Dividing Roles and Decision Rights Between Partners

Partners should write down who runs what and who decides what as two separate lists, before either reaches the agreement's authority and voting clause. In Carta's analysis of 40,228 US startups, 44% of two-founder teams split their equity exactly 50/50 in 2025. Those partners have no majority between them on any question.

The default rule offers no rescue. Under the Revised Uniform Partnership Act, the model most states adopted, every partner has equal management rights unless an agreement says otherwise. That holds for the partner working full time and for the one who visits on Fridays. Silence on authority settles nothing; it moves the argument to a later and more expensive date. The issue arises whenever a co-founder, a joint venture partner or a new owner is being documented, and it decides who can spend, who can hire and who can say no.

Even splits are common and, some lawyers argue, overused. Carta's two-founder median was 51/49 in 2025, and its equal-split share rose from 31.5% in 2015 to 45.9% in 2024. Attorneys at Dinsmore & Shohl counter that "splitting the equity equally among [founders] is rarely the right answer" and want shares to follow anticipated contribution. Carta attributes the fall to 5% equal splits among five-founder teams to role differentiation. These are startups on Carta's platform; the sources carry no comparable figure for small partnerships. An unequal split needs a reason that can be written down, and credit for unpaid work (sweat equity) is one.

The table maps that hypothetical to roles and authority. A has a finance and operations background; B has a sales record. The $25,000 debt line is Causo's example and the reserved items come from Causo and Gabriels. The $10,000 limit and the allocation of owners are illustrations.

Early-stage founders. Causo's template gives the CEO final say on hiring and firing and on anything unreserved, gives the other founder a named domain such as technical architecture, and requires joint consent for the budget, new equity, debt above the sample $25,000, executive hires, a sale and dissolution. It warns against co-CEOs because investors read them badly. Owners with no investors can ignore the reasoning and keep the structure.

Joint ventures. The PTI/CFFT and EPIX/CFFT steering committee clauses give each side one vote, require unanimity, and allow 20 days to resolve a deadlock before it goes to formal dispute resolution. PTI/CFFT carves out one category: disputes over amending the research plan fall to PTI's sole discretion. Unanimity, it turns out, can be rationed by subject.

Service businesses. Gabriels splits a restaurant between a Culinary Director and an Operations Manager, and an accounting firm between tax and audit. A SearchFunder post from a searcher buying a building services company weighs the usual sales and operations split against dividing accounts by size or geography. The stated worry is role clarity for employees when both partners touch the same accounts. The replies are not public, so the outcome is unknown.

A role is not a decision right. The role says who does the work; the decision right says who has the final word when partners disagree.

Ownership is not authority. Under the default rule a 60% partner has the same management right as a 40% partner. Dinsmore & Shohl argue that those contributing most should drive decisions, but that result must be written in; it does not follow from the percentage.

A decision-rights clause is not a deadlock clause. The first limits how many questions can reach an impasse. The second, covered in what happens after an impasse, resolves the ones that still do. Both are needed, since a UK practitioner note from LexisNexis observes that deadlock can arise outside 50/50 ownership wherever reserved matters require unanimity. English law does not govern a US partnership, but the arithmetic travels.

The Finberg Firm wants day-to-day responsibilities, spending limits and reporting duties spelled out for each partner, and LegalNature wants each partner to accept the role formally and early. Visotsky adds an org chart built from the functional map, so that staff can see who owns what. The table above, attached as a schedule, does most of this work, and the authority clause then only has to refer to it. The clause sits among the other terms in what a partnership agreement covers, and the agreement generator produces a full draft for a lawyer to review. This page has no sourced figure for what a lawyer charges to draft it, so none is given.

Visotsky suggests reviewing the split quarterly or semi-annually, as the business grows or new projects arrive, and Gabriels warns against partners retreating into silos even when their lanes are clean. A standing check-in, with sub-tasks clarified in writing as they arise, keeps a schedule from becoming a historical document. Tasks neither partner wants should be listed too. They can be rotated on a fixed calendar or outsourced at the business's cost, and either is better than leaving them to whoever resents them less.

The strongest objection is that all this is over-engineering for two people who trust each other and talk daily. There is force in it. Most decisions in a small business never come to a vote, and Visotsky rightly notes that functional lines sometimes have to blur for the good of the business. A partner covering for another during a crunch should not need a resolution. But crossover concerns who helps; the clause concerns who decides, and it operates only on the days when trust and daily conversation have failed to produce an answer. Partners who never use it have lost an afternoon.

Exits need the same treatment. A departure changes both the ownership table and the role table, so the agreement should say who inherits the departing partner's functions and what happens to any sweat credit not yet earned. The Finberg Firm recommends a buy-sell agreement with a defined valuation method for death, disability or departure, and filed deadlock samples commonly escalate from senior management to mediation to a forced buyout. The partner buyout calculator prices that last step.

Role division is one item in wider partnership planning, and it is the one most easily deferred. The list below can be copied and filled in with each partner's name and figures, and then becomes Schedule A.

The test of the finished draft is cheap: take the most expensive disagreement the partners can imagine and ask the document who decides it. If the answer takes more than one sentence, the clause is not finished.