You're both right. You need the balls to trust your hunches, to put yourself out there, even when there's limited data. But you also need the balls to throw y

You're both right. You need the balls to trust your hunches, to put yourself out there, even when there's limited data. But you also need the balls to throw your previous assumptions out the window, in the face of new data. In the case of startups, it's true that everyone's advice is extrapolated from one or two data points. The key is to look at a lot of those data points -- read Founders at Work or listen to VentureVoice.com, for instance -- and try to find recurring themes and patterns. This is part of the idea behind YC, right? I get that impression from your essays. These days there's less "At Viaweb, we..." and more "Out of the N startups we've funded, X have..."