Before choosing a cofounder, discuss the company you want, the work each person will own and what happens when your interests diverge. Useful questions produce concrete examples and follow-up actions, rather than a reassuring list of shared values.
You are not trying to predict every future disagreement. You are checking whether you can discuss an uncomfortable topic without hiding the important part. A partnership can survive different preferences more easily when both people know what those preferences are.
Start with the life each person is signing up for
Ask: “What would make building this company worthwhile for you?” Let each person answer before debating the response. One may want a small, durable business; another may want a venture-funded company that grows quickly. Those are different operating choices, not differences in personal worth.
Make the question tangible. Would each person be happy if the company became profitable but stayed small? What external commitments shape availability? Under what circumstances would someone need a salary, reduce hours or stop? Respect privacy while requesting enough practical information to plan responsibly.
Do not demand that another person justify family responsibilities or expose their full personal finances. Discuss business-relevant constraints and what each person can realistically commit to. Consent matters in diligence too.
Ask for a past example, then a future response
For each important topic, pair a hypothetical question with evidence from experience. “How do you handle disagreement?” invites an ideal answer. “Tell me about a decision you changed after someone challenged you” can reveal how the answer works in practice.
Useful prompts include:
- When have you missed a commitment, and how did you communicate it?
- What kind of work do you reliably avoid, even when it matters?
- How would we handle a customer asking for something one of us considers unsafe?
- What decision would you want authority over, and where would you want a second opinion?
- What would you do if we disagreed about raising investment?
You are allowed to say that you do not know. Record the uncertainty and design a way to learn, rather than inventing a confident answer because the conversation feels important.
Separate discussion from permission
Agreement that someone should own engineering does not settle every authority question. Can that person stop a release? Spend money? Commit to a delivery date? Hire someone? Change the underlying product direction?
Talk through actual decisions the business is likely to face. Write down which person proposes, which decides and when consultation is required. Revisit the arrangement as the company changes. A practical decision boundary is more useful than repeatedly saying that both founders will “do everything.”
Ownership and governance need appropriate documentation and legal advice. This conversation is preparation for that work, not a substitute for it. YC discusses equity separately from merely originating an idea.
Use separate notes before a joint conversation
An independent written response helps prevent one person from simply adapting to the other's enthusiasm. Each can write a short answer about ambition, availability, responsibilities, money and departure. Exchange the notes, then mark disagreements without immediately trying to solve them.
YC recommends independent answers and experience beyond discussion.
Choose one unresolved issue for a follow-up. If your views on customer support differ, share a support exercise. If availability is uncertain, try working together at the proposed hours. Do not engineer a fake emergency to see how someone reacts.
A hypothetical disagreement worth finding early
Imagine two founders exploring a procurement tool. Both say they want to build an important business. During the discussion, one explains that she wants to keep a consultancy running. The other assumed both would leave other work within a month.
Instead of treating the mismatch as dishonesty, they compare the actual requirements of the proposed customer pilot. They discover that coverage is needed during hours neither can currently guarantee. They postpone the commitment and narrow the experiment.
A different pair might solve the same issue through a clear staged plan. The useful outcome is not that every disagreement ends the partnership. It is that nobody has to discover the real arrangement after a customer is depending on it.
Close with decisions and unknowns
Keep a shared record containing three sections: agreements, unresolved issues and evidence still needed. Include an explicit conversation about ending the collaboration: how to communicate, settle commitments and handle access or work product appropriately.
The process has limits. A thoughtful person can give a poor answer when nervous; a persuasive person can describe values they do not practise. References, obtained with consent, and repeated collaboration provide different information. Use all of them rather than turning one conversation into a personality judgment.
Before your next meeting, choose five questions tied to the company's next six months. Leave enough time to hear the answers, especially the inconvenient ones. A partnership becomes more credible when both people can remain honest without having to perform certainty.