Toyota Failure Proves Quality Isn't Scalable

There's a computer-industry lesson embedded within the bad news surrounding Toyota's unintended acceleration problem. It's out of the same playbook which in an earlier day saw GM morph from an innovator--electric starter, automatic transmission--into a lumbering behemoth. It's this: Quality is not infinitely scalable.I don't want to belabor the explanation of my point, because I think you get what I'm talking about. Sure, there's a caveat to the "quality isn't scalable" rule. Namely, a small operation can bootstrap itself into a medium or even a reasonably large company, via either organic growth or acquisition, and still maintain the quality which was the basis of its initial success.

At some point, though, whether you're talking an automaker or a technology firm, size breeds insularity, which in turn fosters risk aversion. I haven't even mentioned the corner-cutting mentality which occurs when making numbers becomes a (the only?) priority.

[Interestingly, if an operation is build upon low price, rather than quality, it appears a completely opposite dynamic is in play. That's because you can't even beginning to extract supply chain efficiencies until you get uber large. Think Wal-Mart here.]

If our canonical quality-based company is lucky, the end game consists of a prolonged period of slow decline, during which a burgeoning executive class can be supported (Bloomfield Hills) and entire communities can live decently off of the fatted industrial hog. That's the GM story in a nutshell.

Toyota might not have so fortunate a fate, and thus a swifter fall, because we now live in Internet time, where seismic shifts occur in timeframes too tiny for rational thought to stop that Twitter/Facebook/Cable TV train from running the business off the metaphorical track.

It's interesting that only IBM has been able to extricate itself from the ossification typically attendant to the aging conglomerate. That's thanks to the disruptive actions of Lou Gerstner, who made his mark and is rightly remembered for saving Big Blue. On the opposite side of the coin, Roger Smith, the CEO who helmed GM during a comparable period of travail, is remembered mainly as the butt of an amusing Michael Moore documentary.

It's true that the scalability doesn't seem to be a limiting factor for manufacturing, per se. For example, millions of laptops can be cranked out of factories, all within the same parameters of quality. The same is theoretically true for automobiles.

Indeed, it's the case now that it's not the manufacturing which is at fault in the Toyota crisis. I would argue it's not even, or even primarily, the company's engineering staff or technical capabilities.

It's a management failure brought on by an inherent inability of human beings to scale up, beyond a certain point, the social interactions which grease the wheels of a smoothly running society. (And, in a very real sense, a company is a society writ smaller.)

Or, simply put, people have an inherent, psychologically driven tendency to f' everything up eventually. Ergo, quality is not scalable.

P.S. I'm guessing you're wondering why I didn't mention open-source, where a "small is beautiful" ethos has supported the creation of a plethora of handmade, high-quality code.

I'd actually classify open source as an outlier. It's an example of an operation which is to a functioning corporation as a zygote is to a living, breathing human being. It hasn't even grown the basic support infrastructure which you need to create the possibilities of being able to begin scaling things upward.

Which is why, other than Linux APACHE server and MySQL--the latter now essentially taken commercial by Oracle/Sun--open source hasn't scaled up, and won't. (Paging desktop Linux.) And yet even its still-advantageous smallness is being squandered by the petty infighting that's pretty much the dynamic of the open source community these days. Go figure.

Follow me on Twitter: (@awolfe58)

Like this blog? Subscribe to its RSS feed: (here)

  LinkedIn

Alex Wolfe is editor-in-chief of InformationWeek.com.

Network Computing has published an in-depth report on the state of enterprise storage. Download it here (registration required).