 2002 8:56 PM

Index Infant mortality period, 293 Infeasible designs, 183–184 Inflation, 679 and inventories, 682–683 and sales forecasting, 710 Information collecting, 564 in customer/supplier relationship, 22 in expanded partnering, 24 Information brokers, 148 Information phase (job plans), 563–564 cost visibility, 564–565, 568 functions in, 567–568, 573–574 information collection, 564 project scope, 565–567 Information systems and management in business assessments, 139 conformance elements in, 508 nonconformance elements in, 509 in partnering, 22, 24 Information theory, 542 Informative inspection, 217 Inherent availability, 292 In-house reviews, 467 Injuries, costs of, 36 Inland Steel, 99 Inner arrays, 393 Innovations, levels of, 549 In-process inspection, 528 Input, 27, 61 Input output method, 577 Inspections in classifying characteristics, 529 interpreting results of, 530 points, 528 process in, 206 purpose of, 528 stations, 487 techniques in, 217 types of, 528–529 Instruments, 525 Intangible assets, 667 Integrator approach, 199 Intellectual property, 22 Intelligence Tracking Service, 146 Intentional mistakes, 212 Interdepartmental goals, 161 Interest income, 668, 685 Interface matrix, 258–260, 279 Interference, 360 Interference testing, 321–323 Interim design review, 466 Intermittent transfer manufacturing, 206 Internal assessments, 132 Internal benchmarking, 122

751 Internal Internal Internal Internal Internal Internal

best performers, 143 failure costs, 101, 483, 490 manufacturing, 5–6 organizations, for partnering, 14–15 processes, 53 rate of return (IRR) method, 611–612, 709–710 Internal standards and tests, 79 Internal variations, 29 International System of Units (SI), 520 Internet, in customer/supplier communications, 13 Intrinsic value, of assets, 680 Inventories, 682 cycles, 707 determining value of, 682–683 Economic Order Quantity (EOQ) model, 707 as internal failure cost, 490 Inventory control systems, 159 Inventory profits, 458 Inverse power model, 307–308 Inversion method, in problem-solving, 583 Investments, 458 assets as, 680 bonds, 695–696 capital, 661 and depreciation, 685 rating systems for, 695–696 stocks, 696–698 Invoice, 675 IRR (internal rate of return) method, 611–612, 709–710 Ishikawa diagram, 134 ISO 9000 certification program, 42 ISO/TS 19469 certification program, 42

J Jaguar, 296–297 JIT (just-in-time) method, 199 Job plans, 559 creative phase, 582–584 evaluation phase, 585–591 implementation phase, 591–597 information phase, 563–568, 573–574 steps in, 561–562 vs. techniques, 562–563 Job shops, 207 Johnson Controls, 54 Joint costs, 703 Joint stiffness evaluation, 177 Joint ventures, 196 Judgment inspection, 217 Juran, J., 480 Just-in-time (JIT) method, 199

SL3151ZIndex Page 752 Thursday, September 26, 2002 8:56 PM

752

K Kaizen method, 142, 160 Kano model, 68–71 basic quality depicted in, 70 of customer needs, 229 in DFSS (design for six sigma), 715, 718 excitement quality depicted in, 70–71 performance quality depicted in, 70 and transformations, 53 Kepner Trago analysis, 248 Key life testing, 299 K factor, 4 Kolmogov-Smirnov test, 334

L L.L. Bean, 124, 143 Labor, 569 Laboratories, 487 Laboratory errors, 526 Lack of standards mistakes, 211 Landlords, 147 Last-in first-out (LIFO) method, 683–684 Law of maldistribution, 585 Laws of mechanics, 542 LCC (life cycle costs), 348 definition of, 363 in R&M (reliability and maintainability), 356–7 Leadership, in partnering, 21–24 Lead facilitators, in trade-off studies, 472–473 Lead times, 74 Leapfrog approach, 196 Learning curves, 155 Leasing agents, 147 Least cost strategy, 100–101 Ledgers, 676 Legal department conformance elements in, 509 as measure of quality cost, 495 nonconformance elements in, 509 Legislative summaries, 147 Leverage in annual reports, 672 and earnings, 692 in financial comparison, 131 in modified duPont formula, 704–705 in rating bonds, 695 in rating stocks, 697 ratios, 692 Liabilities accrued pension, 667 in balance sheet equation, 664, 674 current, 665–666

Six Sigma and Beyond and current assets, 697 increasing, 669 noncurrent, 667 pension, 667 Life cycles, 133 of companies, 133 definition of, 363 of products. see product life cycle LIFO (last-in first-out) method, 683–684 Limits dimensioning, 523 Lindbergh, Charles, 554 Linear graphs, 382–384, 386 Linear measurements, 527 Linear static analysis, 177 Line elements, 176 Line organizations, 15 Liquid assets, 691 Liquidation value, 679 Liquidity, 665, 671 in financial comparison, 131 ratio analysis of, 691–692 Liquidity ratios, 691–692 Loads, 180, 181 Long-term debts, 667 Long-term process variation, 4–5 Loops, 184, 538 Loop transmission, 538 Losses, 671 and cash flow, 702 controlling, 36 as part of transactions, 672 Loss function, 397–398 calculating, 398–402 for LTB (larger-the-better) situations, 402 vs. process performance (Cpk), 402–403 and signal-to-noise (S/N) ratio, 403–405 for STB (smaller-the-better) situations, 401 Low safety factors, 294 LTB (larger-the-better), 402, 413 Lubricants, 33, 289 Luck, and earnings, 693

M Machine condition signature analysis (MCSA), 363 Machine customers, 229 Machinery FMEA, 224–225, 277; see also FMEA (failure modes and effects analysis) classification in, 279 current controls, 282 detection ratings, 282 failure modes, 277–278 and FTA (fault tree analysis), 348

SL3151ZIndex Page 753 Thursday, September 26, 2002 8:56 PM

Index identifying functions, 277 identifying scopes of, 277 occurrence ratings in, 282 potential causes in, 279 in R&M (reliability and maintainability), 351–352, 361 recommended actions in, 283 RPN (risk priority number), 282–283 severity rating, 279 Machining, characteristic matrix of, 63–64 MacLaurin series, 646, 649 Magnetic disk feeders, 207 Magnetic elevating hopper feeders, 207 Magnetic fields, 289 Magnetic particle tests, 530 Maintainability, 292, 338, 363 Maintenance records, 282 Major parts standards, 338–339 Malcolm Baldrige National Quality Award, 105 Maldistribution, law of, 585 Management benchmarking in, 98, 103–104, 124 and budgets, 662 in business assessments, 138 and earnings, 693 as measure of quality cost, 496 operational, 601 in partnering, 19 roles in customer/supplier relationship, 21–22 and security, 663 systems concept in, 35–-37 in theory of firm, 662 Management process benchmarking, 122–123 Manpower, 483 Manuals, 145 Manufacturing-based view, 114 Manufacturing cells, 206 Manufacturing cost, 569, 571 Manufacturing engineering, 296 Manufacturing engineering sign-of approach, 199 Manufacturing engineers, 269 Manufacturing process, 206 approaches to, 207–208 in business assessments, 136 categories of, 206, 206–207 as cause of product failures, 291 conformance elements in, 506–507 controls, 273–274 costs, 569 costs of, 571 design-related factors in, 197 external, 6–7 factors affecting, 197–198 functions, 269–270 improving, 489

753 internal, 5–6 as measure in TOC (theory of constraints), 462 nonconformance elements in, 507 one in, one out, 207 product design-related factors affecting, 197 in R&M (reliability and maintainability), 350 schematic diagram, 206 secondary, 206 and theory of non-constraints, 464 Margin/fit problems, 177 Marketable securities, 671, 681 Marketing advantages in, 74 conformance elements in, 505–506 as measure of quality cost, 497 nonconformance elements in, 506 Market niches, 54 Market research, 148 in DFSS (design for six sigma), 715, 718 in product development, 295 as source of benchmarking information, 144 in surveys, 118 Market segmentation, 102 in benchmarking, 125 in DFSS (design for six sigma), 717 planning, 45 Market segments, 54 Market share, 696 Mass, 527 Massachusetts Institute of Technology, 541–543 Mass production, 518, 726 Master Belt, in dealing with projects, 661 Master Black Belt, in dealing with projects, 661–661 Materials, 569 analysis of, 176 in business assessments, 136 direct, 569 errors in, 526 handling, 206–207 indirect, 569 as input in team systems, 26 as measure of quality cost, 497–499 properties of, 180 raw, 487, 523 as source of mistakes, 212 in TOC (theory of constraints), 458 Mathematical modeling, 266 Matrix analysis, 587–589 Maturity period, of product life cycle, 699–700 Maytag, 99 MCSA (machine condition signature analysis), 363 Mean deflection, 654 Means, difference between two, 655–656

SL3151ZIndex Page 754 Thursday, September 26, 2002 8:56 PM

754 Mean time between failures. see MBTF Mean time to repair. see MTTR Measurement mistakes, 214 Measurement systems, 524–525 interpreting results of inspection and testing in, 530 mechanical, 527 purpose of inspection in, 528–529 roles of, 525–527 sources of inaccuracy, 526 techniques and equipment, 527–528 testing methods, 529–530 Mechanical design margins, 359–360 Mechanical loads, 178–179 Mechanical measurements, 527 Mechanics, laws of, 542 Medical costs, 36 Mergers and acquisitions, asset values in, 680 Metal detectors, 218 Metric dimensions, 520 Metric tolerance, 520 Metrology, 524–525 interpreting results of inspection and testing in, 530 purpose of inspection in, 528–529 roles of, 525–527 techniques and equipment, 527–528 testing methods, 529–530 Microeconomics, 662 Microinches, 531 Micrometers, 527 Microswitches, 219 Miles, L.D., 555 MIL-HDBK-727 method, 203 Milliken and Co., 143 Millimeters, 520 Minority interest, 667 Mirror image (accounting), 676 Mission statements, 160 Mistake proofing, 208–209 in avoiding workplace errors, 210 devices for, 216–219 equation for success in, 218 inspection techniques in, 217 proactive system approach to, 216 in process FMEA, 274 reactive system approach to, 216 Mistakes, 213–215 causes of, 213–215 detecting, 216–217 examples of, 213 human, 210–212 preventing, 216–217 signals that alert, 215 sources of, 212–213

Six Sigma and Beyond types of, 213–215 Mistakes of misunderstanding, 210 Mitsubishi method, 200–201 Models and modeling, 178 in engineering analysis, 266 in FEA (finite element analysis), 169 finite element, 180 redesigns of, 181–182 as tool of quality cost, 485 Modified duPont system, 157, 704 Money. see funds Monochrome monitors, 358 Monte Carlo method, 169 Moody's, 146, 696 Motion economy, principles of, 201 Motorola Inc., 1 benchmarking programs in, 109–110, 157 six sigma quality programs, 101 Mounting mistakes, 214–215 MSC/NASTRAN software, 180 MTBE (mean time between event), 354–355 MTBF (mean time between failures), 348 conversion to failure rate, 361 definition of, 363 in failure-truncated tests, 318 and inherent availability, 292–293 machine history of, 349 as measure of product reliability, 290 and occurrence ratings, 282 in R&M (reliability and maintainability), 348, 355 in sequential tests, 321 in time-truncated tests, 319–320 MTTF (mean time to failure), 363, 619 MTTR (mean time to repair), 348 definition of, 363 machine history of, 349 in R&M (reliability and maintainability), 292–293, 355–356 Musts and wants, 477

N Nasser, Jacques, 297 National Electrical Manufacturing Association (NEMA), 358 National Institute of Standards and Technology (NIST), 526 National reference, 525 National standards, 525 National Technical Information Center, 147 Need sets, 53 Negative confirmation, in customer satisfaction, 49 Negative feedback, 31, 35

SL3151ZIndex Page 755 Thursday, September 26, 2002 8:56 PM

Index NEMA (National Electrical Manufacturing Association), 358 Net assets, 695 Net income, 671 Net present value (NPV) method, 611, 709 Net profits, 459–460 New products, 710 Newsearch, 146 Newsletters, 145, 147 Newspapers, 145 Nine-level factors, 390 NIST (National Institute of Standards and Technology), 526 Node absorption, 539 Noise factors, 393, 411, 721 Noises, 336–337 Nominal dimension, 531 Nominal group process, 114, 132–134 Nominal size, 522 Non-constraints, theory of, 463–464 Noncontrollable costs, 703 Noncurrent assets, 667 Noncurrent liabilities, 667 Nondestructive testing, 530 Non-disclosure agreements, 17 Nonlinear analysis, 176 Nonlinear dynamic analysis, 177 Nonlinear static analysis, 177 Nonprofit organizations, 668 Nonrecurring expenses, 669 Non-rigid parts, 522 Non-statistical controls, 274 Normal density-like function, 647 Normal distribution, 320 in fixed-sample tests, 320 in sequential tests, 323 Normalizing constant, 308 Normal modes analysis, 177 Not invented here syndrome, 110 NPV (net present value) method, 611, 709 NTB (nominal-the-best), 413–415 in loss function, 399 signal-to-noise (S/N) ratio for, 404–405, 431, 439–441 Nuclear radiation, 289 Numerical evaluation. see paired comparisons

O Objective functions, 183 Object oriented analysis and design (OOAD), 515–516 Observed frequency, 422 Occupational safety laws, 54 Occurrence rating, 249;

755 see also severity rating in design FMEA, 249 and lowering risks, 267, 276 in machinery FMEA, 282 in process FMEA, 250 reducing, 253 Odd part out method, 219 OE (operating expense), 458, 460–461 OEE (overall equipment effectiveness), 349, 356, 363 Office equipment, accounting of, 675 Omega method, 425 One Idea Club method, 124 One in, one out manufacturing process, 207 Ongoing program/project manager approach, 199 Online databases, 145 OOAD (object oriented analysis and design), 515–516 Open systems, 35 Operating characteristic curve, 313 Operating expense (OE), 458, 460–461 Operating hours, 525 Operating instructions, 73 Operating leverage, 682 Operating margin, 668 Operating profits, 484 Operational management, 601 Operational results, 23, 25 Operations mistakes, 214 Operations process benchmarking, 123 Operator errors, 526 Operator-paced free-transfer machines, 206 Operator to operator errors, 526 Opportunity cost, 195, 703 Optical measurements, 527 Optimal inventory cycle, 707 Optimization algorithms, 184–185 Optimization loops, 184 Optimum design, 183 Organizational change, and benchmarking, 126–129 Organizational suboptimization, 36 Organization expense, as noncurrent assets, 667 Orthogonal arrays, 383–384, 386 OSHA (Occupational Safety and Health Administration), 147 Outer arrays, 393 Outliers, 312 Out of pocket costs, 704 Output in process flow diagrams, 61 of teams, 28 Overall equipment effectiveness (OEE), 349, 356, 363 Overarching customers, 54

SL3151ZIndex Page 756 Thursday, September 26, 2002 8:56 PM

756 Overhead costs, 36, 130 Oxidation, 294

P Pace production line, 206, 208 Packaging, as cause of product failures, 291 Paired comparisons, 141, 586–587 Paper pencil assembly, 60 Parallel reliability block diagrams, 323–325 Parameter design, 371 in DFSS (design for six sigma), 715 in DOE (design of experiments), 441–447 in improving reliability, 336–337 Parameter Design approach, 31–32 Parametric variations, 178 Pareto, Vilfredo, 585 Pareto analysis, 44, 132–133, 484 Pareto voting, 585–586 Partial derivatives, 649 Partnering, 9–11 buyer/supplier relationship in, 11–12 checklists of, 21–23 in DFSS (design for six sigma), 13 expanded, 12–14, 23–25 implementing, 14–19 improving, 20–21 principles of, 11 process managers, 15–17 reevaluating, 17 success indicators, 21 typical questionnaire for, 18 Partnering for Total Quality assessment process, 21 Parts, 205 defective, 278 inclusion of wrong, 214 missing, 214 non-rigid, 522 in product design, 205 Parts/component feeding systems, 207 Part worths, 89–94 PASS (production accelerated stress screen), 310–312 PAT (profit after tax), 693–694 Patents, 147, 667 Path transmissions, 535, 538 Payback period method, 612, 708 PDGS-FAST system, 178 P diagrams, 299 in DFSS (design for six sigma), 715, 719 in FMEA (failure modes and effects analysis), 258–260 and team systems, 26 PDS (product design specifications), 717 P/E (price/earnings) ratio, 698

Six Sigma and Beyond Peacemakers, in team systems, 25 Penalty functions, 185 Penetrant dye tests, 530 Pension liabilities, 667 Perceived quality, 113 Perception, 113 Perfect products, 194–196 Performance, 112 vs. importance, 131 index of, 558–559 needs, 229 parameters, 292 product-based view, 113 quality of, 69, 70 reviews of, 19 Performance evaluation review technique (PERT), 604 Period costs, 704 Periodic actions, 549 Periodicals, 145 Perishable tooling, 363 Personal computers, 195 Personnel in business assessments, 139 as measure of quality cost, 499 PERT (performance evaluation review technique), 604 PFIS (plant floor information system), 363 Philip Morris, 99 Phosphate-based liquid, 89–94 Phosphate-free liquid, 89–94 Phosphate-free powder, 89–94 Physical assets, 681–682 depreciation of, 684 inventories as, 682–683 operating leverage, 682 PIMS (Profit Impact of Marketing Strategies), 157 in benchmarking, 119 objectives and benefits of, 312 par report, 130 Pin joint clearance, 181 Planning matrix, 725 Plans up form (forecasting), 710 Plant administration, 504 Plant and equipment, 136, 701 Plant engineering, as measure of quality cost, 494–495 Plant floor information system (PFIS), 363 Plant reports, 480 Plasticity, 177, 181 Plug gages, 527 Plus-minus dimensioning, 523 Pneumatic gaging, 527 Point elements, 176 Poisson distribution, 316, 509–510, 636–640

SL3151ZIndex Page 757 Thursday, September 26, 2002 8:56 PM

Index Poisson process, 635–636 Poka Yoke method, 208, 274, 275, 721 Portfolio analysis, 610 Positioning sensors, 218 Positive feedback, 35 Potential design verification tests, 258 Power supplies, 358 Practice gaps, 158 Predictive maintenance, 363 Pre-feasibility analysis, 38 Preference structure, 89 Preferred stocks, 669, 697 Preliminary design review, 466 Prentice Hall Almanac of Business and Industrial Statistics, 157 Prepaid costs, 704 Pre-planning matrix, 65 Preservation of knowledge, 74 Pressure, 527 Preventers, 216–217 Prevention costs, 482, 488 Preventive maintenance, 350, 363 Price/earnings (P/E) ratio, 698 Price–volume variance analysis, 707 Pricing, 119 and ROI (return on investment), 119 in theory of firm, 661 Primary reference standards, 525 Prime costs, 704 Principal, 685 Priorities, in FMEA (failure modes and effects analysis), 230 Prioritization matrix, 139–140 Proactive systems, in mistake proofing, 216 Probability density function, 618, 628 Probability distribution, 313, 636 Probability of configuration, 637–638 Probability of failure, 618–620 Probability of reliability, 621 Probability paper, 485 Probability ratio sequential testing (PRST), 363 Probes, 219 Process average shifts in, 1–2 short- vs. long-term standard variation in, 4–5 Process benchmarking, 122–123 Process characteristics (CTP), 510 Process Control Methods (book), 6 Process controls, 249, 268, 276 Process customers, 229 Process engineers, 269 Processes, 363 costs of, 571–572 dominance factors, 273 internal, 53

757 parameters of, 255 in partnering, 25 planning with project management, 607–608 in project management, 601–602 quality management in, 23 random vs. identifiable causes in, 133 short- vs. long-term variation in, 4–5 and six sigma, 1–2 special characteristics for, 257 standard deviation, 4–5 static vs. dynamic, 1–2 Process facilitators, in trade-off studies, 473 Process flow diagrams, 61–64, 234, 259 Process FMEA, 224–225; see also FMEA (failure modes and effects analysis) calculating RPN (risk priority number) in, 275 describing failure causes in, 272 describing failure effects in, 271–272 describing failure modes in, 270–271 describing process functions in, 269–270 detection table, 253 estimating detection of failure in, 274–275 estimating frequency of occurrence of failure in, 273 estimating severity of failures in, 273 failure modes, 242–244 forming teams for, 269 identifying manufacturing process controls in, 273–274 linkages to design FMEA and control plans, 258–260 objectives of, 268 occurrence rating, 250 recommending corrective actions in, 275–277 requirements for, 268–269 severity rating, 247 special characteristics for, 257 timing, 268 Process functions, 269–270 Process gaps, 158 Processing mistakes, 213–-214 Processing omissions, 214 Process performance (Cpk), 2, 402–403 Process plans, 72, 201 Process quality, 23, 25 Process redesign, 511–512 Procrustes (Greek mythology), 29–30 Procurement, 10 Producers, 53 Producers' risk, 313 Product assurance, as measure of quality cost, 499 Product-based view, 113 Product characteristic deployment matrix, 72

SL3151ZIndex Page 758 Thursday, September 26, 2002 8:56 PM

758 Product control, as measure of quality cost, 499–500 Product costs, 704 Product demand, and competition, 662 Product design and development, 194 basic vs. secondary processes in, 204 benefits of DFM/DFA (design for manufacturability/assembly) on, 189 case studies, 195–196 as cause of product failures, 291 and costs of engineering changes, 297–298 crash program approach to, 195 and customer satisfaction, 113–114 effects of DFM/DFA (design for manufacturability/assembly) on, 204 factors affecting manufacturing process, 197 focus of, 205 forming and sizing operations in, 204 functions of, 264 fundamentals of, 204 map guide to, 197 as measure in TOC (theory of constraints), 462 minimum performance requirements in, 198 perfect product approach to, 196 primary process in, 204 and product life cycle, 297–298 as product plan, 196–198 QFD (quality function deployment) in, 79–80, 86–88 reducing cost of, 189 reducing risks in, 267 reducing time for, 158 reliability in, 296–297 secondary process, 204 sequential approach to, 191 simultaneous approach to, 191 six sigma philosophy, 5–7 special characteristics for, 257 steps in, 295–296 Taguchi's approach to, 371 TDP (technology deployment process) in, 298–300 Product design specifications (PDS), 717 Product failures, 288, 290 Product flow diagrams, 56–61 Production, 364 costs of, 704 establishing conditions for, 725–726 mass production, 726 as measure of quality cost, 500 requirements in, 87–88 and team systems, 26 Production accelerated stress screen (PASS), 310–312

Six Sigma and Beyond Productivity, 459–460 effects of customs and traditions on, 582 effects of habits on, 582 in theory of firm, 661 Product launching, 189 Product liability, 189, 491 Product life cycle, 133 and cost of engineering changes, 297–298 as a factor in product design, 194 and failure rate, 293–295 maturity period, 699–700 and product design, 297–298 stages of, 699–700 Product plans, 194, 196–198 Product quality, 112–117 eight dimensions of, 112–113 perception of, 117–119 and return on investment, 119 Product quality deployment, 73 Product recall, 491 Product reliability. see reliability Products, 364 characteristics of, 255 defects, 291 durability life, 289 environmental conditions profile, 289–290 expected customer life, 289 function diagrams for, 56–61 functions of, 264 life cycles of, 133 minimum performance requirements, 198 with multiple characteristics, 2–3 nonconforming, 3 non-price reasons in buying, 114 reliability numbers, 290 reliability of, 288 and sales forecasting, 710 Professional associations, 145, 147 Profilometers, 527 Profitability ratios, 693–695 Profit after tax (PAT), 693–694 Profit and loss statements, 667–668 Profit before tax, 693 Profit/equity ratio, 708–709 Profit Impact of Marketing Strategies. see PIMS Profit/investment ratio, 708–709 Profits, 570 analysis of, 704–707 in annual reports, 671 and axiomatic design, 545 calculating, 668–669 vs. cash, 678 in cash flow analysis, 701

SL3151ZIndex Page 759 Thursday, September 26, 2002 8:56 PM

Index direction of, 671 maximizing, 661–662 as part of transactions, 672 planning, 710–712 and productivity, 459–460 rating, 695–696 and ROI (return on investment), 459–460 in theory of firm, 661–662 Program management, as measure of quality cost, 500 Progressive-stress testing, 306 Project decision analysis, 612–613 Project management, 599–601, 604 decision analysis, 612–613 in DFSS (design for six sigma), 605–609 generic seven-step approach to, 603–605 goal setting in, 608 key integrative processes in, 602 processes in, 601–602 and quality, 603 in six sigma, 605–609 succeeding in, 613–615 value in implementation process, 607–608 Projects, 599–601 completing, 605 describing, 603–604 justification and prioritization of, 610–613 planning, 604 planning team for, 604 risk factors, 612–613 scopes of, 565–567 selecting, 597–598 starting, 605 Proprietary information, in expanded partnering, 17 Prospectus, 146 Prototype programs, 296 Proximity detectors, 219 PRST (probability ratio sequential testing), 363 Publications, as measure of quality cost, 500 Public bids, 147 Pugh concept selection, 230–231 in design FMEA, 267–268 in DFSS (design for six sigma), 715 in process FMEA, 275–276 Pulse echo tests, 530 Purchasing conformance elements in, 505 as measure of quality cost, 501 nonconformance elements in, 505 non-price reasons in, 114 Purchasing agents, 54 Purchasing performance benchmarks, 157 Purchasing power, 54

759

Q QAA (qualitative assembly analysis), 199 QFD (quality function deployment), 53, 71–72 benefits of, 73–74 combining with Taylor's motion economy, 200–201 definition of, 73 in design FMEA, 267 development of, 87 in DFM/DFA (design for manufacturability/assembly), 199 in DFSS (design for six sigma), 715, 717–8 function concepts in, 64–68 intangible benefits of, 727 issues with, 75–76 key documents in, 72–73 methodology, 80–84 and planning, 84–86 in prioritizing benchmarking alternatives, 140–141 process management in, 727–730 process overview, 76 in product development process, 79–80, 86–88 project plan, 76–79 stages of, 725–726 summary value, 727 tangible benefits of, 727 terms associated with, 73 total development process in, 75 QOS (quality operating systems), 345–346 QS-9000 certification program, 42, 345 Qualitative assembly analysis (QAA), 199 Quality, 112–117 alternative definitions of, 112–113 basic, 68–69 costs of. see quality costs customer-driven, 107 definition of, 84–85 excitement, 69–71 improving with quality cost, 492 manufacturing-based view of, 114 as measure in TOC (theory of constraints), 462 and operational results, 23, 25 perceived, 113 perception of, 117–119 performance, 69 planning, 22, 24, 41, 102–103 product-based view of, 113 and product reliability, 291–295 and project management, 603 qualitative tool for measuring, 44 and return on investment, 119 and ROI (return on investment), 119 tables, 73

SL3151ZIndex Page 760 Thursday, September 26, 2002 8:56 PM

760 transcendent view of, 113 user-based view of, 114 value-based view of, 114 Quality characteristics (CTQ), 510, 719 Quality control, 206 charts, 72, 478 conformance elements in, 507–508 as measure of quality cost, 501 nonconformance elements in, 508 in SQM (strategic quality management), 103 system, 478 Quality costs, 477–478 analyzing, 484–485 categories of, 481–482 components of, 481–483 concepts of, 480–481 conformance elements in, 502–509 data sources for, 487 and DFSS (design for six sigma), 509–510 improving quality with, 492 inputs, 481–481 inspecting, 487 laws of, 485 measuring, 483–484 nonconformance elements, 502–509 non-manufacturing measurements for, 492–502 optimizing, 483 outputs from, 482 presentation formats for, 485 product control as measure of, 499–500 quantifying, 482–483 tools of, 484 typical monthly report, 486 Quality defects, 291 Quality engineering in DFSS (design for six sigma), 25–33 in measuring methods, 483–484 Parameter Design approach in, 31–32 Quality engineers, 269 Quality failures, costs of, 101 Quality function deployment. see QFD Quality functions, 73 Quality operating systems (QOS), 345–346 Quality ratings, 487 Quality Systems Requirements, Tooling & Equipment (book), 345 Quantitative costs, 572–573, 572–573 Quantum leap — parallel programs, 196 Questionnaires, for evaluating partnering process, 17–19

R R&D (research and development), 137, 501

Six Sigma and Beyond R&M (reliability and maintainability), 345 building and installing, 352–353 concepts, 349–350 bookshelf data stage of, 349 manufacturing process selection stage in, 350 preventive maintenance needs analysis stage in, 350 conversion/decommission of, 353 Department of Defense standards, 337–342 developing and designing, 350–352 and DFSS (design for six sigma), 364–365 implementing, 346–347 key definitions in, 362–364 objectives of, 346 operations and support of, 353 phases in, 346 plans, 364 sequence and timing of, 348–349 targets, 364 tools and measures, 347–348, 354–361 R/1000, 290 Radius, 522 Radius gages, 527 Random variable approach, 632 Random variables constant raised to power of, 653 division of, 651–652 exponential of, 652–653 functions of, 651 logarithm of, 653–654 powers of, 652 in systems failure analysis, 632 Taylor series of, 650 Random variations, 133 Random walk theory, 688 Ranking teams, 473–474, 477 Rate of change of failure (ROCOF), 294–295 Rate of growth, 663 Rate of occurrence, 425–426 Ratings, for partnering process, 17–19 Rating services, 147, 695 Ratio analysis, 688–691 coverage ratios, 692 earning ratios, 693–695 leverage ratios, 692 liquidity ratios, 691–692 return ratios, 693–695 Raw materials, 152–153, 487 Rayleigh distribution, 641 RCA (root cause analysis), 255, 364 R control charts, 274 Reactive systems, in mistake proofing, 216 Recall system, 526 Receivables, 681

SL3151ZIndex Page 761 Thursday, September 26, 2002 8:56 PM

Index Reciprocating tube hopper feeders, 207 Redesign, 511–512 Reengineering, 511 conference method, 513–515 and DFSS (design for six sigma), 516–517 OOAD (object oriented analysis and design), 515–516 process redesign in, 511–512 restructuring approach, 512–513 Reference dimension, 522 Regression analysis, 711, 718 Regulatory requirements review, 259–260 Relationship matrix, 76, 82, 201 Relays, 358 Reliability, 287 block diagrams, 323–325 costs of unreliability, 296–297 and customer satisfaction, 292–293 definition of, 364 in design, 296–297 design, 313 as a dimension of product quality, 112 DOE (design of experiments) in applications, 335–336 environmental conditions profile, 289–290 of equipment, 292 exponential distribution in, 618 gamma distribution in, 627 growth, 364 growth plots, 361 and hazard functions, 634–635 improving through parameter design, 336–337 indicators of, 290 and maintainability. see R&M (reliability and maintainability) parallel, 323–325 probability of, 287–288, 621 and quality, 291–295 reliability numbers, 290 series, 323–325 specified, 312 specified conditions, 289 specified time period of, 288–289 system, 324 in TDP (technology deployment process), 298–300 visions in, 323 Reliability Analysis Center, 339–340 Reliability defects, 291 Reliability demonstration tests, 312–313 attributes tests, 313–314 fixed-sample tests, 314–316 operating characteristic curve, 313 sequential tests, 314, 317–318, 321–322 variables tests, 314, 318–320

761 Reliability function, 632–633 Reliability numbers, 290 Reliability point, 354 Reliability relationships, 632 Reliability standards, 338 Reliability tests, 300 accelerated testing, 305 objectives of, 301–302 planning, 301 sudden-death testing, 302–304 Rents, 668 Repair active repair time, 293 as internal failure cost, 490 planning, 41, 238 Replacement cost, 704 in calculating depreciation, 687 vs. current value, 680 Requirement analysis, 38 Research and development (R&D), 137, 501 Research centers, 145 Resource requirements, 151 Result gaps, 158 Result goals, 159 Retail Scan Data, 146 Retained earnings, 670 Return on assets. see ROA Return on assets managed (ROAM), 695 Return on equity. see ROE Return on gross assets (ROGA), 695 Return on investment. see ROI Return on net assets (RONA), 695 Return on net capital employed (RONCE), 694 Return on sales (ROS), 694 Return ratios, 693–695 Revenues, 668 in annual reports, 671 and costs, 711 in price-volume analysis, 707 rating, 696 Reverse engineering, 148 Revised RPN (risk priority number), 283–284 Revolving credit, 666 Revolving hook hopper feeders, 207 Rewards, 107 Rework, 41 Rigid links, 176 Rise time, 618 Risk priority number. see RPN Risks, 696 and axiomatic design, 546 calculating, 251–253 consumer's risk, 313 and earnings, 693 in manufacturing, 195

SL3151ZIndex Page 762 Thursday, September 26, 2002 8:56 PM

762 in product development, 195 in projects, 612–613 rating, 696 reducing, 267, 275–276 ROA (return on assets), 694 as measure in TOC (theory of constraints), 462 in modified duPont formula, 704–705 in project management, 610 ROAM (return on assets managed), 695 Robert Morris Associates Annual Statement Summary, 157 Robust designs, 543 Robust teams. see teams ROCOF (rate of change of failure), 294–295 ROE (return on equity), 694 calculating, 708–709 in modified duPont formula, 704–705 ROGA (return on gross assets), 695 ROI (return on investment), 693–694 average rate of return, 708–709 as measure in TOC (theory of constraints), 462 and net profits, 459–460 payback period method, 708 and pricing, 119 and productivity, 459–460 in project management, 610–611 and quality, 119 Roll forming, 204 Roman Catholic Church, 672 Rome Air Development Center (RADC), 469–470 Rome Laboratory, 340 RONA (return on net assets), 695 RONCE (return on net capital employed), 694 Roof crush, 177 Roof matrix, 201 Root cause analysis (RCA), 255, 364 ROS (return on sales), 694 Rotary centerboard hopper feeders, 207 Rotary disk feeders, 207 Royalties, 668 RPN (risk priority number), 251–253 calculating, 267, 275 and characteristic/root causes of failure, 276 in machinery FMEA, 282–283 revised, 283–284 Rust inhibitors/undercoatings, 289 Ryan Airlines, 554

S Saab, 296–297 Sabotage, 212 SAE J1730 standard, 245 Safety margins, 359–360 Safety regulations, 54

Six Sigma and Beyond Sales in annual reports, 671 in balance sheets, 668 in benchmarking, 123 in breakeven analysis, 704–705 in business assessments, 135 as cause of product failures, 291 costs of, 703 factors affecting, 710 in financial comparison, 131 forecasting, 710 maximizing, 662 as measure in TOC (theory of constraints), 462 promoting, 157 recording, 675 return on sales (ROS), 694 statistical forecasts of, 711 in surveys, 118 in theory of firm, 662 trend in, 487 Sales goals form (forecasting), 710 Salt spray, 289 Salvage value, 687 Sample data approach, 632 Sample difference, 656 Sample space, 622–623, 632 Sampling, 170, 528 SAVE (Society of American Value Engineers), 556, 593 Savings potential, vs. time, 560 Scale parameter, 625, 641 Scales, 527 Scandinavian Airlines (SAS), benchmarking in, 125–126 Scatter plots, 484 Scheduling, in project management, 603, 604 Schools and universities, 148 Scraps, 41, 278, 490 Screening methods, 585–591 Seat belts, 177 Seating arrangements, in meetings, 33 Secondary functions, 557, 575 Secondary manufacturing process, 206 Securities, 671, 681 Security and business management, 663 as measure of quality cost, 501 Segmentation, 102, 549 Self loops, 538–539 Seminars, 147 Senior management as executive customer/supplier partner, 14 in expanded partnering, 23–24 Sensitivity analysis, 475–476 Sensors, 216–219

SL3151ZIndex Page 763 Thursday, September 26, 2002 8:56 PM

Index Sequence resistors, 219 Sequential tests, 314 for binomial distribution, 317–318 graphical solutions, 318 using exponential distribution, 321–323 using Weibull and normal distributions in, 323 Sequential unconstrained minimization technique (SUMT), 185 Series reliability block diagrams, 323–324, 323–325 Serviceability, 112, 293 Service FMEA, 225 Services in business assessments, 136 and customer satisfaction, 49–51 data on, 282 delivery of, 49–51 hot lines for, 118 and non-price reasons in buying, 114–116 Servo transformers, 358 Sets, 184, 624–625 Setup mistakes, 214 Severity rating, 245–247; see also occurrence rating components of, 279 in design FMEA, 246 estimating, 265 and lowering risks, 267, 276 in process FMEA, 247 reducing, 253 Shape parameter, 625, 641, 643 Shareholder's equity, 667, 670–671 Shareholders, 663 Shewart cycle, 111–112 Shingo method, 208 Shipbuilding industry, 200 Shipping, as cause of product failures, 291 Shock, 289 Shock spectra, 179 Shoguns, in dealing with projects, 661 Short-term process variation, 4–5 Should-cost/total-cost models, 17 Signal factors, 393, 431–432 Signal flow graphs, 535–536 basic operations on, 538 effects of self loops on, 538–539 node absorption in, 539 rules of definitions of, 538 Signals, 27 Signal-to-noise (S/N) ratio, 393 calculating, 404 and loss function, 403–405 for LTB (larger-the-better) situations, 413 for NTB (nominal-the-best) situations, 413–415, 431, 439–441

763 for STB (smaller-the-better) situations, 412 in Taguchi approach, 370 Significant factors, effect of, 424 Simulated sampling, 170–175 Simulation, 169–170 in DFM/DFA (design for manufacturability/assembly), 199 and DFSS (design for six sigma), 185, 715 in sampling, 170–175 software for, 169–170 statistical modeling in, 485 in system and design controls, 266 as tool of quality cost, 485 Simultaneous engineering, 199, 364 Sine plates, 527 Singled station manufacturing, 207 Single-entry bookkeeping, 672 Site inspections, 148 Six sigma, 1; see also DFSS (design for six sigma) and benchmarking, 105–107 equation for, 4 in external manufacturing, 6–7 in internal manufacturing, 5–6 philosophy, 1, 5–7 and product design, 5–7 and project management, 608 short- vs. long- term process variation, 4–5 Six Sigma Mechanical Design Tolerancing (book), 5 Skills development, in partnering, 21 Slow assets, 666, 670 Slowness mistakes, 211 SMEs (subject matter experts), 78 Smith, Adam, 661–662 Social events, 147 Society of American Value Engineers (SAVE), 556, 593 Software, 6, 504–505 Software FMEA, 225 Solid elements, 176 Solid mechanics, 177 Solutions, in partnering, 19 Solver (Excel), 182 Source inspection, 217 Spare parts use growth curves, 485 Special interest books, 145 Special purpose elements, 176 Specified dimensions, 523 Specified reliability (Rs), 312, 318 Spherical radius, 522 Spline gages, 527 Spoilage, 36 Sport weld forces, 177 Springs, 176

SL3151ZIndex Page 764 Thursday, September 26, 2002 8:56 PM

764 SQM (strategic quality management), 102–105 Squared deviations, 91–93 SS (sum of squares), 415–416 SSO (strategic standardization organization), 77 Stainless steel, 531 Standard's optimal inventory cycle, 707 Standard and Poor's, 145, 696–697 Standard cost, 478, 704 Standard deviation, 4–5, 91–93 Standard normal distribution, 647 Standards, 45 hierarchy of, 525 lack of, 211 Startup costs, 74 Startup losses, 278 State corporate filings, 147 Statement of changes, 669 Statement of condition. see balance sheets Statement of financial position. see balance sheets State variables, 183–185 Static process, vs. dynamic process, 1–2 Stationary hook hopper feeders, 207 Statistical analysis, 699, 711 Statistical modeling, 485 Statistical process control, 133 in monitoring team performance, 33 in process FMEA, 274 Statistical tolerancing, 721 Statistics for Experiments (book), 397 STB (smaller-the-better), 401, 412 Steel industry, 200 Steering wheels, 33 Step-stress testing, 306 Stockholders, 667 Stock markets, 688 Stocks, 670, 696–698 Stock size, 522 Stoppage, 278 Stoppers, 219 Storage, as cause of product failures, 291 Straight line depreciation, 685–686 Strain energy distribution, 177 Strain gages, 181 Strategic goals, 19 Strategic Planning Institute, 119 Strategic quality management (SQM), 102–5 Strategic quality planning, 22, 24 Strategic standardization organization (SSO), 77 Stratification charts, 484 Stress, 176 Stress contours, 177 Structural pressure, 128 Sub-customers, 54 Subject matter experts (SMEs), 78 Suboptimization, 35–36

Six Sigma and Beyond Subsidiaries, 667 Substructuring, 179 Subsystem view, 238 Success factors in business, 129–130 Success testing, 316–317 Sudden-death testing, 302–304 Sumerian farmers, 672 Sum of squares (SS), 415–416 Sum-of-the-years' digits (SYD), 686 SUMT (sequential unconstrained minimization technique), 185 Sunk cost, 704 Supermarkets, 116–117 Supervision, as measure of quality cost, 501–502 Suppliers, 10 and benchmarking, 104 councils/teams for, 15 evaluating and selecting, 14 involvement in partnering, 15 partnering managers for, 14–15 in process FMEA, 269 roles in customer/supplier relationship, 13 Supply, factors affecting, 129 Supporting functions, 264 Surface elements, 176 Surface plates, 527 Surplus in capital, 670 Surprise mistakes, 211 Surrogate machinery FMEA, 279, 282 Surveillance equipment, 148 Surveys, 118–119 Survival function, 642 SYD (sum-of-the-years' digits), 686 System/concept FMEA, 262 System controls, 265–266 System customers, 229 System design, 371 System failure, 627–629 System feedback, 31 System FMEA, 224–225, 262 System initial design review, 466 System/part FMEA, 238 System reliability, 324 System requirements review, 465–466 Systems, 34–35 definition of, 34 in management, 35–37 Systems engineering, 34 definition of, 37 design synthesis in, 38–39 pre-feasibility analysis in, 38 requirement analysis in, 38 trade-off analysis in, 39 verification in, 39–40 System view, 238

SL3151ZIndex Page 765 Thursday, September 26, 2002 8:56 PM

Index

T TABInputs, 602 TABOutputs, 602 TABTools and techniques, 602 Tactical goals, 19 Taguchi, G., 481 Taguchi model, 259, 266 vs. axiomatic design, 543 in determining causes of failures, 249 in DFSS (design for six sigma), 716 in DOE (design of experiments), 370–372 loss function in, 397–398 in product design, 371 in QFD (quality function deployment), 725 signals in, 26 Tandy Computer, 195 Tap sensors, 218 Task benchmarking, 122 Task functions, 58–59, 239, 264 Tasks, estimating, 604 Tax deductions, 685 Tax shelters, 702 Taylor's motion economy, 200, 203 Taylor series, 644–649 partial derivatives, 649 of random variable functions, 650 in two-dimensions, 649–650 variance and covariance, 650–651 TDP (technology deployment process), 298–300 Team champions, in trade-off studies, 476 Teams, 26–27 aggressors in, 25 blockers in, 25 boundaries in, 29 conformance in, 29–30 cross-functional, 472 dealing with variations in, 30–31 in DFSS (design for six sigma), 25–26 environment, 28 external variations, 28 feedback to, 31 help-seekers in, 25 input, 27 internal variations, 29 minimizing effects of variations on, 31–32 monitoring performance of, 33 non-systems approaches to, 26 output/response, 28 peacemakers in, 25 ranking, 473–474, 477 signals, 27 system interrelationships in, 33 system structure of, 27–28 in trade-off studies, 472–473

765 Technical abstracts, 145 Technical axis, 79 Technical buyers, 117 Technical system expectations (TSE), 78 Technical targets, 78 Technology deployment process (TDP), 298–300 Technology forecasting, 156–157 Telecommunications industry, 6 Temperature, 29, 289, 526 Templates, 219 Terminus functions, 66–67 Test arrays, 387–389 Testing interpreting results of, 530 methods of, 529–530 reasons for, 529 Testing firms, 148 Test procedure errors, 526 Textbooks, 145 Text databases, 146 TGR/TGW (things gone right/wrong), 349, 364 Theory of constraints. see TOC Theory of firm, 661–662 Theory of non-constraints, 463–464 Thermal analysis, 357–359 Thermal conductivity, 358 Thermal expansion coefficient, 180 Thermal rise, 358 Thermal stresses, 177 Thermodynamics, 542 Things gone right/wrong (TGR/TGW), 349, 364 Thin plates, 176 Thomas Register, 146 Three-level factors, 385, 391, 392 Three-parameter Weibull distribution, 643 Throughput, 458 vs. costs, 461 obstacles to, 461–463 in TOC (theory of constraints), 463 Time, 525 Time interest earned ratios, 692 Time to total system failure, 627–628 Time-truncated tests, 319–320 TOC (theory of constraints), 457 five-step framework of, 464–465 foundation elements of, 463 goals of, 457–458 measurement focus in, 460–461 strategic measures, 458 vs. theory of non-constraints, 463–464 throughput vs. cost world in, 461 Tolerances, 523 bilateral, 523 cost of reducing, 448

SL3151ZIndex Page 766 Thursday, September 26, 2002 8:56 PM

766 in DOE (design of experiments), 371, 447–454 impact of tightening, 449 Tolerance stack studies, 266 Tolerancing, 518–522 conventional, 518 in DFSS (design for six sigma), 716 geometric, 518 and six sigma, 1 statistical, 721 Tooling, 364 Tooling engineers, 269 Tools and equipment design of, 200 wrong and inadequate, 214 Toothpaste industry, 102 Torque, 527 Total cost, 570 Total development process, 75 Total productive maintenance (TPM), 362–3 Total quality management (TQM), 102–105 TPM (total productive maintenance), 362–363 TQM (total quality management), 102–105 Traceability, 39 Tractors, 206 Trade and Industry Index, 146 Trade associations, 145 Trade journals, 145 Trade-off studies, 470–471 checklist of, 476 conducting, 471–475 hypothetical example of, 89 matrix, 477 ranking methods in, 473–474 selection process in, 474 sensitivity analysis in, 475 standardized documentation in, 474 in systems engineering, 39 weighting rule, 475 Trade shows, 147 Traditional engineering, 468 Training, 110 Transactions, recording, 672, 675 Transcendent view, 113 Transfer functions, 51–52, 719 Transformations, 52–53, 396, 717–718 Trend charting, 133 Trial balance (accounting), 676 Triggering events, 150 Trimetrons, 218 TRIZ theory, 230 in design FMEA, 267–268 in DFSS (design for six sigma), 715 foundation of, 548 and innovation, 548 and levels of innovations, 549

Six Sigma and Beyond principles associated with, 550 in process FMEA, 275–276 tools, 549 Tryout period, of product life cycle, 699–700 TSE (technical system expectations), 78 Tumbling barrel hopper feeders, 207 Tungsten carbide, 531 Two-level factors, 390–391 Two-station assembly lines, 170–175

U U.S. Army, 203 U.S. Navy, 555 Ultrasonic tests, 530 U-MASS method, 202–203 Uncoupled matrix, 657 Unemployment insurance, 702 Unequal bilateral tolerance, 523 Uniform Commerce Code filings, 147 Unilateral tolerance, 523 United Technologies, 54 University of Massachusetts, 202–203 Unreliability, cost of, 294 Use value, 558 Useful life period, 293–294 User-based view, 114 User groups, 147

V Vacation pay, 702 Vacuum, 527 Valuation methods, 679–681 current value, 680 historical cost, 680–681 intrinsic value, 680 investment value, 680 liquidation value, 680 psychic value, 680 replacement cost, 680 Value, 557–558 and historical costs, 678 in quality, 114 types of, 558 Value analysis, 555 in DFM/DFA (design for manufacturability/assembly), 199 in DFSS (design for six sigma), 715 function concepts in, 64–68 and transformational activities, 53 Value-based view, 114 Value chains, 54 Value concept, 556

SL3151ZIndex Page 767 Thursday, September 26, 2002 8:56 PM

Index Value control, 553–555 developing alternatives in, 558–559 function analysis in, 573 functions in, 557 history of, 555 implementing, 559 job plans, 559–562 managing, 560 planned approach to, 556 techniques, 562–563 Value engineering, 555 attitudes in, 596 definition of, 556 developing plans in, 592–593 in DFM/DFA (design for manufacturability/assembly), 199 evaluating, 593 goals in, 592 in lowering costs, 581 project selection in, 597–598 selection methods in, 586 setting up organization in, 594–595 understanding principles of, 593–594 value council in, 596–597 Value Line Investment Survey, 697 Values and change management, 127 in goal setting, 160 Variable burden costs, 569 Variable costs, 704–705 Variables, 183–185 Variables tests, 314, 318–320 Variance, 480, 650–651 Variance of deflection, 654–655 Variations compensating for, 30–31 controlling, 30 external, 28 external variations, 28 internal, 29 minimizing effects of, 31–32 random, 133 system feedback, 31 Velcro, 101 Vendors, 10 Verification, in systems engineering, 39–40 Vertical integration, 10 Vibrations, 177, 289 Vibration sensors, 218 Vibratory bowl feeders, 207 Visual inspection, 529 VOC (voice of customer), 73, 83, 201 Voice mail, in customer/supplier communications, 13

767 Volkswagen, 169 Volvo, 296–297

W Wal-Mart Corp., 117 Warehouse operations, 153–154, 157 Warranties, 289 costs of, 101, 294, 297 data, 279 as external failure cost, 491 as measure in TOC (theory of constraints), 462 periods, 289 reducing, 74 Wealth of Nations (book), 661–662 Wear out period, 294 Weather, 289 Web sites, 13, 111–112 Weibull distribution, 640–643 in fixed-sample tests, 320 in plotting and analyzing failure data, 323–333 in sequential tests, 323 three-parameter, 643 using, 334–335 Weibull failure distribution, 642–643 Weibull hazard rate function, 643 Weibull probability density function, 640 Weibull reliability function, 642 Weibull scale parameter, 307 Weibull shape parameter, 307 Weight, 527 Weighted average method, 684 Weightings, 474–475, 477 Welding point indicators, 218 Westinghouse Electric Co., 203 Where to Find Business Information (book), 145 Willful mistakes, 211 Work breakdown structure, 604 defining in projects, 604 improving efficiency in, 199–200 stoppage of, 278 Working capital, 661 and cash flow, 702 format, 666 net changes in, 670 Working standards, 525 Work place, 200, 210 Writing, earliest evidence of, 672

X X-bar charts, 274

SL3151ZIndex Page 768 Thursday, September 26, 2002 8:56 PM

768 Xerox, benchmarking programs in, 97, 108–109, 122, 124, 143 X-moving range charts, 274

Y Yearbooks, 145 Yellow pages, 145 Young's modulus, 180

Six Sigma and Beyond

Z Zero-based budgeting, 712 Zero defects, 483 Zero-growth budgeting, 712 Z score, 699, 721 Z traps, 159