Toyota, 3-Sigma, and Us
It’s not often that I agree with Charles Krauthammer, but in his recent opinion piece, “Toyota and the price of modernity“, he is on the right track when he says,
The question is: How do you distinguish the idiosyncratic failure from the systemic — for example, the single lemon that came off the auto assembly line versus an intrinsic problem inherent in that model’s engineering? How do you separate one patient’s physiology producing a drug side effect versus an intrinsic problem with a drug that makes it unacceptably dangerous?
Perfection is not an option. The question that any manufacturer must ask is if the causes of a problem observed are “assignable” or “common”. If something assignable is going on, for example, a bad batch of gas pedals, they can “fix” the problem and move on. But if the problem is systematic — if for example, the design process discounts safety issues in favor of volume production — the system must be changed as a whole or similar problems will keep reoccurring in various ways.
One method that can help determine if a problem cause is “assignable” or “common” is to chart the events over time and look to see if they fall outside of 3-sigma limits (assignable) or inside those limits (common). This will not provide a certain answer, but it can provide a good indication of the nature of the problem’s cause and how to best address it.
In the case of Toyota, the incidence of sudden acceleration is almost certainly outside 3-sigma. In other words, it is very rare. Toyota has attempted to assign the source of the problem and fix it — dangerous floor mats and bad batches of accelerator assemblies. But in his testimony before the U.S. congress, the leader of Toyota, Aiko Toyoda, takes the position that the problem is common (i.e. systemic) in his company. He says his company became focused on volume at the expense of safety and quality. This is a very interesting twist for a Japanese company that is expert in the use of SPC (Statistical Process Control.)
It is almost certain that Toyoda has internal company data that indicate systematic problems to which we are not privy. Perfection is never an option. But I suspect that the recent problems that have garnered so much attention, are most likely not systemic in nature. The reason Toyoda addresses them as systemic is because American audiences, who have no understanding of the nature of variability, have seized upon a very rare but dramatic event, and erroneously extrapolated from that rare event the conclusion that Toyota products are unsafe and that Toyota is a negligent company. Aiko Toyoda cannot hope nor dare to teach American audiences about the nature of variation. His only option in the face of American hysteria, is to take full responsibility and hopefully move forward.
Are systemic changes needed at Toyota? Toyota is one of the few companies that has the knowledge and methodology necessary to answer that question. But if systemic changes are not indicated, making such changes will likely do more harm than good. In other words, shaking up the whole company as a response to assignable causes may hurt the company, its members, and the consumer in unpredictable ways.
Krauthammer does a pretty good job of pointing out how American audiences consistently confuse common and assignable cause and, one way or another, make decisions that do more harm than good.
Americans habitually confuse common cause with assignable causes, and as a consequence they are constantly looking for, and finding, someone to blame (assign) for their troubles. In America, finding bad guys is pretty easy because we have a system the reliably and predictably produces bad guys — crooks, cons, greedy actors, corrupt politicians, Republicans, Democrats, preachers, atheists, Socialists, Capitalists, and my next door neighbor, Fred.
Once you begin to understand the nature of variation, the cause of the problems that plague us the most, become pretty obvious. Generally speaking, they are COMMON. This tells us that we need to stop looking for bad guys and start working to improve or maybe even transform, our system.
