Always Be Closing: Y Combinator and The Art of the Pitch
Two founders. Two:
When Y Combinator began, Paul Graham set down a formal rule: no funding of startups with only one founder. He later relented a bit, allowing occasional exceptions in cases where the founder seemed extraordinary. It was a good thing he did relent— one if those exceptions was Drew Houston, the sole founder of YC’s largest success: Dropbox. But Houston did not remain the sole founder— Justin. tv’s Kyle Vogt introduced him to Arash Ferdowsi, who became a cofounder. If anyone doubted the need for at least two founders, Graham would list the technology giants whose public identity is bound to a single founder— Microsoft (Gates), Apple (Jobs), Oracle (Ellison)— and then point out that each one initially had two founders, not one. All startups are the same, he maintains. You need two people to spread the load.
Graham could enumerate other reasons for not wanting to fund single founders. The very fact that a founder did not have cofounders was evidence, in his view, that the founder’s friends had withheld placing their confidence in the founder. “That’s pretty alarming,” Graham said in 2006, “because his friends are the ones that know him best.” Even if friends had underestimated the founder’s ability to make a success out of a startup, and even if the founder could handle all of the startup’s tasks alone, the founder still needed others, Graham said, “colleagues to brainstorm with, to talk you out of stupid decisions, and to cheer you up when things go wrong.” More than one founder was essential, but not many more. Harj Taggar, addressing aspiring founders in London in 2011, said that YC’s experiences showed that four founders was too many— decision making was too cumbersome. “You don’t want your startup to be like the UN, some sort of democratic process to get the smallest thing done.”
A startup could change its idea easily, but not its cofounders. It was critically important that the two or three founders knew and liked each other well before starting out. Graham said, “Startups do to the relationship between the founders what a dog does to a sock; if it can be pulled apart, it will be.”
http://foundersdilemmas.com -- very interesting. Quant guy at HBS goes after the question. The free chapter is well worth a read.
Bottomline: Two are better than one or three. They should know each other, but not too close.
Damn sensible, but just like families, they aren't all exactly as they're supposed to be.
David's right that pg revised his thinking and now he says two or three founders.
I agree with you, Barbara. Two is ideal. With three, two will often gang up on one.