Co-Founder Conflict: The Complete Guide to Repairing a Founding Partnership

A note on scope: this guide is for business co-founders — people who built a company together and are not romantic partners. If you are married to your co-founder, the dynamics are different and we cover them separately in co-founder couples.

You did not start this to manage a relationship. You started it to build something, and the person across the table was the reason it was possible: they saw it, they could do the half you could not, and they were willing to be poor and uncertain alongside you.

Now there is a subject you both steer around. Decisions get reopened a month after they were made. One of you has started routing around the other — looping in a third person, sending the Slack message instead of having the conversation, framing a choice as already decided. The company still ships. Board updates still go out. And somewhere underneath, the partnership that made the whole thing possible is running on fumes.

This is the most ordinary crisis in startups and the least discussed. It is also, more often than founders expect, fixable — but not with the tools you have been using.

Short answer: is co-founder conflict normal?

Yes, and that framing is part of the problem.

Disagreement between co-founders is not only normal, it is the point. Two people who always agree are one person with redundant payroll. The founding pair exists so that strategy gets stress-tested before the market tests it.

What is not normal, and what is worth taking seriously, is when disagreement stops producing decisions. The signal is not volume. Loud founding teams can be extremely healthy. The signal is whether a disagreement ends in a decision both of you will actually execute, or whether it ends in a truce that gets quietly relitigated.

A working partnership can argue hard on Tuesday and ship on Wednesday. A broken one is polite in the meeting and negotiating in the corridor.

How common is co-founder conflict?

Common enough to be one of the leading causes of early-stage failure, and far more common than the public record suggests.

The figure most often quoted comes from Noam Wasserman's research at Harvard Business School, published in The Founder's Dilemmas: in his data on high-potential startups, roughly 65 percent of failures were driven by conflict among the founders, not by the market or the product. Wasserman also found that founding teams made up of friends or family were the least stable, largely because people who care about each other avoid the hard business conversations. Post-mortem studies of failed startups put "team problems" in roughly one in five to one in four cases, which is a narrower measure but points the same way.

Two things follow from those numbers. First, the strain you are feeling is not a sign that you chose badly; it is the base rate. Second, almost nobody talks about it while it is happening, because admitting founder conflict feels like admitting the company is at risk. That silence is why it feels anomalous when it happens to you, and why most founding teams reach for help years later than they should.

Why founding partnerships break differently than other working relationships

Three things make the founding relationship structurally unlike any other professional one.

There is no manager. In every other working relationship you have ever had, an unresolved conflict could be escalated. Somebody above you both had the final word. At the founding table there is nobody above you, which means every unresolved disagreement stays in the room forever. Nothing gets settled by authority, so it either gets settled by agreement or it does not get settled at all.

You cannot leave without enormous cost. A bad manager you quit. A bad co-founder is attached to your equity, your reputation, your investors and several years of your life. The exit cost is so high that founders routinely tolerate a partnership they would never tolerate in any other context, for far longer than is good for either of them or the company.

The stakes contaminate the conversation. Every disagreement is also, implicitly, a conversation about who was right about the thing before, who is carrying more, and whose judgment the company should run on. That is why an argument about a hiring decision can feel wildly disproportionate. It is never only about the hire.

Put together: an unleavable relationship with no referee where every ordinary decision carries a hidden referendum on competence. That is not a personality problem. That is a structure that will reliably produce conflict in capable, well-intentioned people.

AET01 Conflict

Fix the fight before you have it.

Rehearse it against an AI partner who doesn't back down. Two minutes, a score, and the line that would have held.

Open AET01 ↗Coming to iOS and Android

The six patterns behind most co-founder conflict

The Six Patterns of co-founder conflict, as we name them at Aether Practice, are: the contribution ledger, the 50/50 deadlock, role drift, divergent ambition, the avoided subject, and trust erosion by small increments. Almost every founding-team breakdown we see is one or more of these. They are patterns, not personalities — which is why they respond to changes in how the two of you operate rather than to one person trying harder to be reasonable. If you want to check which ones are already present, the twelve warning signs are the observable version of this list.

1. The contribution ledger

One of you believes you are carrying more. Possibly you are. The ledger is almost never spoken aloud, because saying it sounds petty, so it gets expressed sideways: in tone, in a pointed comment about hours, in reluctance to let the other make a call.

The dangerous property of a contribution ledger is that it is private, so it never gets corrected. Both founders usually believe they are carrying more, and both are drawing on different evidence. One counts hours. The other counts risk, or relationships, or the decisions that avoided disasters nobody can see. When the ledger becomes the main event, it is usually a conversation about one co-founder not pulling their weight, and that conversation has its own rules.

2. The 50/50 deadlock

An even equity split with no tiebreaker means that any genuine disagreement about strategy, hiring, fundraising or an acquisition offer has no resolution mechanism other than one person giving in. Over time, the person who gives in more starts keeping score — and you are back in pattern one.

This is a design flaw, not a character flaw, and it is worth separating from the emotional content around it. We cover the deadlock and its fixes in detail here.

3. Role drift

You split responsibilities at the start based on who you were then. Two years later the company needs different things, both of you have grown in different directions, and nobody has renegotiated. So you end up with overlapping decision rights in some areas and gaps in others, and every collision in the overlap feels like an incursion.

4. Divergent ambition

One of you wants a profitable company you own. The other wants a venture outcome. Or one wants to sell at a number the other considers an insult. This divergence can stay invisible for years because the near-term work looks identical — until a term sheet or an acquisition offer forces it into the open, at the worst possible moment.

5. The avoided subject

There is one conversation the two of you are not having. Equity. Someone's performance. Whether one of you still wants to be here. The avoidance is rational at first — raising it feels destabilising when you are mid-raise or mid-launch — but the cost compounds. Everything else in the relationship gets slightly more careful, because every conversation now has to navigate around the thing in the middle of it. When the avoided subject is the cap table, renegotiating co-founder equity is its own conversation with its own sequence.

6. Trust erosion by small increments

No betrayal. Nothing you could point to. Just an accumulation: a commitment missed, a decision made without you, a version of events given to the team that was not quite how it happened. Each one too small to raise. Together, they change how you listen to the other person, which means you now interpret neutral things uncharitably, which produces more of the same.

The thing underneath: unspoken terms

Unspoken terms are the implicit agreement every founding team runs on and almost none writes down: how hard is hard enough, who decides when you disagree, what you are building toward, what would make either of you leave, and how you disagree in front of the team. Underneath every one of the Six Patterns is the same mechanism.

When you started, you agreed on the explicit terms: equity, titles, roles, maybe a vesting schedule. You did not agree on the implicit ones, because at the time you did not know they existed. Things like:

• How hard is hard enough, and does that change when someone has a child?
• Who gets the final word when we genuinely disagree, and in which domains?
• What are we building toward, in numbers and in years?
• What does one of us have to do before the other would consider leaving?
• How do we disagree in front of the team?

Those are the real operating agreement, and almost no founding team writes it down. Instead each founder carries a private version, assumes it is shared, and then experiences the other person's different version as a breach rather than a difference. We have written the full list of questions that agreement should answer, for teams that want to write theirs down before it is tested.

Most co-founder conflict is not two people failing to get along. It is two people discovering, several years in and under pressure, that they signed different contracts.

Why advice from investors and advisors usually makes it worse

When the partnership starts to strain, the instinct is to ask the people around the company. This is understandable and mostly counterproductive, for reasons worth naming.

Investors are not neutral and cannot be. Their fiduciary duty is to the asset, not to either of you. A board member who likes you both will still, correctly, optimise for the company. If one founder is the problem for the company, the advice will reflect that, and it will reach you as advice about your partnership. Founders also know this instinctively, which is why they present a managed version of the conflict to the board — which means the advice is based on incomplete information anyway.

Advisors and friends take a side. Usually the side of whoever told them first. Once each founder has their own set of sympathisers, the conflict acquires a constituency and becomes much harder to resolve privately.

Lawyers solve a different problem. A good startup lawyer can fix the structure: vesting, deadlock provisions, buy-sell mechanics, a clean separation. What a lawyer cannot do is tell you whether the partnership is worth keeping. If you bring a repairable relationship to a lawyer first, you will get a structural answer to a relational question, and the process of getting it will often damage the thing you were trying to save.

Each of these has a real place. None of them is the place to find out what is actually happening between the two of you.

Self-check: is this a bad quarter or a broken partnership?

The Repairability Check is eight yes-or-no questions; zero to two yeses is an ordinary stretch under load, three to five is a pattern that will not resolve by itself, and six or more means the partnership is running on avoidance and needs someone outside it in the room. Honest answers only. Nobody is reading them.

1. Is there a subject the two of you have been avoiding for more than a month?
2. When you disagree, does it end in a decision, or in a truce?
3. Do decisions get reopened weeks later?
4. Have you started preparing arguments before conversations?
5. Do you route important things through someone else rather than raise them directly?
6. Does the team know there is tension? Have they started managing around it?
7. Would you choose this person again, knowing what you know?
8. Do you believe they would choose you?

Zero to two is an ordinary stretch under load. Talk about the one subject and move on.

Three to five is a pattern, not a mood, and it will not resolve by itself. This is the point where a structured conversation is cheap and effective.

Six or more means the partnership is operating on avoidance. It may well still be repairable — many are — but not through the normal channels, and not without someone outside the two of you in the room. If you are already asking whether it is over, how to tell a finished partnership from a long bad stretch is the next thing to read.

What does not work

Waiting for a quieter quarter. There is no quieter quarter. There is the current fire and the next one. Teams that wait for calm to have the hard conversation have the hard conversation during a crisis instead, which is the worst possible time.

The offsite. A day away with a whiteboard produces real feeling and no mechanism. Everyone leaves moved and nothing changes, because the thing that was broken was not goodwill — it was the absence of a way to decide.

Working harder to be reasonable. Both of you are already being reasonable by your own lights. More self-restraint from the person who already gives in most produces exactly one outcome, which is a deeper ledger.

Dividing the company into territories. Structuring the work so you do not have to deal with each other reads as maturity and functions as a slow separation. It also makes the company worse, because the areas that need both of you stop getting both of you.

Settling it over drinks. Alcohol is excellent at producing warmth and terrible at producing terms. The reconciliation is real. It is also gone by Thursday, because nothing was written down and nothing was decided.

What works: seven practices for founding teams

The seven practices, in one line each: name the pattern not the person; write the implicit agreement down; give every domain a single owner; separate the disagreement from the scorecard; install a tiebreaker before you need one; schedule the partnership, not just the business; and rehearse the conversation you are dreading. In detail:

1. Name the pattern, not the person. "We reopen decisions" is a problem two people can attack together. "You keep second-guessing me" is an accusation that requires a defence. The first produces work. The second produces a worse Tuesday.

2. Write the implicit agreement down. Actually write it: how hard is hard enough, who decides what, what we are building toward, how we disagree in front of the team. The value is not the document. The value is discovering, in the writing, every place you assumed you agreed and did not.

3. Give every domain a single owner. Not a split, an owner. The other person gets full input and no veto. Shared ownership of a decision sounds collaborative and reliably produces deadlock. Two founders who each truly own half of the decisions argue far less than two who jointly own all of them.

4. Separate the disagreement from the scorecard. Before the argument, say out loud what it is about. "This is about the pricing call, not about whether I trust your judgment." It sounds artificial. It also stops a tactical disagreement from becoming a referendum.

5. Install a tiebreaker before you need one. A trusted third party, a domain-owner rule, a rotating final word — the mechanism matters less than agreeing on it while you are calm. Deciding how to break a tie during a tie is how founding teams end up in lawyers' offices.

6. Schedule the partnership, not just the business. A standing hour, outside the operating cadence, where the only agenda is the two of you; the weekly co-founder sync sets out the agenda and the questions. Most founding teams have four weekly meetings about the company and none about the thing the company depends on. Y Combinator's partners and the coaches who work with its companies give the same advice: most founder conflict is a process problem before it is a relationship problem, and a weekly founder sync is the cheapest process there is.

7. Rehearse the conversation you are dreading. The equity conversation, the performance conversation, the one about whether you still want to be here. People who practise a high-stakes conversation before having it hold their position better and concede less for the sake of quiet. This is the same reason pilots use simulators and litigators run moots, and it is the one preparation almost nobody applies to the conversations that decide how a company goes. You can rehearse in a session with us, or against a model built to hold a position: AET01 Conflict, our two-minute conflict rehearsal app, scores both losing moves — escalating, and quietly giving in.

What co-founder coaching is, and when to bring it in

Co-founder coaching is structured work on the founding partnership itself — how two or more founders decide, disagree, divide responsibility and repair after conflict — with both founders in the room and a practitioner whose only loyalty is to the working relationship. It is not executive coaching, which works on one person. It is not mediation, which settles a defined dispute. It is not therapy, and it is not clinical care; the content is operational, not personal. Some founders call it co-founder therapy, because the methods borrow from couples work, but the subject is decision rights, contribution and direction rather than intimacy.

The useful distinction between the kinds of help is what each one can actually change.

A lawyer changes the structure. Vesting, deadlock provisions, buy-sell terms, a clean exit. Bring one in when you have decided what you want and need it to hold. Not before.

A mediator resolves a specific dispute. They are the right call when there is a defined thing to settle — an equity disagreement, the terms of a departure — and the working relationship is not the thing you are trying to repair.

Co-founder coaching works on the partnership. It is the right call when the company is fine but the partnership is expensive, when you keep having the same argument, or when you cannot tell any more which of your problems are strategic and which are relational. It is also the right call before any of that, which is when almost nobody books it.

We go through what each one fixes, what each one costs, and how to tell which you need, here.

One thing worth saying plainly: bringing in help is not an admission that the partnership is failing. The founding relationship is the single highest-leverage asset in an early company, and it is the only one most teams never invest in deliberately.

For founding teams in San Francisco, New York and beyond

Most of the founding teams we work with are in the San Francisco Bay Area and New York City, and the two ecosystems produce the same conflict with a slightly different accent.

In San Francisco and the wider Bay Area, the ecosystem is small enough that your investors, your advisors and your co-founder's friends overlap, which makes it almost impossible to find a neutral person inside the circle — and makes the board the last place you want a half-formed version of the conflict to land. In New York, founding teams more often sit across industries rather than inside one, and the pressure tends to arrive through fundraising cadence and a culture that treats visible disagreement as weakness. In both cities the fix is the same: a room with nobody from the ecosystem in it.

Because the work is online, a founding team in SoMa, Brooklyn, Palo Alto or Oslo gets the same session, and co-founders in different cities — increasingly the norm — can be in the same conversation without anyone flying. Here is what co-founder coaching looks like for founding teams in San Francisco and New York, what it costs, and how to choose a practitioner.

FAQ: co-founder conflict

How common is co-founder conflict?

Very. Noam Wasserman's research at Harvard Business School, published in The Founder's Dilemmas, attributes roughly 65 percent of high-potential startup failures to conflict among the founders, and post-mortem studies of failed startups list team problems in roughly a fifth to a quarter of cases. Most founding teams hit a serious strain in the first three years. Few of them talk about it publicly, which is why it feels anomalous when it happens to you.

What are the most common causes of co-founder conflict?

Six patterns account for most of it: a private contribution ledger, a 50/50 split with no tiebreaker, role drift as the company changes, divergent ambition about what the company is for, an avoided subject (usually equity or performance), and trust eroded by small increments rather than one betrayal. Underneath all six is the same mechanism: the two of you signed different implicit contracts and are discovering it under pressure.

Can a co-founder relationship actually be repaired?

Often, yes — particularly when the conflict is structural (deadlock, role drift, an unspoken ledger) rather than a fundamental divergence in what the two of you want. The partnerships that do not recover are usually the ones where ambition has genuinely diverged, or where trust was broken by something specific and large rather than eroded by increments.

What is co-founder coaching?

Structured sessions with both founders present, focused on the working partnership: decision rights, the unspoken ledger, the avoided conversation, and how you repair after a disagreement. It is forward-looking and operational. It is not executive coaching for one person, not mediation of a single dispute, and not therapy or clinical care. Some people call it co-founder therapy because the methods borrow from couples work; the content is entirely different.

How much does co-founder coaching cost?

Practitioners who publish prices in 2025 and 2026 typically charge retainers of roughly $2,000 to $9,000 a month or packaged programmes of around $6,000 for six sessions. Aether Practice offers a 60-minute online co-founder session for $170, with a free 15-minute consultation first, so a founding team can find out whether the work fits before committing to anything.

Should we involve our investors?

Not as a first step. Investors are not neutral parties and cannot be; their obligation runs to the company. Work the problem privately first, and bring the board a resolution rather than a conflict.

Is this the same as couples therapy for co-founders?

Some methods overlap, because both deal with two people in a high-stakes relationship with no referee. The content is entirely different. Co-founder work is about decision rights, contribution and direction, not intimacy. This is coaching for a working partnership, not clinical care. If you are married to your co-founder, the two overlap in your case, and we cover that separately.

What if only one of us thinks there is a problem?

That is itself one of the six patterns. The founder who does not see a problem is usually the one whose preferences are currently winning. Start by naming the pattern rather than the complaint, and be specific about what you are seeing rather than how you feel about it. Framed as "we keep getting stuck in the same place and I would like help with it", most co-founders will come to a first session.

When is it time to separate?

When what the two of you want from the company has genuinely diverged and no structure reconciles it, or when trust has been broken in a way neither of you can rebuild. Separating well is its own skill, and done properly it protects the company. We cover that here.

Do you work with founding teams in San Francisco and New York?

Yes. Most of the founding teams we see are in the Bay Area and New York City, and sessions are online, so co-founders in different cities can be in the same room. We also work with teams in Norway and elsewhere in Europe across time zones. Book a free 15-minute consultation to find out whether it fits.

Next step

If the two of you have been circling the same conversation for months, that is the conversation to have — with someone in the room whose only interest is the working relationship.

Aether Practice offers co-founder sessions for founding teams in San Francisco, New York City and Norway. Both of you, online, working on how you decide, disagree and divide the weight — before the company starts routing around it.

The rest of this cluster: Twelve warning signs · The weekly co-founder sync · Glossary of co-founder conflict terms · The 50/50 deadlock · When one co-founder stops pulling their weight · Renegotiating co-founder equity · Mediation, coaching or a lawyer · Co-founder coaching in San Francisco and New York · The co-founder agreement questions · Co-founder breakup · Co-founder couples

Research and further reading

• Wasserman, N. (2012). The Founder's Dilemmas: Anticipating and Avoiding the Pitfalls That Can Sink a Startup. Princeton University Press. The source of the widely cited finding that roughly 65 percent of high-potential startup failures trace to founder conflict, and of the data showing that founding teams built on prior friendships or family ties are the least stable.
• Fetsch, E. Can Relationship Counseling Save Your Startup? Ewing Marion Kauffman Foundation. On the rise of couples-style work with co-founders and the role of unclear roles in founder conflict.
• Gottman, J. M. & Silver, N. (1999). The Seven Principles for Making Marriage Work. On repair attempts and turning toward, the method much co-founder work borrows from.
• Rosenberg, M. B. (2003). Nonviolent Communication: A Language of Life. The observation, feeling, need, request structure used in many founder conversations.
• Also read: Co-founder coaching · Couples therapy for founders · Co-founder couples

Last updated: October 2026.