Cofounder Roles and Responsibilities: How to Divide Them
Most founding teams divide roles once, informally, in the first few months, and never revisit it. That split was correct for the company that existed then and is quietly wrong for the company that exists now, which is why so many cofounder arguments are really arguments about scope that nobody has updated. This page is about dividing roles in a way that survives growth: what to actually split, why the obvious methods break down, how to handle titles honestly, and how to redivide on a schedule so drift never accumulates into a dispute.
What’s inside
Divide accountability, not activity
Three splits founders try, and how each fails
The method
Titles, and what they actually cost
The annual redivision
What should stay shared
Where Blomma fits
Divide accountability, not activity
The instinct is to divide the work: you do the product, I do the customers. The problem is that work changes constantly, so a split defined by activity is out of date within two quarters and has to be renegotiated informally every time something new appears — which is where the friction comes from.
Divide outcomes instead. Not “you own engineering” but “you own whether we ship a product customers keep using, measured by retention.” Not “I own sales” but “I own new revenue, measured by qualified pipeline converting.” The activity underneath can change completely and the ownership still holds.
This has two immediate benefits. First, it survives scale — the outcome is the same at forty people as at four hundred, even though the daily work is unrecognisable. Second, it makes accountability possible without micromanagement, because there is something concrete to be accountable for that is not a list of tasks.
And it forces a useful conversation early: if you cannot name the four to six outcomes this company needs owned, you do not yet have a shared model of the business, which is worth discovering now rather than in an argument.
Three splits founders try, and how each fails
The functional split.You take product and engineering; I take go-to-market. Intuitive, and it fails at the boundaries — which is where the important decisions live. Pricing is a product decision and a revenue decision. Developer documentation is engineering and marketing. A functional split resolves the easy cases and leaves every genuine question ambiguous, which means the hardest calls are also the least owned.
The inside/outside split.One founder runs the company, one runs fundraising, customers, and partnerships. Works well early and creates a specific failure at scale: the outside founder loses context and starts making commitments the inside founder cannot deliver, while the inside founder starts treating the outside founder as a visitor. The relationship erodes through asymmetric information rather than disagreement.
The strong/weak split.Roles assigned by who is better at what, which drifts into one founder owning everything difficult. Efficient in year one, and it produces the load imbalance that becomes the most corrosive form of cofounder resentment — usually surfacing years later as an equity conversation.
Each of these can work for a while. None of them survives without deliberate revision, and the common failure is treating the initial split as settled rather than as a first draft.
The method
A session of two or three hours, done properly once and revisited annually.
Start with the outcomes, not the people.List the four to six outcomes the company must produce over the next eighteen months. Something like: a product customers retain; a repeatable way to acquire customers; enough capital and runway; a team that can execute; and whatever is genuinely specific to your business. Write them without reference to who does what.
Assign one owner per outcome.One name, not two. Shared ownership of an outcome means nobody owns it, and the specific failure mode is that both founders assume the other is worried about it.
Attach a metric and a cadence to each.How you will both know whether the outcome is on track, and when you will look. This is what converts ownership from a title into something real, and it is what makes an accountability conversation possible without it feeling like an accusation.
Then map the decisions.For each outcome, list the significant decisions inside it and confirm the owner actually holds them. An owner without decision rights is a figurehead, and this is where most role splits quietly fail — the scope is granted and the authority is not.
Name what is explicitly shared, and how shared things resolve.A small list, with a rule for deadlock. Everything not on that list belongs to one person.
Write it on one page and send it to your leadership team.A split your executives have not seen does not change behaviour, because people continue routing decisions to whichever founder they expect to get their preferred answer from.
Titles, and what they actually cost
Titles are the part founders handle worst, usually by avoiding them.
The common approach is to leave them vague or make them symmetrical — both co-CEOs, or both simply “founder.” This feels equitable and it exports a cost. Your team cannot tell who decides. Candidates cannot tell who they would work for. Investors will assume one of you is really in charge and work out which, without asking you. And the ambiguity does not remove the underlying power dynamic; it just makes it undiscussable.
A few honest positions:
If one of you is going to have the final call, give that person the CEO title and say plainly what it carries. The ambiguity of a CEO who does not actually decide is worse for both founders than either clear alternative.
If you genuinely intend to split authority, define the split precisely enough that the company can operate — which in practice means naming which decisions each of you holds and what happens on the rest. Co-CEO arrangements work when this is unusually explicit and fail when it is aspirational.
And separate title from status. A cofounder who is CTO rather than CEO has not been demoted; they have a defined job. The reason this needs saying is that founders often resist role clarity because it feels like ranking, and the resulting vagueness costs the company more than the ranking would have.
The annual redivision
The single highest-value habit in this whole subject, and almost nobody does it.
Once a year, sit down and rebuild the split from the outcomes. Not a review of whether the current arrangement is working — a fresh pass. What does the company need owned over the next eighteen months, and who should own each of those things given where each of you actually is now?
Doing this on a schedule changes the conversation completely. A role change proposed during an argument is a judgment. The same change proposed in an annual redivision is planning. Founders who redivide deliberately every year rarely reach a crisis, because they never accumulate two years of drift — and drift, not disagreement, is what most cofounder breakdowns are made of.
Be explicit about what changed and why. Scope moving away from one founder is normal at every stage and reads as demotion if unexplained. And expect some years where the honest answer is that one of you should own less than before, which is exactly the conversation the annual cadence exists to make possible.
What should stay shared
Not everything should be divided, and over-dividing produces two people running adjacent companies.
Keep the standard shared. What quality of work is acceptable, what behaviour is not, who you are willing to hire — that has to be common, or your company develops two cultures along the reporting lines.
Keep strategy shared, with an owner. One of you drives the process; both of you must genuinely be in the direction. A strategy owned by one founder that the other has never committed to is the source of the tactical arguments that never resolve.
Keep the relationship shared. A standing weekly founder conversation that is explicitly not a status meeting. This is the maintenance that prevents the information asymmetry that kills inside/outside splits.
And keep decisions about each other shared and explicit — who decides if one of you should change roles, and how that gets raised. Leaving this undefined is why cofounder performance problems go unaddressed for years: there is no mechanism, so nobody starts.
Where Blomma fits
The hard part of dividing roles is not the framework. It is that an honest division surfaces things founders would rather not say — that a domain has outgrown its owner, that the CEO title has never meant anything in practice, that one of you has been carrying more for two years. Those are the items that make founding teams quietly not schedule the session.
Blomma is an always-on AI career coach with no stake in your company. Use it to draft the outcome list independently before you compare, because the differences between your two lists are more informative than the overlap. Use it to work out honestly which outcomes you should own given where you actually are now, rather than which ones you have always owned. And use it to prepare the two or three moments in the session you expect to be difficult — the title conversation, or the scope that needs to move — rather than deferring them for another year. Where it warrants a human who has been in a founding pair, bring one in..
Role division is not a one-time act of diplomacy. It is a planning exercise you repeat, and repeating it on a schedule is what keeps drift from turning into a dispute.