The New Yorker recently ran an article titled “The Open Secret of Success.” It talked about how Toyota has sold more cars than General Motors has in the first three months of 2008. What makes this significant? G.M. has sold more cars anually than any other company for the past seventy years.

So how did Toyota, a company that makes a line of fairly unremarkable cars, outsell G.M.? Innovation. However, it is innovation in process, on the assembly line, rather then product. The Toyota Production System is a remarkably efficient system that has been continuously fine tuned since the late 1940s. No other car company has been able to keep up with Toyota despite them being very open about their process and even helping G.M. improve its production system. So how does Toyota stay ahead?

The answer has a lot to do with another distinctive element of Toyota’s approach: defining innovation as an incremental process, in which the goal is not to make huge, sudden leaps but, rather, to make things better on a daily basis. (The principle is often known by its Japanese name, kaizen-continuous improvement.) Instead of trying to throw long touchdown passes, as it were, Toyota moves down the field by means of short and steady gains.

This reminded me of what Paul Graham wrote about startups and the future of web applications in his book, Hackers & Painters. In the essay entitled “The Other Road Ahead” he talks about how server-based software, i.e. web applications, is the next big thing and how they are ideal for startups. He wrote this essay in 2001 which says something about his amazing insight. One of the things he touches on is exactly what Toyota has been doing for over half a century, incremental updates.

With server-based software, you can make changes almost as you would in a program you were writing for yourself. You release software as a series of incremental changes instead of an occasional big explosion. A typical desktop software company might do one or two releases a year. At Viaweb we often did three to five releases a day.

In way, Toyota operates like a giant startup in terms of refining its process. It implements millions of ideas a year from its workers, not all of which work, but the net effect is that the process is constantly getting better. Likewise, the hundreds, if not thousands, of web startups that have popped up in the last few years have taken the same approach in developing their products. Take 37signals for example. If you look at their product blog, you can see that they are always adding new features to their products.

The result is that Toyota and 37signals have become leaders in their respective industries. I think this tells us something about how companies work. The most widely used model of leaps of innovation is outdated. Imagine what Microsoft could do if they took the kaizen approach? No, monthly patches don’t count. I want a continuous stream of new features. However, in order for big companies to do this, they would have to change the way they work entirely. They would have to get rid of all the layers of bureaucratic bullshit that often ends up being a complete waste of time. Graham says they need to keep teams small and agile.

As Fred Brooks pointed out in The Mythical Man-Month, adding people to a project tends to slow it down. The number of possible connections between developers grows exponentially with the size of the group. The larger the group, the more time they’ll spend in meetings negotiating how their software will work together, and the more bugs they’ll get from unforseen interactions. Fortunately, this process also works in reverse: as groups get smaller, software development gets exponentially more efficient. I can’t remember the programmers at Viaweb ever having an actual meeting. We never had more to say at any one time than we could say as we were walking to lunch.

But here we reach a paradox. Toyota is a large company, but they seem to be able to operate like a startup (at least in improving their production system). Why can’t other big companies do the same? My theory is the same reason other car companies can’t keep up with Toyota. The reality is that they can, they just aren’t willing to take an approach puts so much responsibility into the hands of the lowly worker. Any change in a big company has to go through a ridiculous chain of approval (I’ve seen this at my school too…) which is just too slow. Kaizen means being quick and agile but this runs counter to the traditional management paradigm.

I’m not sure how one would go about changing the way a company operates in such a big way–or if it’s even possible for that matter. The startup model of operating may not be suitable in all situations. However, I have no doubt that there are some companies that could benefit from it. The existence of startups competing against large corporations proves it. Perhaps one day they will be forced to change in order to compete against startups.