Ben Horowitz, The Case for the Fat Startup

Ben Horowitz has been on a tear recently and I really enjoyed this article. This echoes some of the themes in Albert Wenger’s Winner Take All and Early Stage Valuations. I like a lot of the lean concept precepts, but the problem with that kool-aid is taking your decision making process away from always making choices that optimize for the chance of success (rather than capital efficiency, etc).

I think the point that’s often lost is the idea of product iteration. There’s often a huge disconnect between the big idea in the big market and what you initially release as your first product. I think one of the reasons that the lean startup ideas are so popular is that they give confidence to people that they may possibly be able to find their way to a good product. It’s probably a false confidence, but the push to innovation is a good one.

I guess I’d be really curious to hear from Ben how he used 100 engineers to R&D what eventually became the Opsware product offering, how that R&D operation was different than a big company, and whether the waste in the size of the team was absolutely necessary given the method of product iteration that they employed (beyond copying competitors features and being the junkyard dog in the marketplace).

I also think the thing that needs to be acknowledged by people like Ben is that first time entrepreneurs often don’t have access to “fat capital” until they have significant traction—which happens with a combination of luck, timing, and iteration. Famous entrepreneurs are not the mean.

I guess what I really want is a panel discussion with Ben Horowitz, Angus Davis (or Dave Weiden), Eric Ries, and Steve Blank. And it would be interesting for Ben and Angus/Dave (the Tellme team) to talk about pivoting a startup of considerable size and with considerable amounts of capital and contrast that to the product / market fit methods advocated by Eric and Steve.

If you’re looking for some interesting business reading, try to find out what you can about the RIfle framework employed by Tellme as it switched from a consumer business to powering the voice trees of major corporations. I guess the devilish follow up question to Angus/Dave/Ben is: how much of the acquisition value of Tellme was due to adopting a customer development strategy and the size of their business and how much of it was due to the size of their data asset (voice phonemes)—and if they had enough capital to capture the same amount of data as a consumer tool, what would have happened?