Why Startups Really
Fail: The Cofounder
Data

Most startups don't die from a bad product. They die from the people building it. Harvard Business School professor Noam Wasserman found that 65% of high-potential startups fail because of conflict among cofounders — not because the market rejected them, and not because the code didn't work. The hard thing is the relationship, and almost nobody plans for it.

The number everyone quotes, and what it means

Wasserman studied what he called high-potential startups — companies with real funding, real teams, real shots at scale. His finding, published in his book The Founder's Dilemma, is that 65% of them fail due to disputes among cofounders. Not funding. Not competition. Each other.

That framing matters. These weren't hobby projects that fizzled. They were startups with a genuine chance, undone by the one variable founders treat as an afterthought: whether the people at the top can actually work together.

The conflict rarely arrives as a single blowup. It compounds. As Wasserman describes it, two or more sets of values, motivations, and expectations get merged, and the small misalignments that felt harmless at the start grow into the thing that ends the company. You don't see it coming because you're looking at the roadmap, not the room.

The failure data, side by side

Put the numbers together and a pattern shows up. The proximate causes get all the attention. The root causes are quieter, and more human.

Read that table twice. CB Insights analyzed 431 shuttered venture-backed startups and found that running out of capital tops the list at 70%. But their own analysts flag it as where the story ends, not where it starts. Cash is the last domino. The question is what pushed it.

Sometimes the answer is the market — poor product-market fit shows up in 43% of failures. Sometimes it's timing, at 29%. But sit those alongside Wasserman's 65% and the picture sharpens. A team that can't align burns cash faster, reads the market slower, and stalls on the decisions that would have saved it. The cofounder problem isn't a separate category from the money problem. It's often the reason the money problem happened.

Why people problems beat product problems

Here's the uncomfortable part. Product problems are legible. You can see churn, you can read a sales pipeline, you can run a pricing experiment. People problems hide.

A founding team can look aligned for a long time while quietly drifting apart. One founder assumes they're the CEO forever. The other assumes leadership rotates with the company's needs. Neither says it out loud. The equity split made sense on day one and feels unfair by month eighteen. Decision rights were never written down, so every ambiguous call becomes a small negotiation about power. None of this shows up on a dashboard.

That's why the conflict data is so lopsided toward people. It isn't that products don't fail — they do. It's that teams avoid the conversations that would surface the real risk. Founding partnerships often unravel because the founders stop talking straight with each other, not because the market moved. The tough conversation gets deferred to protect the relationship, and the deferral is what kills it.

Investors know this, even when founders don't. When a fund evaluates a seed-stage company, the team is frequently the deciding factor, because at that stage the team is most of what exists. If you want to see your partnership the way capital sees it, start with how investors evaluate the founding team.

The conflicts that do the most damage

Not all friction is fatal. Some is just how work gets done. The dangerous conflicts cluster around a few predictable fault lines.

Roles and control

Who decides what. It sounds trivial until the company grows past the point where everyone can be in every decision. Ambiguity about leadership is one of the most common sources of founder disputes, and it tends to get worse as the stakes rise, not better.

Equity and contribution

Splits set at the beginning assume everyone will contribute equally forever. They almost never do. One founder goes full-time, another stays part-time. One raises the money, another builds the product. The gap between the cap table and the perceived effort becomes resentment, and resentment is corrosive.

Vision and pace

One founder wants to raise and scale. The other wants to stay lean and profitable. Both are reasonable. Together, unspoken, they pull the company in two directions until something snaps.

The through-line is that all three are avoidable with early, direct conversation — and lethal without it. If you're already feeling the strain, the earlier you name it the better; these are the signs a cofounder relationship is starting to fail before it becomes terminal.

What the data should change about how you build

The lesson isn't "pick a better cofounder." Plenty of failed partnerships started between capable, well-matched people. The lesson is that the relationship needs the same rigor you give the product.

You wouldn't ship a feature without knowing whether it works. Don't run a partnership without knowing whether it's aligned. Get the hard questions on the table before you need the answers: what happens if one of us wants to leave, how do we make decisions we disagree on, what does the equity mean if contributions change. The teams that survive aren't the ones that never disagree. They're the ones that built a way to disagree without it becoming an ending.

Wasserman's 65% is not a prophecy. It's a warning about the gap between how much attention founders give the product and how little they give the partnership. Close that gap and you move yourself out of the majority. When friction does show up — and it will — having a structured way through it matters more than pretending it won't happen; start with a real approach to cofounder conflict resolution rather than hoping it passes.

The startups that last aren't the ones with no conflict. They're the ones that saw it coming.