How to find a team for a pet project

Where to look

Among people you know. Coursemates, former colleagues, people from a course. The simplest route and the most overlooked: you already know how they work.

In topic communities. Chats and forums around languages and technologies. The people there are already building something, which is a good sign.

At hackathons. A day of working together tells you more than a month of messaging. Some teams carry on afterwards.

On sites where people look for teams. The point is that everyone there arrived intending to take part rather than to talk.

Who to take

Not the strongest, but the one who'll finish. They're different people, and the first option lets you down more often.

Signs someone will finish:

• They have finished things. Small ones are fine. One completed project says more than a dozen started.
• They ask about scope and deadlines. That means they're judging their capacity rather than getting swept up in the idea.
• They name their hours honestly. "Six hours a week" is more reliable than "I'll go all in".
• They reply quickly. Someone who vanishes for a week while you're still getting acquainted will vanish during the work.

Signs it won't work

• Lots of talk about the idea, nothing about what they'll personally do.
• Wants to "be the architect" and not write code.
• Can't name a single finished piece of work.
• Insists on a grand plan: a platform, a social network, a marketplace.

How many people

Two to four. Two is the most stable for a first attempt: easy to agree on things, and mutual accountability already works.

Five or more almost always collapses: too much coordination and too easy to hide behind others.

What to agree before starting

Half an hour of conversation saves a month:

1. What we're building. One sentence everyone understands the same way.
2. What counts as finished. The result that tells us we're done.
3. How much time each person has. Honestly, in hours per week.
4. Who takes what. Not "we'll figure it out", but specifically.
5. What happens if someone disappears. Settled in advance, without hard feelings.

Not losing people at the start

Start building in the first week. Teams that spend two weeks discussing usually never begin: talking is more pleasant than working, and it's sticky.

And make the work visible. A shared task board where everyone notes what they've taken is the simplest way to spot a stall in time rather than a month later.