startup needs to test, and then to keep Geez! What were we thinking? track of what they learned—using my customer development process, as well TCS: So the venture capitalists in Silicon as to test the minimal viable products Valley in the 1980s and ‘90s got nervous we were building using Eric’s agile if the founders changed their initial engineering method. business model. Blank: Yeah, and [when that happened] Those three they fired the founders. Any failure was components—business treated as a failure of the individuals, not a failure of the initial hypothesis or model design, customer assumptions. That’s a huge idea. Huge. development, and agile Their belief was, “There can be nothing wrong with the plan. I funded it! It engineering—became must be an execution problem.” the lean startup. By the way, it wasn’t that the founders One key idea of this lean method was were always right. But there was no this notion that we’re learning new notion that they needed to be testing things as we talk to customers and and exploring their [company’s] test minimum viable products. And as hypotheses—about customers, about we learn, we give startups permission features wanted, about pricing, channel, to make substantive changes to their etc. … The other problem was that we business model. These changes are started stang and building a burn called a pivot. It’s hard to remember, but rate per the business plan. I remember in the old days—and this still happens being in companies where the only in corporations—once your plan is thing that [went according to] plan was blessed, any deviation is considered a our burn rate! You kind of go, “I don’t failure instead of as learning. think this is right,” even though you got three-quarters of the plan correct. So I said, “What’s wrong with this movie?”