Picking a co-founder is your most important decision. It’s more important than your product, market, and investors.
The ideal founding team is two individuals, with a history of working together, of similar age and financial standing, with mutual respect. One is good at building products and the other is good at selling them.
The power of two
Two is the right number — avoid the three-body problem. Think Jobs and Wozniak, Allen and Gates, Ellison and Lane, Hewlett and Packard, Larry and Sergei, Yang and Filo, Omidyar and Skoll.
One founder companies can work, against the odds (hello, Mark Zuckerberg). So can three founder companies (hello, @biz, @ev, and @jack). In three founder companies, the politics can be tough — gang-up votes, jockeying for board seats, etc. — but it’s manageable. Four is an extremely unstable configuration and five is right out. When 4–5 founder companies work, it’s because two founders dominate.
Two founders works because unanimity is possible, there are no founder politics, interests can easily align, and founder stakes are high post-financing.
Someone you have history with
You wouldn’t marry someone you’d just met. Date first. Guess which pair of famous co-founders is in this photo:
One builds, one sells
The best builders can prototype and perhaps even build the entire product, end-to-end. The best sellers can sell to customers, partners, investors, and employees.
The seller doesn’t have to be a “salesman” or “business guy”. He can be technical, but he must be able to wield the tools of influence. Bill Gates and Steve Jobs aren’t salesmen, but they are sellers.
Aligned motives required
If one founder wants to build a cool product, another one wants to make money, and yet another wants to be famous, it won’t work.
Pay close attention — true motivations are revealed, not declared.
Criteria: Intelligence, energy, and integrity
It’s not the kid you grew up next to. It’s not the person you like the most. It’s not the hacker most willing to work for free.
It’s someone of incredibly high intelligence, energy, and integrity. You’ll need all three yourself, and a shared history, to evaluate your co-founder.
If it doesn’t feel right, keep looking. If you’re compromising, keep looking. A company’s DNA is set by the founders, and its culture is an extension of the founders’ personalities.
Pick “nice” guys
Avoid overly rational short-term thinkers. There are bounds to rationality. Partner with someone who is irrationally ethical, or a rational believer that nice guys finish first. Be especially careful with the “sales” guy here.
What you don’t know
Business founders who don’t code use bad proxies for picking technical co-founders (“10 years with Java!”). Technical founders who don’t sell also use bad proxies (“Harvard MBA!”). Learn enough of the other side to have an informed opinion. If you’re not seriously impressed, move on.
What if the right guy already has his own startup? Convince him to work on yours part-time — he’ll drop his idea once yours gets traction.
Breakups are hard
If you’re going to fall out with your co-founder, do it early, recover the equity into the option pool to keep the company going, and recruit someone else great to fill the missing slot. Build in founder vesting (a.k.a. the “Pre-Nup”) to keep the breakup from getting messy. Building a great company without a partner is like raising kids without a…
Nearly everything I’ve written on this topic applies to dating and marriage. Coincidence?
Go forth and multiply.
Update: Also see our 40-minute interview on this topic.