Best And Worst Decisions - Paul Graham

A few years ago, for Startupping, I asked several entrepreneurs about their best and worst decisions. With the shuttering of that site, their answers vanished from the web. I’ll be reposting them here. This was originally posted on February 20, 2007:_

Paul Graham is an essayist, programmer, and programming language designer. He is currently a partner in Y Combinator, an innovative venture firm specializing in funding early stage startups. He is also a cofounder of Startup School, which this year is on March 24, 2007 at Stanford. Previously, he co-developed Viaweb, the first web-based application, which was later acquired by Yahoo, and more recently he pioneered the Bayesian spam filter, which inspired most current spam filters.

The best decision I made was to make Viaweb web-based. There were no web-based applications then, so we weren’t sure such a thing would even be possible. Initially what drove us was our dislike of Windows. Writing a desktop application would have meant learning Windows, which we really didn’t want to do. Whereas servers were the same Unix machines we used every day. To make a web-based application, all we had to do was figure out how to let users drive our software by clicking on links on web pages. That was a lot less work than learning Windows.

Hmm, no, actually the best decision I made was to get two fabulously good programmers, Robert Morris and Trevor Blackwell, to start the company with me.

The worst mistake I made, probably, was not being strong enough with investors. I now realize that investors like you to be assertive. It reassures them when founders take charge. But because our investors were so much older than us and had given us what seemed then unimaginably large sums of money, I felt I ought to defer to them. And yet I wasn’t prepared to do things their way in anything really important, like what the software should do or what our strategy should be. This inconsistency led to disputes that sucked up a lot of time and energy.

I realize that’s not really a decision. It was more something I didn’t do than something I did. But I think the worst mistakes startups make are mostly of that kind. Another big mistake I made was not to investigate IP agreements signed by people we hired. That nearly sank us later. But I didn’t decide not to; I just didn’t pay enough attention to it.

_