More than 23% of cofounders are gone by year three. Not because the idea failed — because the partnership did.
That number should unsettle every founder currently shopping for a cofounder on LinkedIn or sizing up a college roommate over coffee. A cofounder exit at the wrong moment — and there is never a right one — doesn’t just leave an empty chair. It spooks investors, destabilizes the team and can gut a company’s momentum at precisely the moment it needs to accelerate.
And yet a lot of founders don’t do the work to prevent it.
The standard cofounder search is essentially a skills audit. One builder, one seller. Technical meets commercial. On a whiteboard, it looks like balance. In practice, it’s an incomplete picture, because startups don’t break down in the whiteboard phase. They break down when the runway shrinks, the product stalls and two people who’ve never disagreed about anything important suddenly have to.
That’s when you find out who you’re actually building with.
Here’s what nobody asks before signing a cofounder agreement: How do you behave when you’re overwhelmed? Do you get quiet or do you combust? When feedback lands hard, do you sit with it or shut down? When you think your partner isn’t pulling weight, do you say so, or do you let it calcify into something neither of you can name?
These aren’t soft questions. They’re risk assessment. The answers tell you whether this partnership will hold when it’s genuinely tested, which it will be sooner than either of you expects.
Take something as mundane as how two people define commitment. One founder counts hours: the 6 a.m. Slack messages, the canceled weekends, the visible grind. The other counts outcomes: the deals closed, the product shipped, the number on the dashboard. Neither is wrong. But in a hard quarter, those two founders are judging each other by different standards and neither one knows it. That’s not a communication problem. That’s a fault line.
The fix isn’t complicated, but it requires the one thing most early stage founders resist: slowing down.
Have the uncomfortable conversations before the company needs you to already have had them. Ask each other what you’re most insecure about. Talk about how you’ve each handled failure — not the polished version, the actual one. Discuss what happens when one of you thinks the other is wrong and won’t budge. Build a vocabulary for conflict before conflict arrives uninvited.
It is the deliberate, pre-commitment work of stress-testing a relationship before the stakes are real. It’s not soft. It’s the most operationally important conversation two cofounders can have.
The strongest founding teams aren’t the ones who never fight. They’re the ones who fight productively, who’ve built enough trust and enough shared language to move through disagreement without torching the relationship in the process.
Speed is the religion of the startup world. Ship fast, iterate faster. Good advice for product. Terrible advice for the person you’re about to share a cap table with for the next decade.
Slow down once. It might be the only decision that actually sticks.