Some Things Just Can't Break

I've been pondering the recent Toyota woes, but once again had to be motivated by someone else to actually type something up. In this case it was a column by Matt DeLorenzo in the latest issue of Road and Track, in which he cites both Toyota's growth and a slight flaw in the Deming quality control philosophy Toyota follows.

He believed that it was better to establish an acceptable level of quality variation rather than waste resources on trying to reach zero-defect levels. While this concept works well when manufacturers are building hundreds of thousands of vehicles, it works less well when those numbers reach into the millions. Even if a manufacturer hits 99.9-percent effectiveness, that 0.1 percent translates into 1 car per thousand. Multiply that by millions and pretty soon it adds up to a large number.

This isn't really a flaw, just an acceptance of reality. No manufacturing process is perfect. Matt is quite right to note that quality gets harder to maintain at mass-production levels; ask anyone who has to cook for large numbers of people. He also makes a valid point that given today's increased demands and pressures perhaps trying for perfection is worth considering. But he leaves out something I believe is important, which is the need for different standards for different parts.

The reasons Toyota's troubles have been such a big deal are which component is failing and the nature of that failure. Accepting a small failure rate for power window motors or air conditioning compressors is fine. But throttle control is mission-critical. Not only should the goal be zero failures, but more importantly, given the ultimate inability to achieve that goal, allowances for failure need to be actively designed into the system to deal with as many potential consequences as possible In this case, the throttle needs to be designed to not stick open no matter what.

I also can't help but notice that this is part of an electronic throttle system, which is something relatively new to cars, especially at the large-volume lower-priced portion of the market. Companies test parts obsessively, but it's tough to duplicate the ravages of time and varying conditions, and difficult to spot impending trouble with inscrutable electronics. We may start seeing new and unexpected failures as the more recent automotive innovations begin to age.

Matt's mention of schadenfreude at other car companies over Toyota's troubles also interested me. I would hope there's also a a feeling of "there but for the grace of God go I." After all, most parts suppliers, like the company that made those throttle controls, make parts for many companies. There's always the chance that one of them could get that next batch of statistical inevitability.