Co-founder conflict often starts as a decision nobody wrote down

By Meet Patel · 2026-10-03 · 6 min read

Harvard Business School professor Noam Wasserman has studied founders at length, and the finding he is best known for is blunt. As summarized by Built In from his work, 65% of high-potential startups that fail do so because of people problems. In a 2012 Stanford talk, the abstract of which is archived by Stanford, he made the same point: while some ventures succumb to product or market fit, the vast majority end because of people problems.

Two limits on that number are worth stating. It describes startups that failed within his sample of high-potential ventures, so it says nothing about the share of all startups that fail for these reasons. And "people problems" covers tension among founders and tension between founders and the people who join them. My argument sits inside that finding: a large share of founder-to-founder tension begins as a decision that two people deferred because discussing it felt awkward.

Why the awkward decisions get deferred

Wasserman names the mechanism for founders who already know each other. He calls it the Playing-with-Fire gap. In his words, "The Playing-with-Fire gap is the greatest for cofounders with a prior social relationship. They are the least likely to deal with elephants in the room and will suffer the most damage if business tensions undo their social relationships; yet their failure to deal with the elephants makes such business tensions all the more likely."

That sentence describes a loop. The closer the founders are, the more a conversation about equity or exit feels like a statement of distrust, so it is postponed. The postponement leaves the business tension in place, and the tension eventually arrives when the company is worth more and the relationship is under more strain. The decision that was cheap to make in month one is expensive in month eighteen.

Five decisions that tend to go unwritten

The pattern I would look for is a decision that the founders believe they made, because they said something about it once, and that neither has written down. Five categories cover most of what I would check.

1. Equity and vesting. Who owns what, and what happens to a founder's shares if they leave in month ten.
2. Roles and decision rights. Who has the final say on product, on hiring, on spending and on the customer, when two founders disagree.
3. Commitment. Full time or part time, other income, outside projects, and what happens if circumstances change.
4. Money. Salaries, loans from founders to the company, personal guarantees, and who covers a shortfall.
5. Exit. What happens when one founder wants out, stops contributing, or receives an offer to sell the company.

Each of these has an answer in both founders' heads. The trouble is that the answers can differ without either founder knowing, and a difference stays invisible until an event forces it into the open.

A worked example with illustrative numbers

Take two founders who split a company 50/50 on day one, each holding 5,000,000 shares. The figures here are hypothetical. Founder A leaves a job and works on the company full time. Founder B keeps a consulting contract and gives it three days a week, intending to move across later. Neither writes this down, because the split felt fair on the day.

At month eighteen, a buyer offers $1.2 million for the company. Founder A believes the work of the last year and a half justifies a larger share. Founder B believes the split was agreed and the days worked are irrelevant. Both are acting consistently with what they remember agreeing. The offer is attractive and the conversation is now a negotiation with a deadline attached.

Vesting is the mechanism that puts a price on this difference in advance. Take a hypothetical schedule of four years with a one-year cliff. Under that schedule, a founder who leaves in month ten keeps no shares, because the cliff has not been reached. A founder who leaves at month eighteen has vested 18 of 48 months, which is 37.5 percent, or 1,875,000 of the 5,000,000 shares. I am using this schedule to show the arithmetic, as an example only. The right terms depend on the founders and the market, and a lawyer should draft them.

A founder agreement checklist

The list below is what I would want written before the company has revenue. Each item should be a sentence or two, dated and signed by both founders, and the whole document should fit on two or three pages.

• Equity split, with the reasoning for it recorded in one line.
• Vesting schedule, cliff and what happens to unvested shares on departure.
• Each founder's role, with the areas where that founder has the final say.
• A tie-break rule for the areas where both founders claim authority.
• Time commitment, and how either founder can change it, including the notice required.
• Outside work and conflicts of interest, including any income that continues.
• Founder salaries and the conditions under which they start.
• Money lent to the company, repayment order and any personal guarantees.
• Intellectual property: confirmation that work done for the company belongs to the company.
• Exit: what happens if a founder leaves, is removed, or receives an offer to buy the company.
• A review date, at six and twelve months, to revisit anything that has changed.

The final agreement is a legal document, and what I am offering is a planning checklist. Have a lawyer in the relevant jurisdiction draft and review the final version. The checklist exists so that the lawyer's time is spent on drafting, with the founders arriving having already agreed on the substance.

How to have the conversation

A method I would use with two founders is to answer the checklist separately before talking. Each founder writes their own answer to every line, without seeing the other's. Then the two sheets are compared line by line. Every line where the answers match takes thirty seconds. Every line where they differ is a decision that was open all along, and the sheets make it visible while the company is still worth little and the stakes are low.

I would expect disagreement on only a few lines, and those lines are the valuable part of the exercise. It is easier to resolve three specific differences than a general sense that something is wrong. If a founder is unwilling to answer a line in writing, that reluctance is information as well.

Put the review dates in the calendar when you sign. Founders' circumstances change: someone gets a better offer, someone has a child, someone discovers they care more about the product than the business. A scheduled review gives each change a place to be discussed before it turns into resentment. This is the same discipline as naming the decisions a company is not making, applied to the two people at the top of it.

The cost of deferring the decision

A founder agreement takes an afternoon of discomfort, a few hundred words and some legal review. It costs more later because by then the company has value, the founders have positions, and every term is a negotiation with a price tag. Founders already pay a loneliness tax, and a co-founder dispute adds the loss of the one person who was supposed to share the load.

The principle I would take from this is that a decision between founders should be written down while it is still cheap, which means before anything has happened that would make either of them want a different answer. A conversation held in month one is a planning exercise. The same conversation in month eighteen is a dispute.

Written by Meet Patel — startup operator and growth strategist in Dubai.

Read on themeetpatel.com