A cofounder trial project should help both people observe how they make decisions, keep commitments and respond to customer evidence. Two weeks can reveal useful working habits, but it cannot certify a long-term partnership or replace proper legal arrangements.
The best trial is a small, real question with a safe scope. “Can we learn why these users abandon this step?” is more useful than “build my complete MVP.” Both people should have meaningful responsibilities, and either should be free to conclude that the partnership is not right.
Agree on the rules before opening a repository
Write a short trial brief together. Include the question, the expected time commitment, the planned output and the date when you will decide what happens next. Explain whether expenses or compensation apply. A trial label does not automatically resolve employment, contracting or intellectual property questions.
Discuss existing employer obligations, confidential information and ownership of pre-existing work. Ask what each person may reuse if the trial ends. Obtain advice from a qualified lawyer when these arrangements matter. This article offers a collaboration process, not an enforceable agreement or legal advice.
Set appropriate access boundaries. Use synthetic or approved test data where possible. Do not exchange customer credentials, production access or private employment materials simply to prove trust. Responsible limits are part of working well together.
A proposed two-week sequence
This example spans ten working days across two weeks, not fourteen daily assignments. It is not a universal assessment method. Adjust the duration and workload to both people's circumstances.
Days one and two: define the question. Each person writes what they believe the customer needs and what would change their mind. Compare those accounts. Agree on the smallest test that can distinguish your assumptions from something observed.
Days three to five: do complementary work. One person might recruit participants and map the current workflow. The other might construct a simple demonstration or test an implementation risk. Both should understand why the tasks matter, even if they do not share the same skills.
Days six to eight: review evidence together. Attend selected feedback conversations together, with participant consent. Record observations separately before proposing changes. Notice whether disagreement turns into a useful question or a contest over who was right.
Days nine and ten: hand over and decide. Produce a concise account of the result, remaining uncertainty and next options. Review access and expenses. Then have an explicit conversation about continuing, extending the experiment or ending it.
YC's matching launch article treats introductions as a beginning rather than a partnership decision. Keep the same distinction here: finishing a small project is a reason to review the relationship, not an obligation to cofound.
Observe behaviour you can describe
A useful observation is “we disagreed about scope and narrowed it using the customer evidence.” A less useful observation is “she has founder energy.” Write down actions, not personality verdicts.
Look for how each person communicates a delay, asks for help and explains an unfamiliar topic. Notice whether invisible work gets recognised. Someone coordinating participants or resolving a permissions question is contributing, even if there is no impressive demo at the end.
Watch your own behaviour too. Did you change the assignment without discussing it? Did you treat a question as disloyalty? A trial should allow mutual evaluation rather than quietly making one person the examiner and the other the applicant.
A hypothetical trial that ends well without a company
Suppose two people in Gurugram are considering software for warehouse handovers. They agree to interview operators and demonstrate one handover screen using invented records. One recruits the participants; the other builds the demonstration. Both commit to clearly stated evening hours.
The interviews show that users care more about accountability than speed. One founder wants to explore the slow integration work that follows. The other enjoys prototyping but does not want long-term operational responsibility.
They decide not to cofound. They settle agreed expenses, remove shared access and retain only the material each is entitled to keep. The trial succeeded because it revealed a meaningful difference without trapping either person in an assumed commitment.
Use a handover checklist even when you continue
Your closing note should answer five questions:
- What did we actually learn, and which claims remain assumptions?
- Where are the permitted outputs and working notes?
- Who has access to which accounts, data and expenses?
- What can each person use after this experiment, under the agreed terms?
- What decision have we made, and when will we revisit it?
YC treats questions as a supplement to working experience.
Continuing the experiment does not settle the partnership. Schedule a separate conversation about authority, commitment, finances, ownership and departure.
There are limits. A calm, short trial will not reveal every response to commercial pressure, illness or prolonged uncertainty. Someone may also have caring or employment commitments that affect this particular fortnight. Make room for context without ignoring repeated broken agreements.
The purpose is not to manufacture pressure. It is to create a fair opportunity to see how a partnership might work, then make a deliberate next decision with evidence and appropriate professional advice.