My last post about the automobile industry was inspired by some comments I had left in response to Cuban’s post on Hacker News, a tech-themed site I often read. The comments weren’t well-received, largely (I think, given the well-received responses to it) because they implied that innovation, at least in terms of improving the product, couldn’t solve this particular dilemma. Now that I think about it, I’m not surprised. A hacker (as the term is meant there) is almost by definition an innovator, so the people there tend to value innovation above all else.
And not just innovation, but a certain sort of innovation, the sort the user sees. Y Combinator’s mantra is "make something people want", with the theory being that if you do that, all else will fall in line. While that’s largely true with software, it’s not the same in the manufacturing world because what people want, often more than the product itself, is to pay less money for it.
Don’t get me wrong, I value innovation as much as anyone else. I think humanity can solve almost all of our problems with science and reason. In fact, the automobile industry, including the American car companies, has knocked out more than their fair share of woes. The price of transportation as a percentage of overall income has been dropping for a long time.
But I came to realize that innovation means different things in different industries. In much of the tech world, especially on the web, innovation means nothing more than making a better product from the end user’s perspective. It’s clear that even if it cost Google 2x as much per search to operate as it does Yahoo, they’d still be making a lot more money. The same is true of most any web startup.
The reason for that is twofold. For one, the apparent cost to the user is $0. Technically users are paying to use both services, and more on Google than on Yahoo, but they’re doing it in a circuitous way by clicking advertisements and then purchasing things on the pages on which they land. It’s sort of a happy payment, since those Google or Yahoo ads are taking you to pages you wanted to spend money on. They’re scratching an itch. The fact that users "pay" more to use Google actually makes them happy. Nobody wants to pay more for a typical mid-sized sedan.
The other reason is that two times zero is still zero, and the cost of serving up a page is, for most web companies, effectively zero. I know a lot of people who run a lot of startups, and not once has anyone ever mentioned to me what it costs for them to serve up a page view. With the exception of the startups that primarily serve up large files (YouTube for instance) most have probably never checked. Hardware and bandwidth are so cheap when sending mainly html and a few images that until you’re serving millions of customers per day it wouldn’t even be worth the man hours it takes to find out.
Because of that, most startups don’t really worry too much about optimizing in terms of hardware costs. They just try to get to a point where scaling is just slapping in more boxes, which granted, is often a hard point to get to but only costs man hours. The marginal cost of a page view is still nothing.
In fact, marginal cost doesn’t even make sense as a metric for most web-based businesses. The vast majority of the expense comes from the engineering it takes to get to where you can slap in more servers. The time spent figuring out how to shard your database costs orders of magnitude more than the hardware you do it on. It’s an industry that’s almost entirely fixed costs.
In manufacturing, marginal cost means everything. That was the primary source of innovation for Henry T. Ford exactly 100 years ago. He didn’t invent the car, and he didn’t even make a higher quality one. He made a cheaper one by perfecting processes, such as the assembly line, that reduced the marginal cost.
Let’s say, for instance, that a Toyota Camry costs $50 million per year to engineer, but sells 1 million units. That’s a cost of only $50 per unit worth of engineering for a car that retails for something like $25,000 on average. The product innovation expense is trivial. It’s a tiny fraction of a percent, essentially a rounding error. For a car that costs $15,000 to build, the labor and raw materials are the overwhelming concern.
Innovation in the automobile industry has, for a century now, been more about reducing marginal costs than improving the product. They’ve been essentially doing the real world equivalent of rewriting your whole website in C, which every startup would be doing if serving up a page view cost them even $1 worth of processing.
So I think that’s the primary difference between the automobile and software industries. One is primarily about making the product better, the other primarily about making it cheaper. Both require innovation, just two very different kinds.
But even with both types of innovation, in an efficient marketplace, one company can’t be expected to significantly out-innovate another because it’s too easy to copy. You can’t patent the concept of an SUV. You can’t even usefully patent something like rain-sensing wipers (presumably, since multiple manufacturers offer them) because someone can trivially improve upon them or anything else you do. It’s not like a drug, where a precise chemical compound could have wild variations from one that was only 90% the same. (Pharmaceuticals are another industry in which the marginal cost is considerably less a factor than research and development.)
And really, I don’t think we’ve seen a discrepancy in terms of product innovation occur in the automobile industry. It’s hard to say because it hasn’t been a fair fight, but Americans have clearly had their share of hits. The largest trend in cars over the past decade would be the SUV, and they’ve led the charge there. Of the top 15 best selling, they produce all but numbers 2, 9, and 15. They own the pickup truck market just as well. The Mustang has been the top selling sports car for 22 years. And GM still has the world’s largest market share, though just barely. They’ve clearly been making cars people want for a long time, even while doing so at a tremendous cost disadvantage.
I view innovation in any competitive industry a lot like Olympic running. You would never expect a runner to finish 10% faster than the second or third best. For instance in this year’s 100m final, Usain Bolt set the world record of 10.2 seconds. Second place was 10.24, or 0.39% behind. The reason for that is simple, you’re taking a selected group of the best runners, all of whom put everything they have into it, all of whom have access to the same technology and practices. Is it therefore surprising that they all tend to perform similarly?
Humans are like that. There’s a lot of variance in a random sample, and Bolt might be able to finish in only half the time I could. But when you take the best tiny fraction of a percent, you find that they’re all pretty damn close to each other in ability. That’s no less true of automotive engineers than it is runners. Unless you believe in some sort of genetic superiority, you can’t expect the best American engineers to perform significantly better than the best Japanese.
What we’re doing to our automobile industry, in forcing upon them such a tremendous cost basis, is taking the second best guy in Usain Bolt’s record-setting race and giving him a 2 second head start. And when we do that, it doesn’t matter how bright our guys are, or how hard they work, or how well they execute, because the competition is going to be at least 99.5% as good. It’s human nature. Our team might finish less than two seconds behind, and in fact I think they have, but the other guys are really fast too.
Even if we’re a little better, and we might be, it’s not by much, and it shouldn’t be, and no matter what we do it never will be. To think otherwise is simply racist, and it’s provably wrong. The Japanese and Germans have their fair share of brigh
t, hard working people (hell, a lot of them are Americans) and we just can’t expect to beat them in a race while starting that far behind the line.
That’s why the Detroit bosses want to level out the playing field. They want everyone starting from the same place. Unless the competition just gets lazy, in the global economy it’s the only way they have a fighting chance. And if we’re going to bail them out and invest American tax payer money into their future, we owe it to our citizens to stop letting our opponents jump the gun.