The Pig Pile Gets Bigger

Critics of the Toyota Production System have been quick to pile on the revolutionary management approach as part of the reason for the recall. This is faulty thinking. Here is why:

The pedal design problem, according to Toyota, is due to a design issue - possible corrosion related. Corrosion, like other engineering problems, is one of a predictive nature.

Back on February 3, I talked with my co-author of the Job Instruction Implementation Manual, Mark Warren, of Tesla2, Inc., about this problem. Mark has been called in to solve failure analysis problems for companies so he had some interesting insight on this matter. After reading my post on "The Pig Pile", Mark had this to say:

"Have any of the critics considered the number of miles on the vehicles in question or the age of the components or the locations of the failures? Given the initial stumble of attributing the failures to floor mats is indicative that they did not use their famous admonition of considering multiple solutions before choosing the final one."

So, at first glance, Mark seems to be falling onto the pig pile, blaming Toyota for not following their principles. Did they simply abandon the questioning method, focusing on the simple, low cost solution of fixing the floor mat? Or were they facing a more complex solution not easily solved by a quality circle or kaizen solution? Mark continues without going down path of the "see, I told you TPS was all wrong! crowd..."

"However, most field failures are rarely a ‘cut and dried’ affair, especially where tragedy has struck. No one likes to hear that human error was the cause or major contributor. While the news reports claim that there are thousands of complaints of sticking throttles, the first step is that we must realize that this is a symptom description. One that was given by a large number of different people. When you take a large number of failures with the same symptom description that are presented to a warranty department, they usually take the component and try to replicate the described failure. Typical results will have a distribution of some of the complaints matching the initial description. You will also have other failure modes that can possibly be described by the initial description, but fit a different failure mode better. Then you have the samples that function perfectly – to the original manufacturing specifications. Granted, some of the parts might exhibit intermittent behavior, but many just get pulled because the dealer is just trying to satisfy a customer that is concerned (the placebo effect). (General rule of thumb was only about half of the components submitted had ANY identifiable defects. People just replace parts until the symptoms go away.)"

This description of the data collection and analysis phase of problem solving feels about right, doesn't it? We can't replicate the problem, or other problems are discovered, or everything works just as expected. Question for the Pig Pilers: is this the "fault" of the Toyota Production System?

"Oops, I forgot to shoot down the people that think a Toyota employee has failed because they did not pull the andon cord on the assembly line. If this is a corrosion problem that caused the sticking of the throttle mechanism, this failure may take months or years to happen. This is no reflection on the Toyota line workers. In addition, most parts undergo a salt spray to test for corrosion. I find it unlikely that the initial design would be the problem. Most cases like this happen where minor changes are made to a component that has been used for years failure free and the contract manufacturer does not complete the lengthy reliability testing on the changes."

Again, the fault of Toyota Production System? Probably not...would we blame this problem on the Ford Production System if this was found with in the Focus, Fusion and Taurus? It seems to be that this problem has presented itself as a convenient way of allowing people to exhibit some good ole' fashion protectionism...but I digress...back to Mark:

"Now, back to the people point fingers…even when you might have a unit with corrosion, this does not mean that you can have the part stick every time you try it. With such a statistically low failure rate there are probably other factors as well that exist to create the situation. Multiple dependent failures are difficult to detect and to duplicate. Because of this the investigators often make misleading assumptions, mostly claiming that they have found THE cause. That is until more failures occur and they must go back and sift thru the data yet again.

"Probably what is going to found is the typical car part has had elevated exposure to moisture and temperature (accelerating the corrosion). That the part dimensions between the sticking components are not near the production means (less clearance than normal). That the return spring might also be corroded, leading to lower than normal C rate to return throttle to idle (or on the lower value if there is no corrosion). That the normal driving habit did not include pushing the throttle to the max position – if they regularly floored the throttle (aggressive drivers), corrosion would be worn away as it accumulated. However, a careful driver is more at risk than an aggressive one (unless they are not the normal driver) – with the random time that they move the throttle into a position that was not typical. Corrosion would have had time to build up to a point that the normal return controls would not be able to function.

I ask you to consider this point about the difficulty failure analysis presents given the wide demographic of consumers, driving conditions, environment, etc. I ask you to consider this because there is also a lot of criticism about the delay in Toyota's response to this problem affecting peoples lives. If Toyota recalled millions of vehicles three years ago and the failure revealed itself after the recall - what would the criticism be then? Hindsight is always 20/20 isn't it?

Mark suggests, "Now that you think that I have defended Toyota, I can take the other side. With all the advances in engine performance maps, I find it unusual that they do not have a safety routine to protect from accidental over-speed events. I’m sure that this sort of close inspection of what they should have done will trigger safety routines to be installed as a standard feature in future Toyotas as well as other brands."

Yes, the armchair quarterbacking of a multi-billion dollar, multi-national company does seem to be setting a certain direction, it seems that Mark's predictions are more or less correct, Toyota is taking a comprehensive approach to the safety problem. Blaming a production system scheme, however, is flat wrong.