How to run a cofounder trial project: a 2 to 6 week plan to test working together

A cofounder trial project is a short piece of real work that two people do together, for a fixed period, to find out whether they should start a company together. The work matters, but what you are really testing is what you learn about each other along the way: how you decide, how you disagree, how fast each of you moves, and whether you both do what you said you'd do.

Conversations can only take you so far. In her account of finding a cofounder, Gloria Lin wrote that "once you start doing some kind of project, you get so much more data on the person" (First Round Review). Steve Blank's advice, in another First Round piece, is to "date first": an intense stretch of work, then "a provisional 30-day working period" before anyone fully commits (First Round Review).

This guide is the practical version. It covers when to start, how long to run it, how to choose the project, how to set scope and roles, what to write down, what to watch for, and how to end it, with a sample plan and a template you can copy.

The trial at a glance

When should you start a trial project?

Start once you've had two or three conversations that went well and covered the hard topics: hours, money, where each of you will be based, what kind of company you both want. If you haven't asked those questions yet, our list of questions to ask a potential cofounder is the place to begin. A trial won't fix a mismatch you could have found in an hour of talking.

Start before anyone quits a job, spends real money or announces anything. The trial's job is to make that decision easier, so it has to come first.

And start with one person at a time, if you can. Running two trials at once halves the hours you give each and makes the comparison unfair to both. Keep other conversations warm, and tell them where you are.

How long should a cofounder trial last?

Published advice ranges from two weeks to several months:

• Hustle Fund's Founder Playbook says a trial project should be finishable in 1 to 4 weeks at most, and calls "two to three weeks with clear milestones" the sweet spot. A single week may be too rushed; more than a month starts to feel like unpaid consulting (Hustle Fund).
• Steve Blank suggests at least a provisional 30-day working period after an intense first burst of work.
• The Whalesync founders ran "a 6-week trial period with weekly check-ins", and agreed the goals, length, check-ins, projects and how they'd communicate before starting (Whalesync).
• Annie Garofalo, writing in The Co-Founder Confidante, describes trials of 1 to 6 months and advises against anything shorter than a month, which she puts at roughly 120 hours of working together (The Co-Founder Confidante).

We suggest 2 to 6 weeks, set by hours rather than by the calendar. What you're measuring is time spent working together, so two people who can each give 10 hours a week need longer than two people working on it full-time.

Whatever you choose, write the end date down at the start. In Lin's experience, what she "most appreciated" in a partnership that wasn't working was "a clear go or no-go date". An open-ended trial tends to drift into a partnership neither of you decided on.

Step 1: Pick the right project

A good trial project passes five tests:

1. It's real. It moves an actual idea forward. Busywork tells you how someone handles busywork. 2. It's finishable. You can reach a clear result within the dates, at the hours you each have. 3. It needs both of you. Each person has a part only they can do well, and at least one piece of work depends on the other. 4. It touches the outside world. Customers, users or a market reply to something you made. Their reaction gives you something real to disagree about. 5. It ends in a decision. The result tells you something about the idea, such as whether to keep going, change direction or drop it.

Cofounder trial project ideas

These fit most early ideas. Pick one, not three.

• Customer interviews and a summary. Book 10 to 15 conversations with people who'd use the product, split them between you, and write one shared page on what you heard. Good for any pair, and especially for testing a business cofounder's ability to get meetings.
• A landing page and a small test. Build a page for the idea, send traffic to it, and compare two ways of describing it. Hustle Fund describes a candidate who did exactly this and came back with data on which message worked better.
• A clickable prototype tested with five users. One person builds it, the other recruits and runs the sessions, and you both watch.
• A manual version for three customers. Deliver the service by hand to a small number of real customers before building anything. You'll learn about the customers and about how each of you handles a promise made to a stranger.
• A pricing test. Put a price in front of potential buyers and see who says yes, maybe or no.
• A technical spike plus a sales list. One of you proves the hardest technical part works; the other builds a list of 50 target customers and contacts the first ten.

Rob Fitzpatrick's short book The Mom Test is a useful shared standard for any trial that involves customer conversations. Reading it together takes an evening.

If you don't have an idea yet

Hustle Fund's second playbook suggests that people starting fresh can test the partnership on something fun or a problem that bugs you both (Hustle Fund). That works for seeing how you split tasks and handle roadblocks. It shows less about how you behave when something you both care about is at stake, so if you have even a rough idea, use it.

Matching the project to your pair

• Technical and business cofounder. Choose work with a clear handoff: the business side brings customer input, the technical side builds something, and the customers react.
• Two technical cofounders. Make sure the project includes talking to users, or you'll both learn a lot about the code and nothing about each other's judgement.
• Two business cofounders. Include something you have to make, even if it's a prototype made with a no-code tool. Otherwise the trial becomes a series of meetings.

Step 2: Set the scope and a finish line

Write the deliverable in one sentence that both of you would describe the same way. "Interview customers" is a direction. "Twelve interviews with clinic managers, notes in one shared folder, and a one-page summary of the top three problems by the 28th" is a finish line.

Then write what's out of scope. Trials grow. A customer suggests a feature, someone has a better idea for the landing page, and three weeks later you're building a product. Hustle Fund calls this "the endless project", and lists it among the traps to avoid. A short "not in this trial" list makes it easy to say "good idea, after the review".

Keep the project smaller than you think. If it turns out too small, finishing early tells you something good about both of you.

Step 3: Split the roles

Give each person one part they lead, with their name on it, and one piece of work you do together. Leading means making the day-to-day calls on that part and telling the other person what you decided.

Two things to avoid:

• One person as the manager. If one of you hands out tasks and the other completes them, you're testing a job, not a partnership. Both of you should make decisions the other has to live with.
• Splitting so cleanly you never meet. If the two halves never touch, you'll each finish your part and learn nothing about working together. Build in at least one moment where one person's work depends on the other's.

Leave one dull, necessary task unassigned on purpose, such as booking the interviews or writing up the notes, and see who picks it up. Our guide to cofounder green flags explains what that tells you.

Step 4: Write a one-page trial brief

The brief is the most useful hour you'll spend on the trial. It turns assumptions into something you can both point to later. Garofalo's agreement has three parts (names and intentions, timelines and accountability, project details), and the Whalesync founders defined goals, length, check-ins, projects and communication before they began. Ours combines both into one page.

Trial brief template

• The question. What do we want to learn about working together? For example: "Can we make decisions quickly when we disagree, and do we both follow through on customer work?"
• The deliverable. One sentence, with a number and a date.
• Not in this trial. Three to five things you agree to leave out.
• Dates. Start date, end date, and the date of the decision meeting.
• Hours. What each of you can realistically give each week, and when (evenings, weekends, two full days).
• Roles. What each person leads, and the piece of work you share.
• Check-ins. Day, time and length of the weekly call.
• Communication. Where you talk day to day, and how either of you flags that something will slip.
• Money and accounts. Any spending allowed (domains, ads, tools), who pays, and whose accounts things are set up in.
• If we stop. Who keeps what each person built, written down before any work starts. If anything here is unclear, especially where one person brings an existing idea or code, take advice from a lawyer.
• Pausing. Either of you can ask to pause or stop early, and how you'll tell the other.

If either of you is employed, add one line on when and where you'll work on the trial, and on whose laptop. Many employment contracts say who owns what an employee invents; our guide to cofounder red flags explains why that matters.

Both of you read the brief, change it, and agree to it in writing. An email reply saying "agreed" is enough. Think of it as a shared record of what you each said you'd do.

Step 5: Set the schedule

A simple weekly rhythm works for most trials:

Hustle Fund suggests regular check-ins every few days and a final review. Garofalo adds check-ins to share appreciation, feedback and experiences, not only progress. Use the weekly call for both: ten minutes on the work, and five on how working together is going.

Weekly check-in questions

1. What did each of us finish, and what slipped? 2. Did anything change since last week, and did we both hear about it in time? 3. What did we decide this week, and how? 4. What's one thing the other person did that helped? 5. What's one thing we'd like to change about how we work next week?

What to watch during the trial

Watch how the work happens, not only what gets delivered. These are the moments that tell you the most.

Decisions. Notice how long small choices take and who makes them. Some people move quickly on small things and slow down for large ones; others agonise over the colour of a button and rush the pricing. Hustle Fund's first trait to watch is speed: do they move forward or stall in analysis?

Disagreements. The first real one usually comes in week one or two, often over which customers to call or what to cut. Watch whether reasons change minds, whether either person goes quiet, and whether the decision sticks the next day.

Pace. Compare what each of you said you'd finish with what got finished. A mismatch in pace is one of the most common reasons partnerships strain, and it's easy to miss on a call.

Follow-through. Does each of you do what you said, by when you said it? When something slips, does the person tell the other early, or does it come out at the check-in?

Customer contact. Does each of you actually talk to customers, and do the notes contain the customer's own words or only conclusions?

Pressure. Garofalo suggests testing how your stress responses interact, for instance by running a high-pressure event together. A demo, a deadline or a difficult customer call does the same.

Feedback. Give one piece of honest feedback on the other person's work in week two, kindly and specifically. Then watch what they do with it.

Don't manufacture conflict or set traps to see how someone reacts. Real work produces enough pressure on its own, and a staged test damages trust you'll need if the trial goes well.

For specific signs to look for, see our guides to cofounder red flags and cofounder green flags. They're written as checks you can run during a trial.

Keep a simple decision log

A decision log is a shared document with a few lines for each real decision you make during the trial. It takes two minutes per entry and does two jobs. It stops you arguing later about what was agreed. And at the review, it gives you evidence instead of impressions.

Log the decisions that mattered, not every message. Five to ten entries over a trial is normal.

For each entry, write:

• Date.
• The decision. One sentence.
• Options we considered. Two or three.
• How we decided. Who proposed it, whether we disagreed, what settled it.
• What we expect to happen.
• What actually happened. Filled in later.

Example entry (made up for illustration): 12 March. Decided to interview clinic managers, not doctors. Options: doctors, clinic managers, both. Priya wanted doctors; Tom argued managers hold the budget. Settled by calling two managers that afternoon, both of whom said they choose the software. Expect: easier bookings. Actual (19 March): 9 of 12 interviews booked in a week.

Keep a second, private note for yourself, separate from the shared log. Write down what you saw, with the date, and keep it apart from what you think it means. "Missed the Thursday deadline, told me on Friday" is an observation. "Unreliable" is a conclusion, and one data point rarely supports it. A month later, memory turns single events into habits and habits into "just that once", so the written note keeps you fair to both of you.

How to end the trial: yes, no or extend

End on the date you set, even if the deliverable isn't finished. An unfinished project is information too.

1. Write separately first. Each of you writes half a page before the meeting: what went well, what didn't, what you learned about the idea, and your answer: yes, no or extend. Writing first stops the more confident person from setting the conclusion. 2. Read both, then talk. Use the decision log to anchor the conversation in what happened. 3. Decide one of three things.

Yes. You both want to keep going. Say it plainly, then agree on the next concrete step and when each of you will start working the hours you discussed. Everything that formalises a partnership comes after this conversation, not in it.

No. One or both of you don't want to continue. Say it kindly and directly. Hustle Fund's advice is to thank the person for the work and explain that the working styles don't fit, without a long justification. For example:

"Thank you for the last four weeks. I learned a lot, and the interviews were better than I'd have done alone. I don't think we work well enough together to start a company, mostly because we want to move at different speeds. I'd be glad to stay in touch, and if I hear of someone who'd fit you better, I'll send them your way."

Then do what the brief said about who keeps what.

Extend. Only when there's a specific new question and both of you want to answer it. Write a new end date, a new question and a smaller project, and extend once. If you're extending because neither of you wants to say no, say no.

Sometimes both people are sure before the end. The Whalesync founders wrote that they "felt so aligned" they were ready to commit before the trial was over. That's a good sign. Still hold the review: it takes an hour and puts on record what you both expect next.

What not to do

• Let the project grow. If the scope changes, change the brief and the end date together, on purpose.
• Turn it into free work. A trial is two people testing each other. If one person does most of the work while the other evaluates, it's unpaid consulting. Hustle Fund calls this "the free labor trap".
• Chase perfection. Hustle Fund lists perfectionism as a trap: someone who can't ship "good enough" work in a trial may struggle with the pace of an early company. Watch for it in yourself, too.
• Quit your job mid-trial. Leave that decision until after the review.
• Skip the brief because you get on well. Liking each other is the reason to write it, not to skip it. Misunderstandings between people who get on are the ones nobody raises.
• Run the trial in silence. If something bothers you in week one, mention it in week one. Raising it at the final review leaves no time to see whether it changes.

Running a trial remotely or part-time

Most trials now happen at least partly remotely, and many around a job. A few adjustments help:

• Agree on working blocks. Name two or three times a week when you're both working on the trial at once, even if it's just two evenings.
• Write more. Remote trials live or die on written updates. The midweek message matters more than the call.
• Meet in person once if you can. A day working in the same room shows things a video call can't, such as how each of you handles interruptions and a long afternoon.
• Allow for time zones. If you're more than a few hours apart, put the check-in at a time that's fair to both and rotate it if it isn't.

A sample four-week trial plan

This is a made-up example to show the shape. Two people, each giving about 15 hours a week around a job, testing an idea for scheduling software for small physiotherapy clinics.

The question. Can we decide quickly together, and does each of us follow through on customer work?

The deliverable. Twelve interviews with clinic owners, a clickable prototype of the booking screen shown to five of them, and a one-page summary by Friday of week four.

Roles. One leads interviews and the summary. The other leads the prototype. Both run the five prototype sessions together.

Week one. Write and agree the brief. Agree the interview questions. Book the first six interviews.

Week two. Run interviews. Start the prototype from what you hear. First check-in, first decision log entries.

Week three. Finish interviews. Show the prototype to five clinic owners together. Give each other one piece of feedback.

Week four. Write the summary. Write separate reflections. Decision meeting on Friday.

Where Foundinity fits

A trial tests what two people do together. It works best when you start it with someone who already looks like a good fit on the things that are hardest to change, such as commitment, pace and how you handle disagreement.

On Foundinity, you state what you need in a cofounder, and answer an assessment about how you work: commitment, working style, ambition, personality, goals, values, alongside your skills and interests. A multi-trait algorithm built on established psychological compatibility models ranks the people you see.

• A compatibility breakdown on every match shows where the two of you align and where you differ. Turn the differences into your trial's question.
• A 10-day clock on each match gets you to a first call quickly, and you book the intro call in the app.
• Your profile isn't public. Only compatible members can see it, and an "exploring" status lets you look while you're still in a job.
• Pricing: free to get started; Plus is $5 a week or $15 a month. See how it works.

FAQ

What is a cofounder trial project?

A short piece of real work that two potential cofounders do together for a fixed period, usually 2 to 6 weeks, to find out whether they work well together before committing to a company. It ends with a decision meeting on a date set at the start.

How long should a cofounder trial last?

Most advice falls between two weeks and a month or two. Hustle Fund suggests two to three weeks; Steve Blank suggests at least a provisional 30 days. Set the length by the hours you each have: two to three weeks full-time, five to six weeks on evenings and weekends.

What's a good cofounder trial project idea?

Something real, small and finishable that needs both of you and gets a reaction from customers: 10 to 15 customer interviews with a written summary, a landing page with a message test, a clickable prototype shown to five users, or serving three customers by hand.

How do you test a cofounder before committing?

Ask the hard questions first, then work together on a trial project with a written brief, weekly check-ins and a decision date. Watch how you decide, disagree and follow through, and keep a simple log of the decisions you make.

Should you pay a potential cofounder during a trial?

Usually neither person pays the other: it's two people testing each other, not a hire. If the work is lopsided, for instance one person is building on top of the other's existing product, talk about it before you start and write the answer in the brief.

Who owns the work from a trial project?

Whatever you agree in writing before you start. Common choices are that each person keeps what they built, or that the person who brought the idea keeps the work if you stop. Put it in the brief, and take legal advice if either of you brings existing code or an idea from a current job.

What if the trial project fails?

A project that doesn't hit its goal can still be a good trial. What matters is how you handled the miss: whether you saw it coming, told each other early and adjusted together. Judge the partnership, not the deliverable.

Can you run a cofounder trial remotely?

Yes. Agree fixed working blocks, write a short midweek update, keep the weekly call, and meet in person once if you can.

How do you tell someone the trial didn't work?

Directly and kindly, soon after the review. Thank them for the work, give one honest reason without a long list, and do what the brief said about who keeps what. Many people who aren't your cofounder know someone who might be.

Is a hackathon a good cofounder trial?

It's a good first step. A weekend shows how someone handles a hard deadline. It says less about week five, when the excitement has gone, so follow it with a trial of a few weeks. That's Steve Blank's order: a "hack weekend" first, then the 30-day period.

What should be in a cofounder trial agreement?

The question you want answered, the deliverable, the dates and decision date, each person's weekly hours, roles, the check-in schedule, how you'll communicate, any spending, who keeps what if you stop, and how either of you can pause. One page is enough.

Sources

• Gloria Lin, "The Founder Dating Playbook", First Round Review, checked 11 November 2026.
• Steve Blank in First Round Review, "Looking for Love in All the Wrong Places: How to Find a Co-Founder", 26 January 2015, checked 11 November 2026.
• Hustle Fund, The Founder Playbook, "Trial projects part 1", 9 October 2025, and "Trial projects part 2", 16 October 2025, checked 11 November 2026.
• Matthew Busel, "Co-Founder Matching Playbook", Whalesync blog, 8 May 2023, checked 11 November 2026.
• Annie Garofalo, "Co-Founder Dating: Try Before You Buy", The Co-Founder Confidante, 22 January 2024, checked 11 November 2026.
• Rob Fitzpatrick, The Mom Test, book site checked 11 November 2026.
• Foundinity plans and how it works from foundinity.com, checked 11 November 2026.