How to Handle a Cofounder Who Isn't Pulling Their Weight
This is the cofounder problem founders sit on longest, and the reason is structural: there is no mechanism for it. You can put an employee on a performance plan. There is no equivalent for a cofounder, no review cycle, no manager, no obvious moment to raise it. So it accumulates — and effort resentment compounds faster than any other kind, because every week adds a fresh instance and none of them get discussed. This page is about checking your read first, identifying which of five things is actually happening, and building the mechanism that does not currently exist.
What’s inside
Check the measurement before the conclusion
Five causes, five different fixes
Why there is no mechanism, and how to build one
The conversation
What a real agreement looks like
If nothing changes
Where Blomma fits
Check the measurement before the conclusion
Your read here is systematically distorted, and it is worth knowing how before you act on it.
You have complete information about your own effort and partial information about your cofounder’s. You know every late night you worked and every difficult call you took. You saw a fraction of theirs. That asymmetry is not a character flaw; it is the default condition of two people working on different things, and it reliably produces two founders who each privately believe they are carrying more.
Visibility also varies enormously by type of work. Shipping code, closing deals, and posting publicly are visible. Recruiting, thinking, managing a difficult person, and preventing problems are close to invisible. A founder doing the second kind can look idle while doing the harder job, and a founder doing the first kind can look heroic while doing the easier one.
So before concluding anything, do a specific test: write down what you believe your cofounder has actually done in the last month, then ask them to tell you. Not as a challenge — as information. Founders who do this frequently discover a category of work they had not counted, and occasionally discover the opposite, which is also useful because it is now evidence rather than an impression.
This is not a reason to dismiss your perception. Unequal load is real and common. It is a reason to establish it as fact first, because a conversation built on a mistaken read is unrecoverable.
Five causes, five different fixes
If the gap is real, five things could be producing it. They look identical from outside and resolve completely differently.
Unclear.They do not know what is expected, because it was never defined. Extremely common in founding teams, where nobody has a job description and “own go-to-market” means whatever each person assumes. The fix is defining outcomes and metrics, not increasing pressure. A founder who has never been told what the bar is cannot reasonably be judged against it.
Unable.They are trying and the role has outgrown them. The output looks like low effort and is actually low capability at the current stage, often compounded by them working longer hours on the wrong things. The fix is a role change or real support, and treating it as an effort problem will destroy the relationship while fixing nothing.
Unwell.Burnout, a health issue, a family situation, depression. Presents identically to disengagement and is far more common than founders assume, partly because a struggling founder is the last person likely to volunteer it. This is the one where getting it wrong causes the most damage, which is why the first conversation should ask before it tells.
Unwilling.They have genuinely disengaged — lost interest, mentally moved on, or are working on something else. Real, and the resolution is an exit rather than a performance conversation. Managing a disengaged cofounder for another two years serves nobody, including them.
Unequal by design, and nobody updated the terms.The split was always lopsided — one of you took the operating load, one took a narrower or more advisory role — and the arrangement was never made explicit. Here the problem is not their effort; it is that the terms do not reflect reality. The fix is renegotiating the arrangement honestly, which may mean adjusting scope, title, or compensation rather than asking for more work.
Ask before you decide which. The first conversation’s job is diagnosis, and founders who arrive with a verdict get a defence instead of information.
Why there is no mechanism, and how to build one
The absence of a mechanism is the actual problem, and it is solvable.
Employees have a manager, expectations, a cadence, and a documented process. Cofounders have none of these with respect to each other, which means raising performance requires inventing a moment, and inventing a moment is itself an escalation. That is why founders wait for a trigger — and the trigger, when it comes, is usually an incident that makes the conversation adversarial.
Build the mechanism before you need it. Three components.
Outcomes with owners and metrics.Each founder owns a small number of outcomes with something observable attached. This is what makes it possible to talk about performance as a fact rather than a feeling. Without it, every such conversation is one person’s impression against another’s.
A standing founder session that is not a status meeting.Weekly or fortnightly, explicitly for things that are drifting. Its existence means raising something does not require creating an occasion.
A stated route for concerns about each other.Agreed in advance: how either of you raises a worry about the other, and what happens next. Written into your working agreement. Deciding this while things are fine is trivial; deciding it during a grievance is another fight.
If you are reading this in the middle of the problem, you can still build these. Proposing the mechanism is a less confrontational opening than the grievance itself, and it creates the room the conversation needs.
The conversation
The shape matters more here than almost anywhere else, because the accusation “you are not working hard enough” is nearly impossible to receive well.
Open with observation and a question, not a conclusion. “Sales hiring has been open for three months and I don’t think it’s moving. What’s going on with it?” That is answerable. “I feel like I’m carrying this company” is a feeling that invites either guilt or denial, and produces neither information nor change.
Lead with specifics, three of them, factual. Instances can be discussed; characterisations can only be disputed.
Ask about cause before you propose a remedy, and mean it. This is where you find out which of the five you are in, and it is the step founders skip because they have already decided. If the answer is burnout or a health situation, everything you had planned to say is wrong.
Own your part. You almost certainly never defined the expectations, never raised it earlier, and let it run for months. Saying so is not a concession that undermines your position; it is what makes the rest credible, and they will raise it if you do not.
And say what you actually want. Not “more effort” — a specific change: this outcome owned, this metric moving, this decision taken by you. Vague asks produce vague compliance and the same conversation in three months.
What a real agreement looks like
End with something concrete, or you will be here again.
Name the outcome that changes, the observable measure, and the date you will both look at it. Write it in a shared doc the same day — two paragraphs, not a document. Verbal agreements between founders reliably evaporate and then return as a dispute about what was agreed.
Include what support is being provided. If the cause is unclear or unable, the agreement should contain something you are doing differently too — a hire, coaching, a scope change, clearer expectations. An agreement where only one person changes is a warning, not a plan.
And set the review close: six weeks, not six months. Short enough that a drift is caught while it is small, and long enough for something to actually change.
If nothing changes
If you have diagnosed honestly, defined expectations, provided support, and set a review, and the review shows nothing has moved — that is real information, and it is the point where the question shifts from performance to arrangement.
The options are the ones that exist for any cofounder mismatch: a role that fits what they will actually do, a renegotiation of terms so the arrangement matches reality, or parting ways. What is not an option is a third identical conversation, which is how founding teams spend years.
One honest note: by this stage many founders have discovered that what they wanted was not more effort from their cofounder but a different arrangement entirely — fewer expectations of them, and terms that reflect it. That is a legitimate outcome and considerably better than either resentment or removal.
Where Blomma fits
The reason this sits unaddressed is not that founders lack nerve. It is that there is no venue, no vocabulary, and nobody to test the read against — your team cannot be asked, your investors would draw conclusions, and mutual friends will take a side.
Blomma is an always-on AI career coach with no stake in the outcome. Use it to test your measurement first, including the honest question of what work you might not be counting. Use it to work through which of the five causes fits the evidence rather than the one your frustration suggests — particularly the possibility that this is unwell rather than unwilling. Use it to build the mechanism, so raising things stops requiring an occasion. And use it to rehearse the opening, because in this conversation the difference between an observation and an accusation is the difference between information and a fight. Where it warrants a human who has been in a founding pair, bring one in..
Unequal load between cofounders is ordinary and it does not stay ordinary. It compounds, quietly, and it is the most common thing founders describe when they explain why a founding relationship ended. Raising it in month three is an awkward hour. Raising it in year three is a different conversation entirely.