What to do when co-founders disagree about company direction

Address co-founder disagreement about company direction through a shared assumptions review and a bounded test where feasible. Separate competing views of the business from questions about loyalty. Confirm who has authority to decide and use qualified advisers when formal rights or obligations are disputed.

Task conflict is disagreement about the content or goals of shared work, which can concern evidence or priorities without necessarily involving personal dislike between the people involved. This definition follows A Multimethod Examination of the Benefits and Detriments of Intragroup Conflict (Karen A. Jehn (1995), defining study). Relationship conflict concerns interpersonal tension rather than the work’s content. The assumptions review separates a disagreement about the work from personal hostility, without claiming that every task disagreement improves the decision.

A private moment that makes the issue visible

Composite scenario, not a client case: You discuss the company’s next move with a friend and describe your co-founder as losing conviction. Back at work, the actual disagreement concerns a different assumption about customer demand. The relationship has become the explanation before the assumptions have been compared.

State the decision before discussing personalities

The decision concerns direction under uncertainty. Task conflict provides a distinction for disagreement about the work; it should not be used to relabel personal attacks or dismiss a real relationship problem.

Proposed method: co-founder shared-assumptions review

State the common decision. Agree on the choice and relevant timing. Identify what each person believes the company must protect. Check that both people are answering the same question over the same period. Separate a near-term operating choice from a wider objective. Confirm the authority needed before the company receives competing instructions.
Record competing assumptions. Describe the evidence and reasoning behind each recommendation. Keep interpretations separate from confirmed facts. Give each recommendation the same fields and evidentiary standard. Ask the other founder to restate it accurately. Keep historical status or effort from replacing the evidence relevant to the current choice.
Find the decision-changing difference. Identify which uncertain condition would lead to different choices. Avoid debating every historical disagreement. Identify an observation that could alter either recommendation. If no evidence would change the positions, examine whether the disagreement concerns incompatible goals rather than an assumption that a test can resolve.
Design a bounded test. Choose appropriate evidence and limits with responsible owners. State what each result would change before beginning. Define the result’s meaning and the commitment limit before work begins. Allow an ambiguous outcome. Obtain appropriate approval for real customer promises or resource use instead of calling consequential action a harmless experiment.
Use the authorized decision route. Confirm the choice, owner, and review condition. Obtain legal or other specialist advice for actual governance or contractual questions. Record the formal choice and one instruction for execution. Preserve remaining disagreement without making compliance a declaration of personal loyalty. Use qualified advice when the issue concerns actual ownership or governance rights.

Ask each founder to complete the same proposal record

Use a short written record before a joint debate. Each founder states the recommended direction, the evidence supporting it, the condition most likely to weaken it, and the work displaced. Exchange the records and ask each person to restate the other’s recommendation accurately. Correct the account before arguing about it. This is a proposed preparation method, not a guarantee that a relationship conflict will disappear when the strategy is written down.

Check whether the founders are answering the same question. One may be choosing a near-term operating response while the other is discussing the company’s long-term purpose. Name that difference. A trial of customer demand cannot resolve incompatible ownership objectives by itself. Confirm which decision is now required and where a formal authority question must go. Keep the company from receiving contradictory instructions while the founders continue a broader discussion.

Locate the assumption on which the recommendations divide

Place the material assumptions side by side. Ask which fact, estimate, or objective differs and which observation could alter a recommendation. Avoid a large catalogue of every disagreement. The useful difference is one that changes the action under consideration. If both founders would choose the same action after a plausible finding, identify the evidence route. If no evidence could change either view, examine whether the dispute concerns objectives rather than an empirical question.

Jehn’s 1995 study distinguishes task conflict from relationship conflict and reports that the consequences of conflict depend on the setting and type. It does not establish that all strategic disagreement is beneficial. Use the distinction to describe what the discussion concerns. Do not relabel personal attacks as productive task debate or dismiss a relationship issue because the participants can also produce competing business cases.

Give the test an interpretation before it starts

For a hypothetical market-entry disagreement, identify the customer requirement that separates the two recommendations. Choose an appropriate way to investigate that requirement, with authorized resources and information handling. State which result would support a narrower step, which would stop it, and what would remain unresolved. Do not use a promotional launch as a supposedly neutral test if the company has already committed to proceed regardless of the response.

Ask someone appropriate to examine the proposed interpretation. A test may supply evidence relevant to one assumption while leaving timing, capacity, or economics unanswered. Preserve those limits. If the result is ambiguous, choose the next decision the available evidence can support. Neither founder needs to announce victory for the work to be useful. Record the change to the recommendation rather than reward the person whose original position sounds closest to the eventual answer.

Keep contribution and authority out of the evidence shortcut

Ross and Sicoly’s research concerns how people recall and attribute their contributions to shared work. It does not determine which founder contributed more or who has authority now. Where historical effort enters the argument, check the relevant records and separate that conversation from the current strategic assumption. Obtain qualified advice for ownership or contractual questions. A founder’s earlier success can inform the discussion without automatically deciding the present commitment.

Edmondson’s work-team study provides a limited reference for making questions and corrections possible. It is not evidence that this co-founder process has been validated or that a comfortable discussion resolves formal rights. Agree on conduct that permits a substantive challenge while preserving the proper governance route. A bounded business decision can close even while personal views differ, provided the authorized instruction and execution responsibilities are clear.

At the review, compare the evidence with the written interpretation from before the test. Record what changed and which issue remains outside the result. If a recurring relationship problem keeps reopening strategy, address that problem through appropriate support rather than commission another trial that cannot answer it. Give the operating team one authorized direction and a condition for reopening, so the business does not carry an indefinite private contest into everyday work.

Put the boundary into words

We disagree about this assumption and the direction it supports. Let us compare the evidence and define what would change either recommendation. We will resolve the business choice through the agreed authority rather than use agreement as a test of loyalty.

Work through a bounded example

Hypothetical application: Co-founders differ over entering a market. The review identifies an unresolved customer requirement and chooses an authorized discovery test. They preserve the possibility that the evidence favors either direction rather than treating the test as a way to confirm one founder’s plan.

Establish what the disagreement is about

Ask each founder to state the recommendation in a form the other can recognize. Include the intended outcome, the main constraint, and the assumption carrying the most weight. Correct inaccurate restatements before debating the recommendation. This helps distinguish a different view of the evidence from a different objective that no single factual check will settle.

Identify any shared condition that both recommendations need. The company may require dependable delivery, continued access to a resource, or compliance with an existing obligation. Verify the actual condition and its owner. A shared commitment can provide a useful boundary for the disagreement without forcing premature agreement on the whole direction.

Keep historical contribution separate from current evidence. A founder’s earlier success can supply experience worth examining, but it does not automatically decide the present choice. The same applies to a past mistake. Ask what the earlier case establishes under today’s conditions. Avoid using the conversation to settle every unresolved question about status, effort, or recognition.

Choose a test that can challenge either recommendation

State the discriminating question before choosing an activity. A test is useful when plausible results could alter a consequential assumption. A promotional launch designed to produce supportive anecdotes is not an independent examination of demand. Use appropriate customer evidence, technical review, or operating work according to the question, and protect participants and confidential information.

Agree on the commitment required for the test. Define capacity, access, and the limit beyond which another approval is needed. Do not describe a consequential market entry or customer promise as a small experiment merely because its budget is limited. Qualified owners should assess reversibility and actual obligations. Preserve a way to stop or revise the work when the relevant condition fails.

Write what each result would mean. The evidence may support a narrower direction, show that both original recommendations miss a constraint, or remain inconclusive. Do not force an ambiguous result into a victory for one founder. Identify the next decision the available evidence can support and whether obtaining more information is worth the delay.

When objectives remain different, acknowledge that the disagreement is no longer only an assumptions question. Use the agreed governance route and appropriate facilitation or legal advice. A business decision can close while personal views differ. Confirm what execution requires and what new evidence would reopen the choice, so the company does not inherit an endless private contest after the formal discussion ends.

Check the failure mode before scaling

Do not claim all task conflict is beneficial or expect one test to resolve incompatible objectives. Avoid informal changes to ownership or governance. Use qualified advisers and facilitation where appropriate.

Review the work, then revise the arrangement

Review the evidence against the original assumptions and record the decision. Address a continuing relationship issue directly rather than keep reopening strategy as its substitute.