Lean and Six Sigma Do the Trick
Thank you for beginning to give lean Six Sigma some equal time in the April 2003 issue (George Alukal, "Create a Lean, Mean Machine," p. 29; Bonnie Smith, "Lean and Six Sigma--A One-Two Punch," p. 37). This combined approach has worked well at a number of my company's locations, and we have only begun to scratch the surface.
We currently have four major businesses, three investment businesses, 59 manufacturing locations and about 5,200 employees worldwide. We have 14 locations registered to either ISO 9000 or QS-9000. The remaining 45 are not registered.
Our various operations have tried many different approaches to business improvement over the last 20 years. While all these approaches have been useful and effective to varying degrees, no one approach produced dramatic, sustained improvements until we combined lean operating principles with Six Sigma project improvement discipline.
Once we started combining work cells, 5S, pull systems, kaizen blitzes and other lean methods with the Six Sigma Black Belt, Green Belt and Champion support structure and the define, measure, analyze, improve and control project management discipline, we began getting the traction we needed.
We are now beginning to aggressively apply lean Six Sigma as a combined approach in all aspects of our business. These include customer interfaces, up-front office systems, manufacturing and supplier interfaces. We have had excellent results with this combined approach at about 12 locations and are beginning to spread it to our other operations.
Please continue to keep this topic high on your list of items to report about.
Lean Has Contributed To Quality Successes
I completely agree with George Alukal in "Create a Lean, Mean Machine." I am a lean practitioner at Kirkhill-TA in California and have been in the quality field for 17 years in varying capacities. In all that time, I have never been able to positively affect quality as I have in the last eight months in my new position as a lean facilitator. Here are a few of our recent successes:
- First-time yield went from 47% to 97% within two months of beginning kaizen.
- Raw material scrap went down by more than 50% after another kaizen event.
- Defective parts went from more than 8% to less than 2% the first month after kaizen.
- Work in process (WIP) went from 21 days of inventory down to only that day's production.
Each of these is a separate success in a different department, and there have been no failures. Lean is an amazing tool. The reduction of the amount of WIP contributed to these great quality related successes. It allowed the people doing the job to see the quality of their work. It also had a psychological impact because people are no longer pressured into believing they must produce everything that's in front of them. Rushing the job is not an option in a pull system.
A caveat: If the chairman of the board, president and those in upper management don't get out and make their support of lean visible to the workforce, it will fail.
HOWARD F. HULIN
"Lean" Is the Quality Buzzword of the Day
One of today's quality buzzwords is "lean," but what exactly is it? It brings back memories of quality circles and just-in-time (JIT)--buzzwords of the '80s. Those earlier philosophies were America's first attempts to face its then looming global competition.
I think we embraced JIT more than quality circles because its financial impact was most evident. Unfortunately, we still haven't matched our international competitors in the quality arena. Hopefully we've learned from our mistakes and will take a long, hard look at lean. It is an evolutionary philosophy, and evolution is usually an advancement.
St. Davids, PA
Taylor Invisioned Lean Concept Long Ago
I wonder what Frederick W. Taylor would think of lean manufacturing? I suspect one thought would occur to him: plagiarism!
Anyone interested in lean manufacturing and its historical origins should read Taylor's book Scientific Management (W.W. Norton, 1911).
Lean Article Provides A Good Overview
For someone who is just beginning to explore lean, the article "Create a Lean, Mean Machine" provides a good overview of the issue. But for beginners and those currently on the lean journey, it is just the same stuff written in a different format.
Internally, my company added a ninth waste: the waste of unactionable information systems. I believe this is an excellent addition to the list and is a problem that plagues many practitioners and organizations. People sit in front of their computer screens with a questionable thirst for data from which no action may stem. They wonder, do we have a measurement system or a management system?
I also would have added more information on problem solving to the article. I agree lean principals work on the identification and elimination of waste, but problem solving works to reduce and eliminate variability in any value stream. Absorbing variability drives questionable and inconsistent quality, which could be reflected in adjustments and inventory.
Leaders Should Take Cues From Sisters
Thanks for your article on SSM Health Care (Susan E. Daniels, "Rx for Excellence," April 2003, p. 42) and for Debbie Phillips-Donaldson's column comparing sister Mary Jean Ryan to Dennis Kozlowski and Kenneth "Kenny-Boy" Lay ("Corporate Ethics Rule," April 2003, p. 6).
Given the nature of our national scandals--from the White House to various boardrooms--perhaps we should consider having our leaders take vows of poverty, chastity and obedience as the good sisters do.
Surely, greed, arrogance and lust are the antithesis of quality leadership.
Juran Center for Leadership in Quality
Personal Attack In Letter Unjustified
I was surprised and disappointed to read a personal attack against me in "QP Mailbag" in the April 2003 issue ("Auto Manufacturers Not Focused on Real Change," p. 8). Allen Huffman alleges I have "made millions of dollars off the standard [I] forced down the throats of the supply base."
I count two errors in this statement. First, I have never made millions of dollars or anything remotely close to that from QS-9000, my investments or my salary.
Second, I forced nothing down anyone's throat. QS-9000 was developed at the request of a number of automotive supplier CEOs in response to their concern about multiple quality assessments by their customers. Chrysler, Ford and General Motors responded to these requests by chartering the supplier quality requirements task force and directing it to develop harmonized reference manuals and requirements.
My colleagues on the task force and I, with the strong support of our respective corporations, concluded that a single harmonized group of requirements and a single assessment to those requirements would be most responsive to the supplier CEOs' request. We also believed, perhaps naively, that using the third-party registration process would give suppliers a sense of ownership of their quality management system.
Huffman's ad hominem attack no doubt arises from some real or imagined slight he has suffered. It is surprising, however, that QP would honor such an emotional diatribe by publishing it.
While I am no longer a member of ASQ, I was during much of my 30-year career with Ford Motor Co. and my nine-year career with KPMG Quality Registrar. I would never have thought of using ASQ or QP to voice my personal grievances or make allegations about a specific company or person.
RADLEY M. SMITH
Ann Arbor, MI
Who Determines Key Characteristics?
Could you please clarify something in the March 2003 article "Survey for Action, Not Satisfaction" by John Cravenho and Bill Sandvig (p. 63)? The authors say the survey focuses the customers' comments on 20 to 25 key characteristics. Are these characteristics decided by the surveying party or by the respondents?
According to the article, it should take the respondent less than five minutes to answer three questions about 25 characteristics. I assume the customer would not be required to come up with the 25 characteristics, but that the characteristics are already provided and the customer would just need to pick the top five for each of the three questions.
Our discussion of this article at a department meeting led to two different interpretations of how the characteristics were determined. We are thinking about using this survey format for ourselves, but need some clarification before moving forward.
TINA BOARDMAN, CQA
Author's Response: The 20 to 25 questions are developed through a collaborative effort of the staff and management of the company conducting the survey, with guidance from their consultants.
Figure in 'Survey' Article Needs Clarification
I'm confused by something in "Survey for Action, Not Satisfaction" and think one of the figures could possibly be wrong.
In Figure 4 (p. 66), the target line ascends from right to left across the chart, and the authors state the bars move from right to left in descending order of importance.
These two taken together seem to indicate the most important issues for customers (far right on the chart) have the lowest expectation in terms of target performance. This is contrary to the way I would view this feedback from customers.
Author's Response: Yes, you are correct. The most important bars in Figure 4 are on the left-hand side of the chart, and the least important bars are on the right-hand side of the chart.
Miscalculated Outlier In March Article
I believe Joseph Conklin miscalculated the outlier criteria for the box and whisker plot in "Smart Project Selection" (March 2003, p. 81). In step 3 on p. 82, 9.75 should not be subtracted from the median; it should be subtracted from Q1 (88.25 - 9.75 = 78.5). The project score of 74 still appears to be out. In step 4, 9.75 should not be added to the median; it should be added to Q3 (94.75 + 9.75 = 104.5). Given that 100 is the highest score possible, this simple calculation should raise some serious questions about this treatment of the data.
The author makes comments about the normality assumptions in the Dixon test but doesn't stress the problems with the box and whisker plot when dealing with bounded data such as 0 and 100, which do affect this data set.
Recognizing the skewness of these data, analysis of Ln [100 - project score] reveals a bell shaped distribution and indicates the project score of 74 is not unusual.
Author's Response: Step 3 for low outliers should read, "Subtract 9.75 from Q1, 88.25 (88.25 - 9.75 = 78.50)." Step 4 for high outliers should read, "Add 9.75 to Q3 (94.75 + 9.75 = 104.5)."
I wanted the data set to look like something the majority of readers might have seen before, and I thought the project scores stood a good chance of filling that bill. I also wanted to use outlier detection techniques that have a wide application. I believe the ones selected for the article fill that bill, too, though that doesn't mean they are the only ones available.
By their nature, generally applicable techniques are not the best choice for all individual data sets. Whenever theory or experience suggests a particular transformation is appropriate for a data set, it usually deserves precedence over the general technique in data analysis.
I will add the transformation Kittlitz suggested to my arsenal for naturally bounded data sets. My thanks to him for keeping me on my toes and for giving me an idea for another article.
U.S. Bureau of the Census
Predictive Maintenance Article Misses the Point
I think it would have been helpful if Carmen Carnero Moya, author of "How To Set Up a Predictive Maintenance Program" (March 2003, p. 56), established a clear definition of what a predictive maintenance program is and outlined the goals of such a program early in the article.
I don't think the article addressed the title, and it provided limited detail on how to successfully set up a predictive maintenance program. The article focuses more on how to promote industry standards for predictive maintenance.
Preventive maintenance was briefly mentioned, but it means many things to many people, from lubing tools to implementing an advanced replacement, analysis and testing program.
It is unclear if the author is at all familiar with the statistical approaches to maintainability engineering for establishing preventive maintenance programs. Such statistical techniques are used in aerospace and in the military.
JAMES D. HERARD
Endicott Interconnect Products
We welcome your letters. Send them to EDITOR, ASQ/QUALITY PROGRESS, 600 N. PLANKINTON AVE., MILWAUKEE, WI 53203-3005; or e-mail them to firstname.lastname@example.org. Please include address, daytime phone number and e-mail address. All letters will be published on the QP Discussion Board, or you can post your comment on the QP Discussion Board directly at www.asqnet.org. We reserve the right to edit letters for space and clarity. Opinions expressed are those of the writers and not necessarily of ASQ or Quality Progress.