How to Stop Starting New Side Projects

I have started over a hundred side projects. I have finished four. The failure was never technical, and it was never a lack of time. Each project died the week a newer idea showed up.

Starting is a reward, finishing is a cost

A new project pays you immediately. You get to name it, pick the stack, draw the schema, see the first screen render. None of that is the work. The work is the unglamorous middle, and the middle never feels as good as a blank repo. So the brain does the rational thing and goes back for another hit. That is not weak discipline, it is an incentive problem, and you cannot willpower your way out of an incentive problem.

Why the usual advice fails

Most advice about how to finish side projects is a feeling, not a mechanism. Focus more. Want it badly enough. Timebox it. All of that evaporates the moment a better idea arrives, because none of it physically prevents the next start.

What works is a constraint you cannot negotiate with at midnight. Not a reminder, not a habit tracker. A rule that refuses.

One active project, enforced

The constraint that changed it for me is allowing exactly one active project at a time. Every new idea goes into a queue and waits. Nothing is promoted out of that queue while something else is live.

The important word is enforced. A rule you can click past is a suggestion. When I built One Project OS I put that rule in the database as a unique index rather than in the application code, so two fast clicks cannot produce two active projects. There is no override button anywhere in the app, because I know I would use it.

What actually changes

The queue also does something unexpected: most ideas die in it, on their own, without you having to argue with yourself. More on that in the 14 day rule for side project ideas.

One Project OS enforces all of this: one active project, a 14 day gate on new ideas, and no way to close a project without saying why.

Try it free
All posts