The following article was published on the Ford Communication Network in 2004, when I was the Henry Ford Technical Fellow for Quality Engineering
“The path to wisdom begins by calling things by their right names.” Taiichi Ohno, founder of the Toyota Production System
Anyone who cares about Ford’s future should make sure they understand what Dr. Timothy Davis, the Ford Technical Fellow for Quality Engineering, is saying to everyone who will listen. You start by learning the name he uses for the outcome of doing things in a less than perfect way. That name is “failure mode.” He uses it a lot.
If you design an underhood layout with a power steering pump next to a hot exhaust manifold, you’re in for a failure mode, because the heat will cause the pump’s seals to dry out and leak. If you design a defrost unit that can melt a glacier but it is hard to figure out how to use, you’re in for another failure mode, because even though it will defrost the windshield, it will frustrate the customer.
Now learn Davis’s name for whatever fixes a failure mode. It’s called a “counter-measure.” He uses that word a lot, too. Ford’s success, Davis says, depends on finding failure modes as early as possible in design and product development, then applying countermeasures to fix them. Early in the process, it’s hard to find failure modes, but it’s easy to apply counter-measures. Later on, the failure mode becomes obvious -- like when the power steering pump is leaking all over the road -- but the counter-measure is difficult or even impossible to implement, and expensive. Without careful thinking and planning, failure modes are inevitable. The trick is to find them as early as you can.
When Davis was a teenager in Warwick, England, he often ran his father’s bakery. One time he mixed 300 pounds of dough to make 200 loaves of bread. But he forgot to add salt. If he had discovered this failure mode while he was still mixing the dough, he could have applied a counter-measure easily and cheaply, just by adding some salt to the dough and re-mixing. But he discovered the failure mode only after the bread was baked. The only possible counter-measure then was to throw all the bread away and start over -- a costly and wasteful fix. “This is exactly what I teach,” Davis says. “You may think that, to engineers, this is obvious and trivial. But, truthfully, it’s not.”
It’s not obvious because manufacturing automobiles is so much more complex than baking bread. And the U.S. auto industry grew up with the age-old engineering tradition known as cut-and-try, which relied less on careful planning than on intuition and trial-and-error. Then came Toyota, Honda and Nissan. Because natural resources were so scarce in Japan, those companies maximized planning to minimize waste, a system that Henry Ford himself did much to inspire. American engineers, following the lead of such quality gurus as W. Edwards Deming and Genichi Taguchi, began to understand that cut-and-try is inherently wasteful, expensive and slow. And intuition, as Davis points out, is often simply wrong.
“Throughout history, many of the intuitively obvious solutions to scientific problems have been wrong,” he says, “and it’s taken people with passionate advocacy to prove it. Take the example of Copernicus, who postulated that the earth went round the sun. Up until that point, the intuitive interpretation was that the Earth was at the center and everything else rotates around it, because people couldn’t tell they were moving. Their intuition was wrong. “Automotive engineering is so complicated that I simply don’t subscribe to the theory that gut intuition is going to be right all of the time. If that was true, we would be completely unique in the whole gamut of human understanding of the way the world works.”
In manufacturing, the alternative to intuition is the field known as quality engineering -- the application of scientific and statistical analysis to design and production. In nearly 20 years at Ford, Davis has become one of the world’s leading figures in the field, coordinating a vanguard of quality-and-reliability engineers who work throughout the company. Of course, statisticians in a factory are often greeted the way a carpenter greets an architect -- with a roll of the eyes and the expectation of a lofty but irrelevant lecture. But Davis has made his mark by applying his ideas collaboratively on many Ford vehicle programs. As a quality manager with Ford of Europe, he helped bring the Focus and Mondeo to launch, among others, and as quality director for North American Truck from 1999 to 2002, he was instrumental in the quality planning that ultimately led to an excellent launch for the new F-150.
Earlier this year, when the older F-150 and the Ranger earned top marks for three-year reliability from J.D. Power and Associates, Davis and his team recalled with satisfaction the 150-some counter-measures they applied to failure modes on those two trucks three years earlier. (Based on that work, he and his colleagues were recently awarded a U.S. patent for their process of finding and fixing failure modes in the field, and avoiding them on the next models.) “I think we have here some tangible evidence that theory matches experiment!”, he says. “It’s always nice when your plan actually does deliver.”
Although he holds a Ph.D. in statistics from the University of Birmingham (U.K.), Davis isn’t much interested in arcane formulas. What he cares about is how Ford designers and engineers think about what they do. This is what the talk about failure modes and counter-measures is all about. These, he says, are the fundamental quantities that engineers should worry about. Any other way of thinking about manufacturing is a distraction. “As soon as we’ve said, ‘We’re going to do this car or that truck,’” he says, “from that point forward, we’re actually managing a failure mode factory, as bizarre as it sounds, because what we’ve got to do is find all the failure modes -- all the things that are going to make my design go wrong. “Now, do you find them early or do you find them late? Our current state is that we find too many failure modes late, exactly at the point where there isn’t much latitude to effectively deal with them. What we’re trying to do is find as many failure modes as we can early, where the latitude to take the countermeasures is widest.
“I’m very keen to get engineers talking in terms of fundamentals. Failure modes are very fundamental things. Countermeasures are very fundamental things. If you talk in those terms, it’s very difficult to say something that will distract people from understanding the true state. You don’t talk about ‘issues,’ you don’t talk about ‘current assumptions,’ you don’t talk about ‘unresolved problems.’ You say: ‘What is the failure mode? Do we have a countermeasure for it? And can we implement it before we release the drawing?’ It’s actually that simple.”
Praise for Davis’s approach comes from no less an engineer than the quality-and-reliability pioneer Don Clausing, the longtime technical chief at Xerox Corp. who became a leading engineering professor at MIT and a close collaborator with Genichi Taguchi. Now retired, Clausing was a founder of the theory of failure-mode avoidance, and he has watched Davis apply and refine the idea in the automotive world for ten years. “The problem with most statisticians is that they have good answers to whatever they think the question is, but because they know very little about engineering, they’re not very good at formulating questions,” Clausing said. Davis, he said, excels at both. “I actually don’t know anybody who’s as strong at this interface of engineering and statistics as Tim is.”
(Indeed, Davis’s work was recognized earlier this year when he was named a Fellow of the Institute of Mechanical Engineers in London; he is now a Chartered Engineer as well as a Chartered Statistician.)
“The challenge,” Clausing adds, “is to get Tim’s thinking implemented throughout the organization. If they can, it would enable Ford to catch up with Toyota. But that’s a huge challenge in an organization as big as Ford.”
Davis is well aware of the difficulty. “This is not for the faint-hearted,” he says. “To turn around an engineering organization and get it to think differently – you don’t put that on your list of annual objectives, because if you do, you’re going to fail. This is something that you have to passionately believe in, and you have to be consistent over a number of years as to what the right thing to do is. And eventually you start to see tangible results.” He works at it day by day. He gives seminars to engineers. He chairs design reviews—though he prefers the term “planning and assistance" meetings, because he thinks the word “review” implies that he’s inspecting the engineers’ work and giving them grades, when in fact he spends his time asking questions to uncover problems so that he can help. His tool of choice is conversation, not edict.
“You have to be a little bit of an evangelist,” he says. “Of course, management can always say to people, ‘Thou shalt do this,’ and you get some sort of immediate impression that people really are doing it. But I’ve always subscribed to the view that people really only do what’s in their hearts. In the long run, you get better results if you engage people’s intellect and passion, and get them to contribute to the solution.
“Now, I don’t claim that all my ideas are correct. In fact, many of my friends will tell you that some of my ideas are wacky. But you have to have the odd wacky idea to uncover the good ideas. And I like to think that I have had a couple of good ideas.”
Failure mode avoidance is useful for authors as well as engineers…..
From: James E. Tobin
Sent: 12 November 2004 09:26
To: Davis, Tim (T.P.)
Subject: Revised version of the article
I hope the article will do some good. Your ideas have even helped me to think about my own work in a new way.