shared what the pattern recognition skills were, but they could predict in some form—“Oh, I’ve seen that problem before, that’s Heuristic Number Whatever.” But no one ever wrote that down.
And so I started writing. To make a long story short, what I managed to extract, easily, was the four steps. When I say easily, I mean a year or two. I must have read everything written about corporate strategy. My head was spinning. But what took another year and half was that the model only worked about a third of the time when you tried to apply it to companies. This notion of market type, which was another signal on top of it, was a little more obscure. Understanding the difference between executing in an existing market, where the customers are known, versus executing in what Clay Christiansen would call a disruptive market and what I would call a new market, even though the customers were known [is important]. Even though you’ve got four steps, how you do them and how much money you spend dramatically changes the nature of the company. And in fact, it explains the Internet bubble. People were spending money like they were in an existing market, when in fact the Internet, in almost all its forms, in 1999, was a new market. We had just proven that infinite cash could not accelerate adoption. We actually ran that experiment, without actually having the words. So customer development, I would say, had a three or four year gestation, as I was literally trying to create a pattern.
X: Were you trying to retroactively test the model on stories you knew about, or applying it to people you know who were in business at the time?
SB: Yes and yes. The first question was, does this explain any of those war stories that I had been through myself. Second was, does it explain any of the technology companies where I was a board advisor. Third was, I had some friends—Bill Davidow, Jon Feiber, Kathryn Gould at Foundation Capital, people who have been around since if not the beginning of VC at least the second generation of VC—who you could get to tell twenty or thirty years of war stories. I would say, “Bill, how do you know this?” and he’d say “Well, we’ve seen it a million times,” and I’d say, “Well have you guys written it down?” and they’d say, “Why would we do that?” Which was another observation. I think there should be a war crimes trial for VCs for lack of sharing of knowledge. When you push them, of course, they say “Well, that’s not our business.”
X: They’re holding on to their trade secrets.
SB: I’ve been making the argument—and now people are starting to believe it—that if we enrich the pool, then their trade secrets will move up the stack [to paraphrase Blank here: if venture partners shared their basic tricks more freely, they would go on to find more elaborate ways to offer value to entrepreneurs and investors]. That’s why I’m teaching these classes at Stanford and Berkeley, and they’re also finally catching on.
I’ve had a bigger insight, 10 years later, about what customer development actually was. I never would have described it this way until the last year, and it was only due to the work of Eric Ries—and I’ll skip that for a second, but I just want to give him immense credit. Eric’s observation as my student, and when I was on his board, was “Steve, customer development is great, but it needs to be paired with an agile development model for engineering. You can’t just assume that this works with waterfall, because you know that’s great but they’ll ignore you for a year and a half.” [Speaking very roughly, the “waterfall” software development model emphasizes a sequential process of defining requirements, designing, implementing, verifying, and maintaining. In agile software development, by contrast, requirements and solutions are defined iteratively in collaboration with users.] So now we have customer development and agile development. But really, what we’ve done is we’ve started to replicate an entrepreneurial management science stack of tools and techniques that parallel an MBA.
It’s a big idea. And the insight I’ve had in the last year or two—it might be wrong, but I think it’s right—is companies were around since the 1600s. The East India Company, whatever. But the first MBA in the United States was not until Harvard in 1908. So it wasn’t that companies were sitting around waiting for some university to codify management science. We had it anyway. But it wasn’t for a couple of hundred years that we took all that knowledge and said wait a minute, we could now teach executives a course stack—finance, strategy, accounting, management, et cetera.
I’ll contend that entrepreneurship in its modern form with venture capital has only been around for 40 years. And up until now we’ve kind of assumed that all the skills and science we needed was just like an MBA. And the observation I’m making is that customer development was the first piece of a new stack that’s education, management knowledge, that is completely different from an MBA. I’ll suggest: agile development, business model generation (Alexander Osterwalder‘s stuff), user-centric design, HR for small companies, finance and metrics. We could take that equivalent stack and say “What is it that we’re going to build not for a B-school but for an E-school.”
And so my prediction is, E-schools are emerging. And I’m doing everything I can now to be a catalyst, a Johnny Appleseed. And Eric is doing the same at incubators, venture firms, universities, et cetera, to [help people] recognize that this is not an MBA with different features, this is a unique stack, because of this second insight: Corporations execute a known business model. Startups search for a business model. Startups, when they find a business model, eventually become execution machines. But the search requires different skills. That’s the underlying distinction. (By the way, if you stay a startup forever in the search mode, you’re a failed startup. There’s no such thing as a 12-year-old startup. There’s a two-year-old startup and a 10-year-old failure.)
So a startup is, I have found, a temporary organization designed to search for a repeatable and scalable business model. Now we have a definition of a startup. So all the tools and techniques of customer development and agile development and business model design are all around that search.
