Six Sigma :LESSONS LEARNED
1. The time is right.
2. The enthusiastic commitment of top management is essential.
3. Develop an infrastructure.
4. Commit top people.
5. Invest in relevant hands-on training.
6.Select initial projects to build credibility rapidly.
7. Make it all pervasive, and involve everybody.
8. Emphasize Design for Six Sigma (DFSS).
9. Don’t forget design for reliability.
10.Focus on the entire system.
11.Emphasize customer critical to quality characteristics (CTQs).
12. Include commercial quality improvement.
13.Recognize all savings.
14. Customize to meet business needs.
15. Consider the variability as well as the mean.
16. Plan to get the right data.
17. Beware of dogmatism.
18.Avoid nonessential bureaucracy.
19.Keep the toolbox vital.
20. Expect it Six Sigma to become a more silent partner.
When Jack Welch brought Six Sigma to General Electric (GE) in 1995, he proudly said he would build on the successful experiences of other organizations such as Motorola and Allied Signal.
Since then, nimble companies have introduced or improved Six Sigma by adapting best practices from GE and elsewhere. In the same vein, those who today want to make Six Sigma happen can learn from the accumulated experiences of their predecessors.
1. The Time Is Right
Six Sigma makes sense. Most Six Sigma Forum Magazine readers take this statement as self-evident. Nevertheless, it warrants repeating.
A key difference between Six Sigma and other approaches is the integration of a highly disciplined process (such as DMAIC for Define, Measure, Analyze, Improve, Control or DMADOV for Define, Measure, Analyze, Design, Optimize, Verify) with one that is very quantitative and data oriented. This is a winning combination—as evidenced by the results of the companies that have used it.
The concepts underlying Six Sigma have always made sense. What makes Six Sigma especially timely today is the combination of the following:
Intense competitive pressures, including those resulting from globalization.
Greater consumer demand for high quality products and management recognition of the cost of poor quality.
The accessibility of large databases and our ability to digest and analyze data.
The last point is especially important. I have witnessed other initiatives with some similarity to Six Sigma. Although generally moderately successful, they have floundered for two reasons. One was the inability to harness data speedily. This has changed radically with the computer revolution, in general, and the increased accessibility of user oriented software, in particular.
2: The Enthusiastic Commitment of Top Management Is Essential
This gets to the second reason many initiatives have floundered in the past. Many quality improvement efforts have been promoted by lower or middle management and took a bottom up, rather than a top down, approach.
It is unlikely Six Sigma would have succeeded at GE without CEO Jack Welch’s (and now Jeff Immelt’s) unflinching leadership. Tacit approval is not enough. What is needed is both an up-front investment in funding and a system that actively rewards successful implementation and implementers.
Welch required employees to be “lunatic” about quality and made Six Sigma a major criterion for incentive compensation and promotion. Enthusiasm spread from management to the entire workforce.
Local business leaders, called Six Sigma Champions, play an important role in making Six Sigma happen in large companies. They establish the mechanism, select the local Master Black Belts (MBBs), provide the needed resources, ensure everybody in the workforce pays attention and take overall responsibility for successful implementation and continued commitment.
3: Develop an Infrastructure
Six Sigma cannot be an informal extracurricular activity. A formal supporting infrastructure is required. Management’s commitment includes establishing and supporting such an infrastructure.
This commitment includes setting up the formal organization, defining key objectives and responsibilities, developing a budgeting process and specifying a solid process for measuring results. This process needs to be closely integrated with the existing system and must be an important element of it.
4: Commit Top People
There is a natural inclination to assign to Six Sigma people that happen to be available, rather than those that would be best suited for the job. Instead, Six Sigma merits the assignment of the most capable people to be MBBs and Black Belts (BBs). They should not only excel technically but also be imaginative and persuasive leaders.
At GE, for example, MBBs are likely candidates for subsequent high management positions. Also, But they need to be relieved of existing responsibilities so they can, at least initially, give Six Sigma their full commitment.
5: Invest in Relevant Hands-on Training
Once the structure for implementing Six Sigma is in place, the hard work follows. It begins with providing the workforce the training needed for successful implementation. The following are some key learningsSome key points about training include the following:6, 7
Ensure trainers are knowledgeable and outstanding communicators.
Customize the training, especially the examples, to the needs of the specific business. This often requires deviations from and extensions to previously developed canned materials; , as advised insee lesson 14.
Ensure the common vocabulary of Six Sigma is retained—an essential for expediting communications.
Incorporate hands-on involvement. The classical Six Sigma curriculum has four three-day sessions (approximately corresponding to DMAIC) offered four weeks apart. Six Sigma projects are conducted by the class participants during the intervening time. Green belt (GB) and BB certification requires successful completion of two Six Sigma projects, as judged by a MBB.
Engage Consider engaging external gurus to expedite Six Sigma introduction. GE followed this formula did this in introducing both DMAIC and the DMADOV processes for DFSS. Within a few years, these efforts were, however, fully transitioned to people within the company.
6: Select Initial Projects To Build Credibility Rapidly
Total management commitment, appropriate infrastructure, involvement of top people and the right training set the stage for success and result in high expectations. These expectations need to be met—and met rapidly—to maintain momentum. All Six Sigma projects need be selected judiciously. This is particularly critical for start-up projects. They need to meet the following criteria:
Their importance is evident or can be readily demonstrated.
They are viable and doable in a short time (preferably less than three months).
Their success can be readily quantified.
This often translates into addressing the proverbial low-hanging fruit. Initial Six Sigma applications have frequently been in manufacturing, for which the DMAIC process is especially suitable. Key metrics of success, such as the reduction of end of line rejects as measured by scrap and rework, can generally be rapidly impacted in manufacturing and are usually well-publicized and well recorded.
A successful start, built on accepted and highly demonstrable successes, will expedite the path forward.
7: Make It All Pervasive, and Involve Everybody
Once credibility for Six Sigma is established, you need to leverage the momentum. The goal should be to make Six Sigma all pervasive within the organization and beyond.8 The initial distinction between Six Sigma projects and other projects should disappear as Six Sigma becomes “the way we work.”
Short-term successes in manufacturing generally represent just the tip of the iceberg. Six Sigma also must include vendors (product can be only as good as its raw materials) and customers (more on that starting in lesson nine).
The all pervasiveness of Six Sigma is reflected in many of the subsequent lessons learned.
8: Emphasize DFSS
The quality of a product is fundamentally determined by its design. What can be done in manufacturing is generally reactive and limited., and the In contrast, the impact of high quality in design goes well beyond improving end of line yield. It can also heavily influence the quality experienced by the customer. Thus, DFSS has become an integral element of Six Sigma, with long-term payoffs that may well exceed those in manufacturing.9, 10
The importance of DFSS is being recognized in the The training of those who can impact design is recognizing this importance of DFSS. Thus, the basic implementation process is DMADOV, with its emphasis on design, in place of the closely related DMAIC, which emphasizes improvement. Implementation has rapidly followed.
For example, according to GE’s 2000 annual report, its medical systems business alone introduced 22 DFSS products that year. Most significant among these were the Senographe and Innova proprietary digital x-ray systems, claimed to revolutionize breast cancer detection and interventional cardiac imaging.
An important element of DFSS is robustness of design—ensuring the product performs well under diverse operational environments (from the tropics to the Artic), varied customer use (and misuse) scenarios and subtle changes in production conditions (such as variability in raw materials).
9: Don’t Forget Design for Reliability
Design quality has both short- and long-term elements. Short-term quality is reflected by customer delight in on-time delivery and initial experience with the product, but most major products (for example, automobiles, washing machines, locomotives and aircraft engines) are purchased for the long haul.
For such products, quality is reflected by high reliability (problem free operation) over many years. Such long-term performance of a product is most vivid in the minds of customers at the time of product replacement.
Thus, a key goal in an all pervasive Six Sigma program is to design products for long life and high reliability. Methods for reliability improvement need to be an integral part of the Six Sigma toolset for design and product engineers.
10: Focus on the Entire System
An important part of making Six Sigma all pervasive is to focus on the entire system. Initial Six Sigma projects traditionally address individual critical to quality characteristics (CTQs), such as the elimination or drastic reduction in the occurrence rate of a specific defect.
In an uncompromising attack on the CTQ at hand, little attention might be given to other CTQ’s. In many cases, such as ones dealing with the removal of defects, projects that improve one CTQ will also improve others, but this is not necessarily so.
As an extreme example, Oone way to decrease the rejection rate is might be to broaden product specifications. This could have a deleterious impact on other CTQs—such as the reliability of the product in the field. Moreover, while the positive impact of yield improvement is usually evident, the possible negative effect on reliability is often delayed, and, therefore, may not be recognized readily..
You need to recognize that Because CTQs are frequently closely related, so you must evaluate the impact and associated cost of any action on all important CTQs. The focus needs to be More generally, you need to focus projects on overall system improvement rather than on the improvement of any individual CTQ.
11: Emphasize Customer CTQ’s
The initial successes at GE and elsewhere were due principally to involved internal improvements that even though they had a significant impact on the bottom line. However, Ccustomers felt left out. They were quoted by Welch and in the GE 1998 annual report, as asking “When do I get to see the benefits of Six Sigma?”
This question led to the introduction of the outside in thinking concept, with the goal of focusing on customer directed CTQs, such as time to delivery, waiting time to respond to customer enquiries and general customer satisfaction.
These concepts were recently extended by GE through its ACFC (at the customer, for the customer) initiative. Its purpose is to train customers in Six Sigma and help them in the start-up process, including guiding them on specific projects.
Interestingly, GE’s annual reports from 1996 to 1999 quantified the mostly internal savings from Six Sigma (for example, $2 billion for a $550 millionin 1999). In contrast, the GE 2000 annual report emphasized savings to the customer.
In contrast, the GE 2000 annual report emphasized savings to the customer. For example, GE helped airlines undertake more than 1,2000 customer Six Sigma projects, realizing more than $300 million in savings. The company also generated more than $100 million in benefits for medical systems customers, serving as a catalyst for 1,000 Six Sigma projects aimed at improving patient throughput and reducing variability in healthcare delivery.
12: Include Commercial Quality Improvement
The broadening of Six Sigma to encompass commercial and transactional quality almost coincided with GE’s drive to address customer CTQs. Successful applications includeddealt with portfolio acquisition and management, pricing, marketing strategy and collections. Again the basic design of such processes (DFSS for commercial quality), as well as the improvement of existing processes, was emphasized.
It soon became evident evident that these concepts provided payoffs to backroom operations as well as to customers, matching that could match those from product design and improvement.11
13: Recognize All Savings
A significant element of Six Sigma is the quantifying of hard number savings. The relatively easy and speedy quantification of dollar savings is, as already suggested, why many companies focus initial Six Sigma projects on manufacturing.
But here’s the rub: How can you make Six Sigma all pervasive, systems oriented and customer responsive if you cannot readily achieve financial recognition of the resulting gains? In particular, how can you quantify the added sales resulting from customer goodwill and word of mouth recommendations created by DFSS or improved commercial quality?
Even more complex, how can you ensure proper recognition of the savings from averting problems resulting from DFSS, such as a design change that significantly reduces the probability of a product failing during its first five years of operation?
Best practices for answering these questions are needed. One approach might be to establish CTQs that, in their own right, recognize such factors as premature failure. Clearly, for Six Sigma to be all encompassing youwe need to develop ways of recognizing all savings and loss avoidances, including long-term ones—even if those savings cannot be immediately or even ever quantified precisely by the finance department.
14: Customize To Meet Business Needs
The basic concepts of Six Sigma, such as a highly disciplined data oriented approach to quality improvement, are directly applicable to all operations, but the relative importance of specific tools varies from one operation and one business to another.
This variance needs to be recognized in keeping training fully relevant as Six Sigma moves from the manufacturing floor to other functions. For example, the proper planning of investigations is universally relevant. For most product development and manufacturing situations, therefore, design of experiments (DOE) merits close attention. In addition, for chemical processing businesses, mixture experiments (evaluating the impact of product ingredients that need to add up to 100%) also warrant consideration.
In contrast, in marketing or servicing operations you might provide only minimal training in formal DOE and instead focus on the planning and conduct of customer surveys.
It is also important to adjust the tools to the sophistication of those involved and remain cognizant of practitioners’ ability to learn these while pushing forward new and improved approaches.
15: Consider the Variability as Well as the Mean
As Six Sigma has matured, there has been increased recognition of the importance of improving the mean and reducing variability, in addition to the traditional goal of improving the mean. “Variability is evil” has become a a Six Sigma slogan. A typical application is the reduction of variability in the time from the placement of an order for a product or service to its delivery.
A typical application is the reduction of variability in the time from the placement of an order for a product or service to delivery.
Some even claim variability, not the mean, is what principally impacts customers. I think this is an overstatement.
For example, Ccustomers seek a consistently long life for a product. A consistently short life would generally not do. Neither would a long life for some units, and early failure for others. In this and many other applications, both the mean and variability are important and need to be addressed. Their relative importance varies from one application to the next.
There are some situations for which variability is inherent and can only be partially eliminated. In teaching the workforce to perform specified tasks, for example, you can reduce, but not completely remove, variability by judicious hiring practices and targeted training. The
Some variability due to inherent differences between workers will likely remain. The challenge then becomes that of helping ensure the process is maximally robust to such remaining variability.
16: Plan To Get the Right Data
A Six Sigma slogan goes, “In God we trust, all else others bring data.” Unfortunately, the data that are readily available, even though sometimes voluminous, are often insufficient for the task at hand. Frequently, this is because the data were obtained for reasons other than analysis, such as to meet accounting requirements. In fact, the most relevant information might not even be recorded, or the data that are available may be inconsistent or ambiguous.
Say, for example, you wish to learn about the root causes of field failures that have not been eliminated by DFSS. Data may be available only on units that failed under warranty, and nothing may be known about the others those that failed out of warranty. In addition, there may be little information on unfailed units. Even for
Even for units with failure data, knowledge may be limited to the date of failure, the cost to the manufacturer and perhaps a general description of the failure.
Detailed failure descriptions may be at the discretion of repairpersons and thus can be highly inconsistent. Important information about such factors such as production date and shift, time or number of cycles in service, operating environment, degradation of the unit= and the specific cause of failure are often missing.
Thus, it is often necessary to obtain added information, even though When added information is required, this costs money and takes time.
The data that are available may be useful in helping decide what added information would be most useful. Data Data on every unit is not always needed. Instead, a carefully selected sample often suffices. This is because it is the quality rather than the quantity of the data that counts.
In obtaining a better understanding of refrigerator failures, for example, good information on a statistically selected sample of 10,000 units is generally more useful than incomplete and inconsistent data on a million units.
Because there is generally a significant delay between planning the data acquisition system and getting the data, it is important to establish a good data procurement system proactively, rather than wait until it is required by a problem. In some situations, a well-designed experiment can also provide needed information.
17: Beware of Dogmatism
Six Sigma concepts and tools are sound and have proven their worth. You can apply them with confidence, but you often need to adapt them they often need to be adapted to be maximally responsive to the issue at hand. For example, the concept of 3.4 defects per opportunity can sometimes be vague. And how do you define an “opportunity?” Does a shift of exactly 1.5 sigma in the mean from the short-term to the long-term really always apply?
Instead of debating these questions, it may be better to directly use the percentage of defect in the final product or in the process as your criterion. Also, you should not insist on the use of any specific tool from the Six Sigma toolkit.
The goal is to gain significant quality improvement—not to use DOE, gage repeatability and reproducibility (R&R) or whatever just for the sake of using it. These tools are very powerful and have much applicability, but they are not equally relevant for all situations and to need to be used judiciously.
18: Avoid Nonessential Bureaucracy
One of GE’s key initiatives prior to Six Sigma was “work out.” Its goal was to do away with bureaucracy and all else that impeded progress. Welch surprised many employees in introducing Six Sigma when he urged proceeding even if it might sometimes result in adding a small amount of bureaucracy.
This addition, indeed, is a price one pays for a highly disciplined process (such as DMAIC, tollgates and establishment of an infrastructure). It has, however, resulted in some criticism, especially from those in the trenches who perhaps experience some frustration in the added work or possible delay that might result from Six Sigma.
Bureaucracy, moreover, can build on itself. You need to be vigilant to ensure only the most essential bureaucratic elements are added and to carefully scrutinize these to keep them to an absolute minimum. As Six Sigma develops into the way you work, you should even be able to chip away at the bureaucracy that remains. some Elementsese.
19: Keep the Toolbox Vital
Statistical and other tools form the backbone of Six Sigma. These often involve technical concepts—such as emphasis on the measurement system (gage R&R), DOE, quality function deployment—that are important in practice but are unlikely to be encountered in standard college training.
What has helped make Six Sigma powerful is the tools being imbedded into a process and immediately put to practical use. Six Sigma is therefore much more than a toolbox. With experience, some tools have been found to be more important than others, and some not included originally have been found to warrant addition.12, 13
Statistical topics not included in the original training that have proven their value include:
Handling of nonnormal data.
DOE as a step-by-step process.
Simulation analysis (especially for determining sample size).
Quantifying individual sources of variability.
Expressing uncertainty by statistical intervals in place of hypothesis (or significance) testing.
Analyzing product life data using special methods.
Tools for analyzing large data sets, such as CART (classification and regression trees) and MARS (multivariate adaptive regression splines).
Graphical methods for supplementing formal statistical analyses.
20: Expect Six Sigma It To Become a More Silent Partner
The introduction of Six Sigma is often accompanied by much fanfare. This needs to be rapidly followed by solid evidence of successful applications resulting in important quality improvements. The concepts are then extended to broader, longer-term, more systems focused and customer oriented improvement challenges.
As Six Sigma becomes engrained within an organization and applied to vendors and customers, it may no longer be front-page news. This should not come as a surprise and certainly should not imply it is any less relevant.
The momentum can be retained by continued successful and broader applications. Best practices need to be sought and implemented as Six Sigma truly becomes the way we you work.
Where Six Sigma Is Headed
Six Sigma got started and has been applied, with much success, in large corporations—perhaps because these have the greatest resources. But it is clearly adaptable to the needs of medium-sized and small businesses.
Its relevance, moreover, is not limited to the corporate world. It has, for example been applied recently to farming. and is even being used on a farm.14 Six Sigma can also be valuable to the public sector, including government, hospitals, nonprofit organizations and schools.
Six Sigma’s deployment in these new arenas represents its new frontier and presents some important opportunities and exciting challenges. I hope the pointers I’ve presented provided will help expedite successful jump-starts in many new and diverse applications.
1. The time is right.
2. The enthusiastic commitment of top management is essential.
3. Develop an infrastructure.
4. Commit top people.
5. Invest in relevant hands-on training.
6.Select initial projects to build credibility rapidly.
7. Make it all pervasive, and involve everybody.
8. Emphasize Design for Six Sigma (DFSS).
9. Don’t forget design for reliability.
10.Focus on the entire system.
11.Emphasize customer critical to quality characteristics (CTQs).
12. Include commercial quality improvement.
13.Recognize all savings.
14. Customize to meet business needs.
15. Consider the variability as well as the mean.
16. Plan to get the right data.
17. Beware of dogmatism.
18.Avoid nonessential bureaucracy.
19.Keep the toolbox vital.
20. Expect it Six Sigma to become a more silent partner.
When Jack Welch brought Six Sigma to General Electric (GE) in 1995, he proudly said he would build on the successful experiences of other organizations such as Motorola and Allied Signal.
Since then, nimble companies have introduced or improved Six Sigma by adapting best practices from GE and elsewhere. In the same vein, those who today want to make Six Sigma happen can learn from the accumulated experiences of their predecessors.
1. The Time Is Right
Six Sigma makes sense. Most Six Sigma Forum Magazine readers take this statement as self-evident. Nevertheless, it warrants repeating.
A key difference between Six Sigma and other approaches is the integration of a highly disciplined process (such as DMAIC for Define, Measure, Analyze, Improve, Control or DMADOV for Define, Measure, Analyze, Design, Optimize, Verify) with one that is very quantitative and data oriented. This is a winning combination—as evidenced by the results of the companies that have used it.
The concepts underlying Six Sigma have always made sense. What makes Six Sigma especially timely today is the combination of the following:
Intense competitive pressures, including those resulting from globalization.
Greater consumer demand for high quality products and management recognition of the cost of poor quality.
The accessibility of large databases and our ability to digest and analyze data.
The last point is especially important. I have witnessed other initiatives with some similarity to Six Sigma. Although generally moderately successful, they have floundered for two reasons. One was the inability to harness data speedily. This has changed radically with the computer revolution, in general, and the increased accessibility of user oriented software, in particular.
2: The Enthusiastic Commitment of Top Management Is Essential
This gets to the second reason many initiatives have floundered in the past. Many quality improvement efforts have been promoted by lower or middle management and took a bottom up, rather than a top down, approach.
It is unlikely Six Sigma would have succeeded at GE without CEO Jack Welch’s (and now Jeff Immelt’s) unflinching leadership. Tacit approval is not enough. What is needed is both an up-front investment in funding and a system that actively rewards successful implementation and implementers.
Welch required employees to be “lunatic” about quality and made Six Sigma a major criterion for incentive compensation and promotion. Enthusiasm spread from management to the entire workforce.
Local business leaders, called Six Sigma Champions, play an important role in making Six Sigma happen in large companies. They establish the mechanism, select the local Master Black Belts (MBBs), provide the needed resources, ensure everybody in the workforce pays attention and take overall responsibility for successful implementation and continued commitment.
3: Develop an Infrastructure
Six Sigma cannot be an informal extracurricular activity. A formal supporting infrastructure is required. Management’s commitment includes establishing and supporting such an infrastructure.
This commitment includes setting up the formal organization, defining key objectives and responsibilities, developing a budgeting process and specifying a solid process for measuring results. This process needs to be closely integrated with the existing system and must be an important element of it.
4: Commit Top People
There is a natural inclination to assign to Six Sigma people that happen to be available, rather than those that would be best suited for the job. Instead, Six Sigma merits the assignment of the most capable people to be MBBs and Black Belts (BBs). They should not only excel technically but also be imaginative and persuasive leaders.
At GE, for example, MBBs are likely candidates for subsequent high management positions. Also, But they need to be relieved of existing responsibilities so they can, at least initially, give Six Sigma their full commitment.
5: Invest in Relevant Hands-on Training
Once the structure for implementing Six Sigma is in place, the hard work follows. It begins with providing the workforce the training needed for successful implementation. The following are some key learningsSome key points about training include the following:6, 7
Ensure trainers are knowledgeable and outstanding communicators.
Customize the training, especially the examples, to the needs of the specific business. This often requires deviations from and extensions to previously developed canned materials; , as advised insee lesson 14.
Ensure the common vocabulary of Six Sigma is retained—an essential for expediting communications.
Incorporate hands-on involvement. The classical Six Sigma curriculum has four three-day sessions (approximately corresponding to DMAIC) offered four weeks apart. Six Sigma projects are conducted by the class participants during the intervening time. Green belt (GB) and BB certification requires successful completion of two Six Sigma projects, as judged by a MBB.
Engage Consider engaging external gurus to expedite Six Sigma introduction. GE followed this formula did this in introducing both DMAIC and the DMADOV processes for DFSS. Within a few years, these efforts were, however, fully transitioned to people within the company.
6: Select Initial Projects To Build Credibility Rapidly
Total management commitment, appropriate infrastructure, involvement of top people and the right training set the stage for success and result in high expectations. These expectations need to be met—and met rapidly—to maintain momentum. All Six Sigma projects need be selected judiciously. This is particularly critical for start-up projects. They need to meet the following criteria:
Their importance is evident or can be readily demonstrated.
They are viable and doable in a short time (preferably less than three months).
Their success can be readily quantified.
This often translates into addressing the proverbial low-hanging fruit. Initial Six Sigma applications have frequently been in manufacturing, for which the DMAIC process is especially suitable. Key metrics of success, such as the reduction of end of line rejects as measured by scrap and rework, can generally be rapidly impacted in manufacturing and are usually well-publicized and well recorded.
A successful start, built on accepted and highly demonstrable successes, will expedite the path forward.
7: Make It All Pervasive, and Involve Everybody
Once credibility for Six Sigma is established, you need to leverage the momentum. The goal should be to make Six Sigma all pervasive within the organization and beyond.8 The initial distinction between Six Sigma projects and other projects should disappear as Six Sigma becomes “the way we work.”
Short-term successes in manufacturing generally represent just the tip of the iceberg. Six Sigma also must include vendors (product can be only as good as its raw materials) and customers (more on that starting in lesson nine).
The all pervasiveness of Six Sigma is reflected in many of the subsequent lessons learned.
8: Emphasize DFSS
The quality of a product is fundamentally determined by its design. What can be done in manufacturing is generally reactive and limited., and the In contrast, the impact of high quality in design goes well beyond improving end of line yield. It can also heavily influence the quality experienced by the customer. Thus, DFSS has become an integral element of Six Sigma, with long-term payoffs that may well exceed those in manufacturing.9, 10
The importance of DFSS is being recognized in the The training of those who can impact design is recognizing this importance of DFSS. Thus, the basic implementation process is DMADOV, with its emphasis on design, in place of the closely related DMAIC, which emphasizes improvement. Implementation has rapidly followed.
For example, according to GE’s 2000 annual report, its medical systems business alone introduced 22 DFSS products that year. Most significant among these were the Senographe and Innova proprietary digital x-ray systems, claimed to revolutionize breast cancer detection and interventional cardiac imaging.
An important element of DFSS is robustness of design—ensuring the product performs well under diverse operational environments (from the tropics to the Artic), varied customer use (and misuse) scenarios and subtle changes in production conditions (such as variability in raw materials).
9: Don’t Forget Design for Reliability
Design quality has both short- and long-term elements. Short-term quality is reflected by customer delight in on-time delivery and initial experience with the product, but most major products (for example, automobiles, washing machines, locomotives and aircraft engines) are purchased for the long haul.
For such products, quality is reflected by high reliability (problem free operation) over many years. Such long-term performance of a product is most vivid in the minds of customers at the time of product replacement.
Thus, a key goal in an all pervasive Six Sigma program is to design products for long life and high reliability. Methods for reliability improvement need to be an integral part of the Six Sigma toolset for design and product engineers.
10: Focus on the Entire System
An important part of making Six Sigma all pervasive is to focus on the entire system. Initial Six Sigma projects traditionally address individual critical to quality characteristics (CTQs), such as the elimination or drastic reduction in the occurrence rate of a specific defect.
In an uncompromising attack on the CTQ at hand, little attention might be given to other CTQ’s. In many cases, such as ones dealing with the removal of defects, projects that improve one CTQ will also improve others, but this is not necessarily so.
As an extreme example, Oone way to decrease the rejection rate is might be to broaden product specifications. This could have a deleterious impact on other CTQs—such as the reliability of the product in the field. Moreover, while the positive impact of yield improvement is usually evident, the possible negative effect on reliability is often delayed, and, therefore, may not be recognized readily..
You need to recognize that Because CTQs are frequently closely related, so you must evaluate the impact and associated cost of any action on all important CTQs. The focus needs to be More generally, you need to focus projects on overall system improvement rather than on the improvement of any individual CTQ.
11: Emphasize Customer CTQ’s
The initial successes at GE and elsewhere were due principally to involved internal improvements that even though they had a significant impact on the bottom line. However, Ccustomers felt left out. They were quoted by Welch and in the GE 1998 annual report, as asking “When do I get to see the benefits of Six Sigma?”
This question led to the introduction of the outside in thinking concept, with the goal of focusing on customer directed CTQs, such as time to delivery, waiting time to respond to customer enquiries and general customer satisfaction.
These concepts were recently extended by GE through its ACFC (at the customer, for the customer) initiative. Its purpose is to train customers in Six Sigma and help them in the start-up process, including guiding them on specific projects.
Interestingly, GE’s annual reports from 1996 to 1999 quantified the mostly internal savings from Six Sigma (for example, $2 billion for a $550 millionin 1999). In contrast, the GE 2000 annual report emphasized savings to the customer.
In contrast, the GE 2000 annual report emphasized savings to the customer. For example, GE helped airlines undertake more than 1,2000 customer Six Sigma projects, realizing more than $300 million in savings. The company also generated more than $100 million in benefits for medical systems customers, serving as a catalyst for 1,000 Six Sigma projects aimed at improving patient throughput and reducing variability in healthcare delivery.
12: Include Commercial Quality Improvement
The broadening of Six Sigma to encompass commercial and transactional quality almost coincided with GE’s drive to address customer CTQs. Successful applications includeddealt with portfolio acquisition and management, pricing, marketing strategy and collections. Again the basic design of such processes (DFSS for commercial quality), as well as the improvement of existing processes, was emphasized.
It soon became evident evident that these concepts provided payoffs to backroom operations as well as to customers, matching that could match those from product design and improvement.11
13: Recognize All Savings
A significant element of Six Sigma is the quantifying of hard number savings. The relatively easy and speedy quantification of dollar savings is, as already suggested, why many companies focus initial Six Sigma projects on manufacturing.
But here’s the rub: How can you make Six Sigma all pervasive, systems oriented and customer responsive if you cannot readily achieve financial recognition of the resulting gains? In particular, how can you quantify the added sales resulting from customer goodwill and word of mouth recommendations created by DFSS or improved commercial quality?
Even more complex, how can you ensure proper recognition of the savings from averting problems resulting from DFSS, such as a design change that significantly reduces the probability of a product failing during its first five years of operation?
Best practices for answering these questions are needed. One approach might be to establish CTQs that, in their own right, recognize such factors as premature failure. Clearly, for Six Sigma to be all encompassing youwe need to develop ways of recognizing all savings and loss avoidances, including long-term ones—even if those savings cannot be immediately or even ever quantified precisely by the finance department.
14: Customize To Meet Business Needs
The basic concepts of Six Sigma, such as a highly disciplined data oriented approach to quality improvement, are directly applicable to all operations, but the relative importance of specific tools varies from one operation and one business to another.
This variance needs to be recognized in keeping training fully relevant as Six Sigma moves from the manufacturing floor to other functions. For example, the proper planning of investigations is universally relevant. For most product development and manufacturing situations, therefore, design of experiments (DOE) merits close attention. In addition, for chemical processing businesses, mixture experiments (evaluating the impact of product ingredients that need to add up to 100%) also warrant consideration.
In contrast, in marketing or servicing operations you might provide only minimal training in formal DOE and instead focus on the planning and conduct of customer surveys.
It is also important to adjust the tools to the sophistication of those involved and remain cognizant of practitioners’ ability to learn these while pushing forward new and improved approaches.
15: Consider the Variability as Well as the Mean
As Six Sigma has matured, there has been increased recognition of the importance of improving the mean and reducing variability, in addition to the traditional goal of improving the mean. “Variability is evil” has become a a Six Sigma slogan. A typical application is the reduction of variability in the time from the placement of an order for a product or service to its delivery.
A typical application is the reduction of variability in the time from the placement of an order for a product or service to delivery.
Some even claim variability, not the mean, is what principally impacts customers. I think this is an overstatement.
For example, Ccustomers seek a consistently long life for a product. A consistently short life would generally not do. Neither would a long life for some units, and early failure for others. In this and many other applications, both the mean and variability are important and need to be addressed. Their relative importance varies from one application to the next.
There are some situations for which variability is inherent and can only be partially eliminated. In teaching the workforce to perform specified tasks, for example, you can reduce, but not completely remove, variability by judicious hiring practices and targeted training. The
Some variability due to inherent differences between workers will likely remain. The challenge then becomes that of helping ensure the process is maximally robust to such remaining variability.
16: Plan To Get the Right Data
A Six Sigma slogan goes, “In God we trust, all else others bring data.” Unfortunately, the data that are readily available, even though sometimes voluminous, are often insufficient for the task at hand. Frequently, this is because the data were obtained for reasons other than analysis, such as to meet accounting requirements. In fact, the most relevant information might not even be recorded, or the data that are available may be inconsistent or ambiguous.
Say, for example, you wish to learn about the root causes of field failures that have not been eliminated by DFSS. Data may be available only on units that failed under warranty, and nothing may be known about the others those that failed out of warranty. In addition, there may be little information on unfailed units. Even for
Even for units with failure data, knowledge may be limited to the date of failure, the cost to the manufacturer and perhaps a general description of the failure.
Detailed failure descriptions may be at the discretion of repairpersons and thus can be highly inconsistent. Important information about such factors such as production date and shift, time or number of cycles in service, operating environment, degradation of the unit= and the specific cause of failure are often missing.
Thus, it is often necessary to obtain added information, even though When added information is required, this costs money and takes time.
The data that are available may be useful in helping decide what added information would be most useful. Data Data on every unit is not always needed. Instead, a carefully selected sample often suffices. This is because it is the quality rather than the quantity of the data that counts.
In obtaining a better understanding of refrigerator failures, for example, good information on a statistically selected sample of 10,000 units is generally more useful than incomplete and inconsistent data on a million units.
Because there is generally a significant delay between planning the data acquisition system and getting the data, it is important to establish a good data procurement system proactively, rather than wait until it is required by a problem. In some situations, a well-designed experiment can also provide needed information.
17: Beware of Dogmatism
Six Sigma concepts and tools are sound and have proven their worth. You can apply them with confidence, but you often need to adapt them they often need to be adapted to be maximally responsive to the issue at hand. For example, the concept of 3.4 defects per opportunity can sometimes be vague. And how do you define an “opportunity?” Does a shift of exactly 1.5 sigma in the mean from the short-term to the long-term really always apply?
Instead of debating these questions, it may be better to directly use the percentage of defect in the final product or in the process as your criterion. Also, you should not insist on the use of any specific tool from the Six Sigma toolkit.
The goal is to gain significant quality improvement—not to use DOE, gage repeatability and reproducibility (R&R) or whatever just for the sake of using it. These tools are very powerful and have much applicability, but they are not equally relevant for all situations and to need to be used judiciously.
18: Avoid Nonessential Bureaucracy
One of GE’s key initiatives prior to Six Sigma was “work out.” Its goal was to do away with bureaucracy and all else that impeded progress. Welch surprised many employees in introducing Six Sigma when he urged proceeding even if it might sometimes result in adding a small amount of bureaucracy.
This addition, indeed, is a price one pays for a highly disciplined process (such as DMAIC, tollgates and establishment of an infrastructure). It has, however, resulted in some criticism, especially from those in the trenches who perhaps experience some frustration in the added work or possible delay that might result from Six Sigma.
Bureaucracy, moreover, can build on itself. You need to be vigilant to ensure only the most essential bureaucratic elements are added and to carefully scrutinize these to keep them to an absolute minimum. As Six Sigma develops into the way you work, you should even be able to chip away at the bureaucracy that remains. some Elementsese.
19: Keep the Toolbox Vital
Statistical and other tools form the backbone of Six Sigma. These often involve technical concepts—such as emphasis on the measurement system (gage R&R), DOE, quality function deployment—that are important in practice but are unlikely to be encountered in standard college training.
What has helped make Six Sigma powerful is the tools being imbedded into a process and immediately put to practical use. Six Sigma is therefore much more than a toolbox. With experience, some tools have been found to be more important than others, and some not included originally have been found to warrant addition.12, 13
Statistical topics not included in the original training that have proven their value include:
Handling of nonnormal data.
DOE as a step-by-step process.
Simulation analysis (especially for determining sample size).
Quantifying individual sources of variability.
Expressing uncertainty by statistical intervals in place of hypothesis (or significance) testing.
Analyzing product life data using special methods.
Tools for analyzing large data sets, such as CART (classification and regression trees) and MARS (multivariate adaptive regression splines).
Graphical methods for supplementing formal statistical analyses.
20: Expect Six Sigma It To Become a More Silent Partner
The introduction of Six Sigma is often accompanied by much fanfare. This needs to be rapidly followed by solid evidence of successful applications resulting in important quality improvements. The concepts are then extended to broader, longer-term, more systems focused and customer oriented improvement challenges.
As Six Sigma becomes engrained within an organization and applied to vendors and customers, it may no longer be front-page news. This should not come as a surprise and certainly should not imply it is any less relevant.
The momentum can be retained by continued successful and broader applications. Best practices need to be sought and implemented as Six Sigma truly becomes the way we you work.
Where Six Sigma Is Headed
Six Sigma got started and has been applied, with much success, in large corporations—perhaps because these have the greatest resources. But it is clearly adaptable to the needs of medium-sized and small businesses.
Its relevance, moreover, is not limited to the corporate world. It has, for example been applied recently to farming. and is even being used on a farm.14 Six Sigma can also be valuable to the public sector, including government, hospitals, nonprofit organizations and schools.
Six Sigma’s deployment in these new arenas represents its new frontier and presents some important opportunities and exciting challenges. I hope the pointers I’ve presented provided will help expedite successful jump-starts in many new and diverse applications.
Posts
No comments:
Post a Comment