Elimination of 'Failure' Causes Confusion
I was quite amused and a little disappointed regarding the explanation that corrective action is a form of preventive action in the September 1999 ("Preventive, Corrective Definitions Do Differ," Debra Dyer, p. 14) issue of Quality Progress. It reminded me of two similar events that happened in the past.
In the early 1960s, computer systems personnel liked to develop their own standards without any regard for other standards. Their biggest faux pas had to do with differentiating between the number zero (0) and the letter O in writing programs. As a ham radio operator, I followed the universal standard and put a slash through the zero (Ø) to conduct my statistical analysis. I was then cautioned by computer systems personnel that I must follow computer protocol, so I showed them that they weren't following the universal standard for writing and printing letters and numbers. They were amazed, but it still took them three years to correct their mistake.
Also in the early 1960s, the U.S. government actually published a quality system standard that identified quality control as a management system process. The government identified the inspection and testing process as quality assurance because it assured that the agreed-on quality was present in the product or service delivered. Then came MIL-Q-9858 and the old system standard disappeared. I know that original system standard existed because I used to show it to my students.
The present situation regarding preventive and corrective action could have been prevented if the quality profession hadn't gone conservative and eliminated one important word. Preventive and corrective action used to be called failure prevention and failure correction. These terms later became failure preventive action and failure corrective action. The latter contained the need for contingency action. Of course, that was a long time ago. Maybe we need to take another look at our quality assurance process. It's never too late!
EDWARD M. DUKE
San Jose, CA
Emphasis on Ethics Is Exciting to Reader
I have been an ASQ member for over 30 years, and participating in the evolution of quality methodology has been an exciting trip for me. The current emphasis on ethics, however, is the most exciting phase I have yet to witness. It's been a long time coming.
The article "The Call of Quality: Doing Right Things Right" in the September 1999 issue (March Laree Jacques, p. 48) is an excellent treatise on the new and recurring pressures placed on those making decisions to make ethical ones. However, the idea that a tool is needed to motivate people is contrary to my experience. Age-old management theory states that motivation is the result of placing individuals in environments conducive to unleashing their inherent motivation. The honest, ethical and productive effort emanating from this environment will surpass the effort produced by any reward and punishment system. Sometimes self- motivated people simply develop their own intrinsic reward systems in order to survive when challenged with really bad management systems.
Similarly, the article "On Leaders and Leadership" in the October 1999 issue (Rick L. Edgeman, Su Mi Park Dahlgaard, Jens J. Dahlgaard and Franz Scherer, p. 49) presented an excellent model for systematic leadership. The study and its presentation were humbling in their excellence. One of the biblical lesson comparisons, however, makes a seriously erroneous assumption. It was not the wicked workers who should have been removed by corporate downsizing. Removal of the evil leaders in industry would reveal that the workers were forced to bury their motivation, caring, creativity, ability and drive. They needed to be free to unleash their true power. They were not wicked.
It is amazing how many leaders do not recognize the destructiveness of unethical strategies. Good business bases all interactions on ethical values.
BILLIE R. MARCUN
New Millennium Begins On January 1, 2001
Miles Maguire's Up Front column in the December 1999 issue of Quality Progress (p. 6) does a great service to those of us who will correctly celebrate the millennium on January 1, 2001.
Maybe Maguire can also help set the record straight on another misconception: the Y2K problem. It's Y2k, not Y2K! A capital K is the international symbol for the Kelvin temperature scale and not the symbol for 1000s as we've been led to believe by almost every major newspaper, magazine and television show.
Six Sigma Is Actually Nuts and Bolts Quality
I read Jay Velury's letter ("ASQ Should Stop Perpetuating Fads," December 1999, p. 8) regarding the article "Cowboy Quality" (Miles Maguire, October 1999, p. 27) with interest. I am a Six Sigma Black Belt but learned most of the course's content prior to taking it. While certain aspects of Six Sigma may seem faddish, the nuts and bolts of the applications are solid quality engineering principles.
Mikel Harry's statement that you can't improve what you can't measure was not intended to apply to nebulous areas such as customer satisfaction. I think all quality professionals agree with Deming that there is no way to measure the cost of a dissatisfied customer. However, it is very difficult to improve what you can't measure in relation to process improvement. All quality professionals are at times faced with the dictum to improve quality when quality is not clearly defined. If what is measured is trivial, why is anyone concerned about process capability? My main goal throughout most of my quality career has been to find the relevant measures of quality so improvements can be made.
As for intrinsic motivation, I agree that quality professionals should strive to make high quality products because that is the right thing to do. But if there is no market for those products, there is no motivation to make them. Quality departments do not exist in a vacuum. They exist as a part of business that is measured. Why would any business invest in technology, methods or people if it did not believe that there would be a return on its investment in the long run?
And, yes, total quality management (TQM) is important. But without the basic process improvements needed to create robust processes, there won't be a reduction in failure costs. TQM is the framework in which process improvement can thrive. I don't think Harry was saying that TQM is useless, but rather that it's not the complete answer to being competitive in a global market.
TQM, statistical process control, Six Sigma, zero defects and many other programs could have been called fads when they were introduced. That does not mean that there is no value to gain from learning the methods. Harry does come across as a little eccentric, but he has helped many companies and people through his systematic application of proven quality engineering techniques. Unless quality professionals try to learn those techniques, how can the effectiveness of the techniques be measured?
JOHN C. FINLEY
ISO 9000 Quality System Is Not Logical To Apply
The article "Quality System Harmonization for Clinical Laboratories" (William I. White, September 1999, p. 57) discusses the good news that the Clinical Laboratory Improvement Amendments have been harmonized with the ISO 9000 requirements. While I agree with the basic premise that an ISO 9000 quality system forms a good start for a measurement laboratory, it is not the most logical system to apply.
Any discussion of the quality assurance systems applicable to measurement laboratories in any field must give space to the current ISO Guide 25 system--soon to be replaced with the ISO/IEC 17025 document. These documents were created to provide a quality assurance system for measurement laboratories. Producing Table 2 in the article would be easier if ISO 9000 requirements were replaced with the ISO/IEC 17025 requirements.
The need for synergy with ISO 9000 should be yesterday's news. Effort is best expended in promoting mutual agreement between all the regulatory and accreditation bodies on the ISO/ IEC 17025 requirements. I haven't seen any of the regulatory and accreditation body requirements, but I am sure most of them would agree with a major portion of the ISO/IEC 17025 document.
Surrey, United Kingdom
Management Success Is in the Details
I enjoyed reading the November 1999 One Good Idea column by Steven Stein ("A Recipe for Well-Done Procedures," p. 136). As the first step to take when writing a procedure, Stein recommends assembling "one subject matter expert (SME), four to six staff members, one technical writer and one well-seasoned facilitator." He would have hit a home run in his brief article if he had placed more emphasis on management's key role in writing procedures. The decisions made regarding how work is accomplished in an organization or between organizations rest solidly with management. Procedure writers are often left to coordinate and determine how to accomplish important management processes before obtaining direction from senior level management.
Finalized procedures are often coordinated at lower levels in an organization and presented to busy managers for their rubber stamp approval. This process creates managers who have a lack of knowledge regarding important management processes under their jurisdiction. Who is running the farm--SMEs or management?
JOHN P. JACKSON
San Diego, CA
Six Sigma Gives ASQ A Bad Name
I was disgusted when I saw Mikel Harry glamorized on the cover of the October 1999 issue of Quality Progress (Miles Maguire, "Cowboy Quality," p. 27). My only encounter with this hard-riding cowboy was at his booth at the Annual Quality Congress (AQC) show last spring. The booth featured what appeared to be has-been hookers sitting atop a mechanical bull. That's not exactly the kind of image you expect from a company in the quality industry, especially one that has been embraced by ASQ as a partner. (Or is that "pardner"?)
As a professional woman familiar with male dominated industries, I'm not easily offended. But as I ventured into the Six Sigma "corral," I was immediately insulted by caustic remarks from the so-called cowboys. They ridiculed the quality industry, the show's attendees and the other exhibitors at AQC. The entire cowboy routine was in bad taste, and Harry led the pack as a self-important, loudmouthed egomaniac. Following the show, I almost picked up the phone to call ASQ and voice my opinion.
If that wasn't enough, I came to find out that QP had devoted a cover story to Harry. It was all I could do to read it, but amazingly, the trumped-up editorial almost made Harry sound like a decent person. From one quality professional to another, I wouldn't let the good-old-boy act fool you. Harry knows nothing about humility and is riding tall at everyone else's expense.
I just have to say it: The boys at the Six Sigma Academy give ASQ a bad name and make a mockery out of the quality profession. If you've ever known any real cowboys, you'd know Harry is an embarrassment to the cowboy profession as well.
Sector-Specific Versions Of ISO 9000 Are Natural
I read Quality Progress regularly, including QP Mailbag. The letter from David Hutchins in the November 1999 issue ("New Changes to ISO 9000 Will Have a Large Impact," p. 8) prompted me to write.
Hutchins sees QS-9000 and other variants of ISO 9000 as alternatives about to swamp ISO 9000 unless it radically changes. He's worried that all these new choices will make quality strategy choices "chaotic for the next two or three years."
With all due respect, I think Hutchins misses the point. The basic ISO 9000 framework is a well-designed, generic quality system, and the upcoming 2000 revision will be an evolutionary improvement. ISO 9000 is a generic template for a good reason. It is an elaborate, real world checklist created for the full range of considerations to be incorporated into the design of a specific quality system for a specific enterprise.
QS-9000 is not an alternative to ISO 9000. It is a sector-specific version of ISO 9000 and was created for entirely legitimate reasons relating to the inherent nature of the auto industry. The names we all know in this industry (Ford, GM and so forth) are basically design, assembly and service firms. A large portion of the total value added in a car comes from component suppliers that ship highly engineered subsystems into car assembly plants. ISO 9000 says that for any manufacturing facility incoming raw materials' quality and supplier reliability should be monitored with programs designed for continuous improvement. But, when possibly as much as 75% of the value added is from incoming components, something more prescriptive is clearly required. Until QS-9000 each car company had its own prescriptive requirements. These requirements were similar, but they weren't identical, and this lack of harmony created chaos for the suppliers. QS-9000 creates a large degree of unity for suppliers and should be thought of as a sensible sector-specific version of ISO 9000.
I was personally involved in the early process that eventually produced a similar variant for the aerospace industry: AS 9000. Because the aerospace industry is an assembler of very high cost components from numerous suppliers, people frequently asked why the industry couldn't simply use QS-9000. The sensible answer arises from the essential differences between the two industries: Cars are produced in the hundreds of thousands, while planes and missiles are produced in the thousands and hundreds and space vehicles are produced in the tens and ones. The quality methods and techniques, the failure risk and the importance of cost are different for each industry. As in the auto industry, the generic ISO 9000 was not prescriptive enough, but in this case the QS-9000 remedy was too prescriptive. So another sector-specific quality system variant came into being. Think of both QS-9000 and AS 9000 as ISO 9000 with the supplier section bulked up with different steroids.
On another note, the National Sanitation Foundation merged the prescriptive requirements of Hazard Analysis Critical Control Point (HACCP) with the ISO 9000 format to create HACCP 9000. This makes excellent sense for companies in the food processing industry. Also, ISO Guide 25 is emerging as a more appropriate standard for special issues related to testing and certification companies. As with QS-9000 and AS 9000, these tailored variants of the basic ISO 9000 template are positive developments--they extend the ISO 9000 concept and demonstrate its power.
Any qualified ISO 9000 auditor knows if the enterprise being audited just installed its quality system to get a certificate or intelligently applied the template to create a quality system designed to support and reinforce the unique nature of its business. When entire industries come together and use the ISO 9000 template to create a more meaningful shared system, we should expect better industry results and real benefits for clients and other stakeholders. Sector-specific versions of ISO 9000 are a natural phenomenon and will continue to spread.
ERIC N. BIRCH
QA, QC Staff Are Critics, Not Playwrights
The subtitle--"It's easier to be a critic than a playwright"--to the article "The Art of Creating a Standard" (Charles A. Cianfrani, December 1999, p. 59) is quite amusing and ironic. By definition quality assurance (QA) and quality control (QC) staff members do not manufacture products. They only evaluate and attempt to persuade others to improve their efforts; therefore, all QA and QC staff members are critics, not playwrights. While this may be an unpopular concept in QA and QC land, it is still true. When QA and QC staff members are actually responsible for getting the product or service out the door, they will become manufacturers or playwrights.
The real lament Cianfrani and the other authors of the article have is that many people looked at what was written and found that it was sadly lacking. This distresses the authors, and I can feel their pain. Unfortunately, they should have listened more to the customers than to themselves.
The customers for the new ISO 9000 standard are the firms that actually pay to have the standard installed, not the customers of the firms or the QA and QC staff in the firms. If the new ISO 9000:2000 standard does not reduce costs by several multiples annually, then it is utterly worthless. This is a concept Cianfrani either misunderstood or ignored. ISO 9000:2000 is more complex and costly to operate and install than ISO 9000:1994. It will be impossible to persuade top management to fund these costs when there aren't any clear benefits.
A quality system must make the installing firm more efficient and profitable. If it doesn't, it should be removed or never installed at all. The article lists six major tasks that working group 18 pursued. None of the tasks involve making the installing organization more efficient.
1. Providing standards compatibility. This concept offers no financial utility to existing or prospective installers.
2. Providing management structures emphasizing customer satisfaction. This concept ignores profitability as a valid goal.
3. Accomodating all sizes and all industries. They are not all the same, and one size does not fit all. A single standard is guaranteed not to fit anyone well.
4. Connecting business process with the QA system. This is a meaningless concept since the QA and QC process evaluates existing business processes and their results. Business processes and QA are already connected.
5. Facilitating compatibility with other management systems. This is a nice goal, but failure to facilitate compatibility is certainly not a major flaw in ISO 9000:1994.
6. Providing simple and clearly understood standards. This concept was apparently discarded because ISO 9000:2000 is longer and more complex than ISO 9000:1994, and the new standard is basically incomprehensible.
Cianfrani states that these concepts were carefuly formulated, but he neglects to inform us by whom and for what reason. One thing is certain--they did not come from the management of the firms who were the installed base of ISO 9000 users. If they had, Cianfrani would have gotten a different set of results. The new standard would be simpler, clearer, less prescriptive and without jargon.
The most amusing concept in the article is the assertion that an organization shouldn't do anything just to meet a standard requirement. While this is true, it is impossible to become registered without complying with the standard. It doesn't cost a firm anything to install ISO 9000:2000 if the firm ignores the areas that aren't useful. However, this means that it may not comply with the standard, and if a firm is not going to comply with the standard then the standard is merely a toothless set of platitudes.
A major thrust of ISO 9000:2000 is to include the service industry. Consequently, some very clumsy language was inserted into the standard. The reason ISO 9000 is not popular in the service sector is that customers aren't powerful enough to mandate it. It is too complex and isn't focused on increasing the efficiency of the installing firm. ISO 9000:2000 is even worse than ISO 9000:1994. It too will be unpopular with the service sector.
It is not easy to be a competent critic. One needs intimate knowledge of the underlying concepts--what was done in other trials and how well the efforts went. This requires an expertise all its own. As a critic of ISO 9000:2000, I think it's a failure. But my judgment is not final--the market will ultimately decide on its viability.
RALPH J. TEETOR
Boca Raton, FL
Statistics Article Ignores Topics For the Future
As we approach the new year, we can expect to see many articles regarding quality for the next century. J. Stuart Hunter's discussion in Statistics Roundtable ("Statistical Process Control," December 1999, p. 54) is one such piece, though it was a disappointing one.
In his article, Hunter addresses the simplest control charts, the regular Shewhart and the exponentially weighted moving average (EWMA) for the mean and the Box-Jenkins chart for adjusting process settings given the current process output level. The descriptions given for each are adequate enough for QP's audience. But perhaps I misunderstood the purpose of the article, because the scope hardly suffices for a glimpse of the future of statistical process control (SPC). Two of the topics Hunter ignored are discussed below, but by no means is this list meant to be exhaustive.
Hunter talks about the Box-Jenkins chart, which combines SPC with engineering process control for autocorrelated processes. A lot of recent work was done in this area, but there are still plenty of areas available for further research. Hunter mentions the property of robustness for Shewhart and Box-Jenkins charts, areas that merit study with respect to charts for autocorrelated data. Such studies are necessary due to the plethora of processes that do not exhibit independence over time.
The area of multivariate SPC also warrants further study. This control chart is for shifts in the mean and covariance matrices for processes with several outputs, and it should be used in most applications. Many people argue that multivariate charts, such as Hotelling's T2 charts, reduce a complex process down to one number and oversimplify the monitoring process. Such an objection is not warranted.
One does not need to use a multivariate chart and eschew the use of Shewhart charts--they can be used together. The multivariate chart is necessary due to the correlation between output variables. Consider a process used to produce fabric, for example. The output measures of interest include fabric thickness, weight and tensile strength. A light spot in the fabric will tend to have a low weight and tensile strength. Such a combination of low ratings might not show up on the Shewhart charts but would show up on a T2 chart. Simply monitoring the individual variables potentially ignores dangerous changes in the process. The interpretation of signals in multivariate charts remains an important area for future research.
There are still other topics that deserve mention here, including variable sampling interval charts, nonparametric charts and diagnostic procedures for assessing the causes of chart signals. Again, this list is not intended to be exhaustive. The need to study these topics cannot be overstated if we are to move into the next century prepared to improve.
J. DOUGLAS BARRETT
Bravo to Barrett's concern over the multivariate and time dependent aspects of most production data. I agree that such data offer vast opportunities for research and application. However, my paper was intentionally limited to univariate data. Its purpose was to contrast models appropriate to the Shewhart chart for time independent data with those appropriate to the EWMA and the Box-Jenkins chart for time dependent data.
There are two kinds of correlation: contemporary (children's weights are correlated to their heights) and temporal (as children grow older, weight and height will be autocorrelated and cross autocorrelated). Hotelling's T-square statistic is only appropriate for contemporary correlated data, because it assumes independence between the time adjacent multivariate points. Quality control charting for data that are multivariate and time dependent is a topic that requires development. Barrett is correct in emphasizing the need for further research in all aspects of handling quality control data.
J. STUART HUNTER
Maintenance FMEA Used To Evaluate Equipment
I am writing in response to an article in the December 1999 issue of Quality Progress (Teodor Cotnareanu, "Old Tools--New Uses: Equipment FMEA," p. 48). A team at my organization took the initiative to implement a program similar to what was described in the article.
The team created a cross functional team consisting of quality, manufacturing, engineering, production and maintenance representatives. They met and covered every aspect of the machine, in much the way an organization conducts a process or design failure maintenance and effects analysis (FMEA). Then they developed the following list of core process functions that were evaluated on the pieces of equipment:
* Work space.
The team examined these areas on the equipment and followed the FMEA format to develop a new list of severity, occurrence and detection ratings (see Table 1 on page 16).
I enjoy reading your magazine every month, and I especially enjoyed this article.
Author's Response To Letter in Dec. Issue
We are happy to respond to Donald Ermer's question ("What Is a Negative Lower Control Limit?," December 1999, p. 14) regarding our article in the August 1999 issue of Quality Progress (Carl B. Montano and Glenn H. Utter, "Total Quality Manage- ment in Higher Education," p. 52). The fact someone even had a question at all pleases us--and we thank you--because it indicates that there are colleagues out there who are in-terested in and serious about improv- ing the quality in higher education.
In our article, Figures 3, 4 and 5 are all i-MR charts, with the MR part not shown. This type of control chart plots individual observations (i) and their moving ranges (MR) over time. By contrast, x-R control charts plot the means (or x) of the sample observations (of sample size, say n = 4) and their corresponding ranges (R) over time.
Regardless of the kind or type of control chart, however, both the upper and lower control limits are calculated equal to mean ±3 standard deviation. The upper control limit equals mean or average + 3 standard deviation, and the lower control limit (LCL) equals mean 3 standard deviation.
Software used to generate control charts usually places the i or x plot on the upper part of the page and the corresponding MR or R plot on the lower part of the page. But to save space in this article, the plot of the MR wasn't shown.
In all of the above figures, the negative values of the LCL are meaningless and should be ignored. We all know that the lowest observation in this study should be zero, meaning no student applied for admission to Lamar University that particular week. The negative LCL appeared on the plot because, indeed, that is the value of the mean 3 standard deviation. But we, the authors, so understand that negative numbers are meaningless. We could very well have cut the control charts to show only zero and positive numbers. But you can also imagine that doing so would have made the control charts look unbalanced and ugly.