Vous êtes sur la page 1sur 10

III

DESIGN FOR SIX SIGMA


Although a powerful strategy, Six Sigma has limitations; and after a few years of dramatic quality improvements, companies may nd themselves faced with decreasing returns on their quality improvement efforts. The problem is not Six Sigma itself but the fact that effort is being expended to perfect awed processes and products. Design For Six Sigma (DFSS) takes a different approach and helps companies build in quality from the beginning. Using another case study of the ctitious Global Widget Company (GWC), Section III illustrates the use of DFSS tools and techniques in the creation of a totally new widget. Chapter 10 provides an introduction to DFSS, explaining the differences between it and classic Six Sigma. Similar to classic Six Sigma, DFSS is normally divided into phases; however, unlike Six Sigmas DMAIC, those phases are not consistently named or dened. Chapter 10 introduces the IDDOV methodology and the background for the new case study. The I in IDDOV represents the Identication of Opportunities phase and is discussed in Chapter 11. In addition to the steps associated with classic Six Sigmas Dene phase, Identication includes the use of a critical new tool: the Quality Function Deployment (QFD) matrix. During the Denition of Initial Design phase (Chapter 12), the reader is introduced to a second new technique, the Pugh Concept Selection, as the team quanties the benets of several potential solutions and selects the most suitable. In the Development of Concept phase (Chapter 13), the team models its proposed design and uses the classic Six Sigma tool FMEA to identify and mitigate potential risks in that design. Optimization (Chapter 14) demonstrates methods for ensuring that the project achieves the highest possible quality at the lowest possible cost by introducing the

2007 by Taylor & Francis Group, LLC

130

Six Sigma Software Development

reader to parameter and tolerance design. During the nal phase, Verication (Chapter 15), the team uses prototypes and pilots to demonstrate that its design is a robust one and that the product will meet all customer requirements.

2007 by Taylor & Francis Group, LLC

Chapter 10

Introduction to DFSS
As discussed in Chapter 1, although classic Six Sigma is powerful, it has limitations.

The Need for DFSS


Many companies have discovered that once they identify key processes and reduce variation in those processes, it becomes increasingly difcult to continue improving quality. For the rst few years after they implement Six Sigma techniques, the companies achieve tangible benets, and those benets are substantial. Not only has the quality of their products and processes improved, but the corporate culture has also changed. Employees of Six Sigma companies have a common vocabulary, clearly focused priorities, and an appreciation of the power of teamwork. The companies are stronger, and so too are their bottom lines. Quality continues to improve as the companies continue to rene their processes and reduce variation. Unfortunately, the reality is that the upward spiral cannot continue indenitely. Although companies have accepted the iterative nature of Six Sigma and the need for continuous improvement as shown in Exhibit 1.3, the fact is that they hit barriers, frequently after they reach the 4.5 sigma level. They have picked all the low hanging fruit. That is, they have xed the largest and most egregious problems. They have increased quality but they have not yet achieved the goal of ve sigma. At some point, typically before they reach ve sigma, the effort required to increase quality becomes cost-prohibitive.
131

2007 by Taylor & Francis Group, LLC

132

Six Sigma Software Development

Writing about classic Six Sigma, Harry and Schroeder aptly entitled their book Six Sigma: The Breakthrough Management Strategy Revolutionizing the Worlds Top Corporations.1 They were not exaggerating. Six Sigma has provided signicant breakthroughs and it was revolutionary; but as time passed, it was apparent to Six Sigma veterans that another breakthrough was needed. Enter Design For Six Sigma (DFSS). The premise of DFSS is simple: begin earlier in the process and design in quality rather than add it after the design has been implemented. This concept has its roots in ancient history. Archimedes2 claimed that if he had a long enough lever (and, of course, a place from which to wield that lever), he could move the earth. While companies are not trying to move the earth, they do want to break through the 4.5 sigma barrier, and they can. Archimedes statement was based on the fact that the effect of a lever is directly proportional to the distance between the fulcrum and the force applied. When it is not possible to increase the force, lengthen the lever. By increasing the length of the lever and the distance from the fulcrum, it is possible to achieve the desired results with less effort. Because they had hit a barrier and could no longer afford to expend increasing effort to improve quality, Six Sigma companies needed a longer lever. As shown on Exhibit 10.1, DFSS is that longer lever. By its very nature, being employed at the beginning of the process, DFSS places the would-be earth mover further from the problem. The result is that less force is required to achieve the desired change, and what was not costeffective becomes feasible. This is no surprise to IT professionals, who know that the earlier in the system development life cycle that a problem is detected, the less it costs to correct it. DFSS, however, is concerned with more than correcting problems it seeks to prevent them.

Dening DFSS
As its name suggests, Design For Six Sigma (DFSS) has as its core the design of a product, process, or for IT a piece of software. The objective is to create a design that can be produced at the six sigma level. Consider the effect that such a design would have. If the resulting product, process, or software module operated at the six sigma level, it would have virtually no defects. The cost of support, which IT professionals know is the largest portion of a software products life cycle, would be minimized. Similarly, the rework and redesign that often occur during the development process would be reduced, if not totally eliminated. That is, the goal of DFSS is to start at the beginning and do everything right the rst time, thus eliminating the all-too-frequent iterations in the design and production process, as well as substantially reducing support
2007 by Taylor & Francis Group, LLC

Chapter 10 Exhibit 10.1 The Effect of the DFSS Lever

133

Design

Develop

Test

Implement

Support

Eort required

The DFSS lever In this case, substantially less eort is required to achieve the same process improvement, since the eort is applied at the beginning of the process, creating a longer lever. Process improvement Design Develop Test Implement Support

2007 by Taylor & Francis Group, LLC

Process improvement

The classic six sigma lever Substantial eort is required to increase sigma level beyond 4.5, because the eort is applied after the process has been implemented.

Eort required

134

Six Sigma Software Development Forces Affecting Design


Process capability The voice of the process

Exhibit 10.2

Customer requirements The voice of the customer

Design

and warranty costs once the product is in production. Although the concept is simple, implementation requires rigorous attention to details and a consistent focus on what is most important. As Exhibit 10.2 shows, there are two forces that drive a design: (1) the customer and (2) the process. For a design to be successful, it must accomplish two things. It must: 1. Meet the customers expectations and 2. Dene a process or product that can be produced at the six sigma level. Although these might sometimes be opposing requirements, the combination of the two is what distinguishes a Six Sigma design from others. A design that satises every customer requirement but cannot be produced effectively is one that will require constant tweaking and improvement and one that will not be protable for the company. Conversely, a process that produces defect-free products that are not what the customer wants or needs is of little value. It also will not be protable for the company. Six Sigma companies are well aware of the importance of building the right product the right way and ensuring that customer requirements are satised. They are also well-versed in the need to reduce variation and in techniques to do that. DFSS has the same goals. It can be viewed as an addon or a front end to classic Six Sigma, because it provides a set of tools and techniques that helps reduce variation in the design process at the same time that they work to ensure that both the voice of the customer (the functional requirements) and the voice of the process (the capability to manufacture at the desired quality level) are fully understood and are in sync.

Phases of DFSS
Classic Six Sigma is a mature strategy. As such, it has been standardized to the extent that any Six Sigma practitioner recognizes the acronym DMAIC and knows it stands for the ve phases of Six Sigma. DFSS is younger. While companies have been using Six Sigma techniques since the mid-1990s, Chowdhurys book, entitled Design For Six Sigma, which
2007 by Taylor & Francis Group, LLC

Chapter 10

135

he explains was the rst book on the subject,3 was rst copyrighted in 2002. As might be explained by that relative immaturity, although the goals of DFSS are consistent, the way it is implemented is not. Most DFSS training courses and textbooks divide the process into between four and six phases. The names vary, as do the steps included within each one. Chowdhury4 uses the ve-letter acronym IDDOV but divides it into four phases: 1. 2. 3. 4. ID: Identify the opportunity and Dene the requirements D: Develop the concept O: Optimize the design V: Verify the design

Stamatis, author of the seven-volume Six Sigma and Beyond series,5 prefers a ve-phase process that he calls RDCOV: 1. 2. 3. 4. 5. R: Recognize D: Dene C: Characterize O: Optimize V: Verify

Honeywell International DFSS training features six phases with the acronym UDMOVC:6 1. 2. 3. 4. 5. 6. U: Understand customer needs D: Dene initial solution M: Model design performance O: Optimize design V: Verify design C: Control

A search of the Web uncovers yet another variation: ve phases named DMADV:7 1. 2. 3. 4. 5. D: Dene M: Measure A: Analyze D: Design V: Verify

A casual study reveals the similarities among the phase names. Rather than create a new acronym, this book uses Chowdhurys acronym of
2007 by Taylor & Francis Group, LLC

136

Six Sigma Software Development

IDDOV. It should, however, be noted that IDDOV will be treated as ve different phases, and that the steps described in the following chapters do not necessarily coincide with those of Chowdhury.

Distinguishing between Six Sigma and DFSS


As Chapter 1 pointed out, although classic Six Sigma and DFSS ar e complementary strategies and employ some of the same tools and techniques, there are differences between them. Exhibit 10.3 outlines those differences. It is important to note that the two strategies are neither mutually exclusive nor dependent on each other. Each has its strengths and its most effective uses. Although it is important to consider each of the factors shown in Exhibit 10.3 when deciding whether to use DFSS techniques or to treat the project as a traditional Six Sigma initiative, the critical question is often whether the project involves a new product (or process) or an existing one. DFSSs strengths are best employed on new products, while classic Six Sigma is used to improve existing products or processes. The remainder of Section III uses a case study to illustrate the IDDOV process and the use of DFSS tools and techniques.
Exhibit 10.3
Element Focus Goal

Differences between Classic Six Sigma and DFSS


Six Sigma Existing process Reduce variation DFSS New process Reduce variation and optimize performance Longer Higher Longer Higher

Time required to implement improvements Potential nancial results Payback period

Shorter Lower Shorter

Disruption to internal organization Lower Best suited for

Maximizing current process Developing new products or reengineering existing processes Cp (reducing variation) Reactive Cpk (centering within customer requirements) Predictive

Major eect is on One word description

2007 by Taylor & Francis Group, LLC

Chapter 10

137

Background to the Case Study


GWC manufactures one product, metal alloy widgets, in two sizes. Large widgets have an interior diameter of 114 inches, while the diameter of small ones is 18 inch. Although they are the market leader with a 40 percent market share, GWC is concerned about increasing competition from Consolidated Asian Widgets (CAW), which has a reputation for producing widgets of equal or better quality than GWCs but at a lower cost. The CEO of GWC was playing golf with two of GWCs major customers, Great Auto and Little Telecom, both of whom expressed their frustration because their businesses were changing, and neither of GWCs current widgets would meet their new requirements. They needed medium-sized widgets. Recognizing both the opportunity and the risk if GWC did nothing, the CEO convened a special meeting of the Executive Committee. His staff conrmed that the current production lines were running at full capacity but there was unused space in the factory, which meant that it would be possible to add a new production line. They could get that new production line up and running in a relatively short time if they duplicated their current manufacturing process, changing nothing other than the widgets size. Although that was a possibility that GWC might have consider ed seriously several years earlier, the Executive Committee was vocal in insisting that there had to be a better way. The VP of Marketing was concerned that the medium-sized widget would be too expensive for Great Auto and Little Telecom, while the VP of Customer Relations feared that the existing quality level might not satisfy the customers new needs. The answer, the Executive Champion of Six Sigma said, was a DFSS project. The Committee, all of whom had been trained in DFSS concepts, recognized that there were risks associated with making this a DFSS project. There would be a longer implementation time because of the front-end planning involved in DFSS. Furthermore, there was a potential for disruption to the organization because the design selected through DFSS might be fundamentally different from the current manufacturing process and might demand different skills or fewer employees. The rewards, however, could be substantially higher than GWC would achieve with a replication of the current process. The Committee balanced the risks against the rewards and decided that this would be a DFSS project. Because her department would be the most heavily impacted, Veronica Major, the VP of Manufacturing, was asked to serve as the champion. GWCs rst major DFSS project had begun.

2007 by Taylor & Francis Group, LLC

138

Six Sigma Software Development

References
1. Harry, Mikel, Ph.D., and Richard Schroeder, Six Sigma: The Breakthrough Management Strategy Revolutionizing the Worlds Top Corporations. New York: Doubleday, 2000. 2. Archimedes, Pappus of Alexandria Collection, bk. VIII, prop. 10, sec. 11 from Bartletts Familiar Quotations. Boston: Little, Brown and Company, 1980, p. 93. 3. Chowdhury, Subir, Design For Six Sigma. Chicago: Dearborn Trade, 2005, p. 177. 4. Chowdhury, Subir, Design For Six Sigma. Chicago: Dearborn Trade, 2005, p. 18. 5. Stamatis, D.H., Six Sigma and Beyond: Design For Six Sigma. Boca Raton, FL: St. Lucie Press, 2003, Preface. 6. Honeywell International, Inc., 2002. 7. www.iSixSigma.com

2007 by Taylor & Francis Group, LLC

Vous aimerez peut-être aussi