The second outcome, was we looked back internally and every quarter we do this thing called an expectations analysis. We ask everyone at the firm to give a low, medium, and high expectation for every single company we’ve invested in. What’s interesting about that exercise is that you’re talking about a very diverse, eclectic, group of people, meaning that we have data scientists, engineers, associates, partners, investors, CFO, GC… What you get with that is a very asymmetric amount of information. The companies where I’m on the board, I have a lot of information. Even though I’m sharing the information out with everyone else, they don’t have time. And other people have different senses of it. Some people may use our products, some people may not.
What’s amazing now is that we have 18, 20, people, that’s enough signal to produce these really interesting data sets that occur. And we’ve been doing this now for four years.
When you look at all of the companies that are generating all of the returns for us, they’re all the ones with these large upper bounds and lower bounds of outcomes; the error bars are huge.
So, wherever there’s disagreement, there’s the ability to generate massive outcomes.
Wherever there’s consensus, there’s still a lot of value, but they tend to not surprise anybody.
Typically what happens is that companies spend many years in these periods of disagreement and lack of consensus about what they’re really doing, and then they get into this breakout velocity and all of a sudden everyone says, “Oh!” They anoint them, whatever.
So I said, “Look. As a series A and B investor, we need to live in that world of ambiguity. How do we do that?”
The only way that I saw that we did that is by putting a bunch of different people together that don’t necessarily think the same way and act the same way.