Tire Problem Lies in Quality, Reliability Documents
Thank you for your rundown on the Bridgestone/Firestone fiasco in the December 2000 issue of Quality Progress (Susan E. Daniels, "Tire Failures, SUV Rollovers Put Quality on Trial," p.30). It is evident that all of the authors are well-established in the quality profession.
I want to add that the real problem lies in how the quality and reliability documents, such as customer rejections, loss of life statistics and pertinent correspondence, are received and filtered by an organization. In some cases, the manager of quality, vice president of quality and reliability, or the director of quality assurance is unable to take immediate action. In the Bridgestone/Firestone fiasco this situation resulted in the loss of lives.
Each person in one of these positions has the authority to shut down the line. When the top quality and reliability employee is informed of the gravity of the situation, he or she should suggest to top management that the line must be shut down until a corrective action is in place. At this point, the president or chief executive officer must agree to the action. If quality is not supported, the manufacturing or finance manager will rescind the order.
The management of two successful
programs, Saturn/Apollo (Boeing-NASA) and the Polaris Program (Lockheed-Navy),
supported the quality position, and quality was notified of rejections
and customer complaints during construction, test and field
performance. Top management supported the quality decision to shut the line down until a corrective action plan was implemented. The suppliers were required to do likewise.
In the case of the tire situation, the data were suppressed, and the line was not interrupted or shut down until the loss of life was common knowledge. It is alarming that the U.S. Congress was forced to pass a law to ensure the timely disclosure of life threatening product deficiencies when it is well within the power of existing organizational structure to save lives.
Saturn/Apollo SIC Program
Sun City West, AZ
Tire Customers May Share Responsibility
The December 2000 issue of Quality Progress contained many interesting articles on the tire failure and SUV problem (Susan E. Daniels, "Tire Failures, SUV Rollovers Put Quality on Trial," p. 30). Surprisingly, only one article even remotely referred to driver's habits as a possible contribution to the problem.
The article by D.H. Stamatis, "Design and Manufacturing FMEAs Are Called For," stated that "the tire should have been designed appropriately and tested under real-world usage conditions." I hope that "real-world usage conditions" does not mean excessive speed, overloading the vehicle, reckless driving or failure to perform proper maintenance.
I am sure that the tire manufacturers are equally to blame for this situation, however, the customer must also share the responsibility to properly use the product in a safe manner. It is interesting that neither ISO 9000 nor QS-9000 addresses customer responsibility as a quality system requirement. Quality should be used as a tool to produce a product that is functional and reasonably safe for its intended use. Manufacturers should not be forced to compensate for every conceivable abuse of the product. To do so would only drive up costs. In every situation product safety has a limit. Any product that is abused or operated in a negligent manner will become a danger to the user.
El Dorado Hills, CA
What's Necessary For Customer Satisfaction
I have not been inside the Ford or Bridgestone/Firestone plants, but I would like to add to the comments in the December 2000 issue (Susan E. Daniels, "Tire Failures, SUV Rollovers Put Quality on Trial," p. 30).
"Mr. Pareto Head" offered the quality plan "PDPQ or Plan, Delay, Pray and Quit" (Mike Crossen, December 2000, p. 12). It echoes what I heard during a recent company visit: "We tried total quality control and total quality management. Neither of these worked. Now we are working on Six Sigma, and that is probably not going to work either." Each of these quality programs was sold by a quality manager to top management or was picked by the top manager who said to put it in place. What is the missing ingredient that prevented these good programs from succeeding?
Aristotle said, "It is possible to fail in many ways ... while to succeed is possible in only one way ... ." Providing safe and effective products and services requires interactions among many people and attention to detail. ISO 9000, QS-9000 and Six Sigma programs alone are not able to provide the one way for success. To be successful requires recognition that it is the leaders' attitudes that ultimately determine if devices will be safe, reliable and effective.1
Cues come from the leaders, starting with top management and extending to the supervisor on the line. When the leaders have the right attitudes, it does not matter which quality program is used. Companies produced quality products and services well before the advent of ISO 9000, QS-9000 or Six Sigma. People produce products and services. Standards, programs and certifications do not. Standards cannot cause people to reach out and search for ways to improve and maintain better control over the processes involved in delivering a product or service. It is the attitudes of the leaders that lead to excellence.
Quality and Regulatory Affairs
Poor Processes Turn Out Poor Products
The article "Tire Failures, SUV Rollovers Put Quality on Trial" (Susan E. Daniels, December 2000, p. 30) was timely and fair in its overview. It was also quite revealing of the hidden agendas in the quality profession. In particular, I noticed a strong defense of the emperor, ISO 9000, who has no clothes. A good analogy is a police department that boasts of the community jaywalking rate when the felony rate is out of control. Quality professionals have a tough sell convincing customers that an ISO 9000 quality management system is effective even though they must fear for the safety of their families.
Convincing engineers is an even tougher job. Amy Zuckerman's contribution, "Some Believe QS-9000 Plays No Role in Situation," states that QS-9000 auditors examine only the process, not the product. But the whole justification for process testing is the belief that if the process is good, the product will be good. As a systems engineer, I believe that a process that turns out a bad product is a bad process.
QS-9000 and ISO 9000 have taken a hit as a result of the Ford-Firestone fiasco. A few of the contributors hold that ISO 9000 is a first step toward quality. If so, QS-9000 is a second step and is still not good enough. Perhaps a series of graded standards such as the capability maturity model (CMM) is needed. As I understand it, the CMM has five levels of quality.
With respect to this sorry tale of industrial failure and death, it does not behoove quality professionals to say as Pontius Pilate did, "Lavabo inter innocents manus meas," or "I will wash my hands among the innocents." We look foolish endorsing a process that can kill because the documentation is nice and neat. We risk relegating ourselves to irrelevance.
WILLIAM A. STIMSON
End Result Depends On Corrective Actions
The role of ISO 9000 in "Tire Failures, SUV Rollovers Put Quality on Trial" (Susan E. Daniels, December 2000, p. 30) was handled by several experts. Each expert had a different opinion, and an outsider cannot know if ISO 9000 and internal standards were followed or not. Therefore, I want to pinpoint two aspects of the ISO 9000 standards.
Harold P. Greenberg summarized the situation in "Ethical Considerations in the Firestone-Ford Situation" by stating that the root causes are found in the design or manufacturing process. My experience is that 80 to 90% of all quality problems originate in the R&D phase.
I recommend studying how well the tire manufacturer applied the reviews in the design phase. The organization needed to ask, "Do these specifications meet actual driving conditions in the harshest environment?" If the ISO 9000 requirements were followed, then it is a matter of know-how, and quality procedures cannot substitute for know-how.
Many writers wondered why Ford and Firestone failed in performing corrective actions. ISO 9000 defines the procedure for taking corrective action: Once a nonconformity has been detected, a root cause must be found, and it must be eliminated. If this is the whole truth, one can say that organizations have not followed this ISO 9000 requirement.
The problem is that the corrective action procedure in ISO 9000 is impossible for any organization to follow. If nonconformities can be eliminated via ISO 9000, a good number of registered companies would have zero defects.
Unfortunately, system failures cannot be eliminated that way. ISO 9000 considers that every nonconformity has a special cause. Because this approach won't work in reality, organizations and registrars use the loophole in ISO 9000 that states "to the degree appropriate to the magnitude of problems and commensurate with the risks encountered." This means that any practice can be good enough for organizations and registrars. ISO 9000 requirements are too loose, and the end result depends on the company instructions for corrective actions.
What Was the Appropriate Approach to the Election?
The recent presidential election resulted in one of the closest margins in history. Those of us in the quality profession know all measurements are subject to some level of uncertainty. It's ironic that the pollsters properly addressed the margin of uncertainty when predicting results before Nov. 7, but voting results were treated as truth following the election. Was that an appropriate approach?
The data are that 100 million votes were cast, with a Gore victory of 100,000 votes. The victory margin was less than 1%. In the state of Florida, Bush won by 537 votes out of 6 million cast, a margin of about 0.01%. Was that decisive or was that within the margin of error for the measuring device? Experience shows that most calibrated measuring devices are accurate to about 0.1%. A device with a measuring accuracy of 0.01% is considered highly accurate.
Literature states that the measuring device should be anywhere from four to 10 times as accurate as the measurement required. The margin of uncertainty should be small when compared to the specification limit. So the national poll margin of error should be about 0.01%, and the Florida margin should be about 0.001% to give confidence that the tally accurately reflects the true result.
In order to calculate the measurement uncertainty, we need multiple counts of the same data to determine a standard deviation of the repeat counts. It was reported that Bush carried Florida by margins of 1,850; 950; 250 and 400. This provided an average Bush victory of 862.5 votes, with a standard deviation of 723.8 votes. So the upper and lower six sigma control limits showed a Bush victory by as many as 3,896 votes and a Gore victory by 1,309 votes. The Florida result is, therefore, within the margin of error, and there is no winner. Statistics prove the surprising and unsettling news that nobody won Florida.
ITS Intertek Services
Gulf Breeze, FL
We welcome your letters. Send them to EDITOR, ASQ/QUALITY PROGRESS, 611 E. WISCONSIN AVE., PO BOX 3005, MILWAUKEE, WI 53201-3005; or e-mail them to firstname.lastname@example.org. Please include address, daytime phone number and e-mail address. Whenever possible, the e-mail addresses will be included with published letters. Due to space restrictions, Quality Progress will publish a selection of letters in the magazine. All letters will be published on QP Forum, or you can post your comment on QP Forum directly at www.asqnet.org. We reserve the right to edit letters for space and clarity.