I have one rule for side-projects: ship something usable in the first week, or kill it.
Most projects die because the gap between "the idea is good" and "this thing now exists" is 90% of the work, and that work is unglamorous. Plumbing, deployment, the boring parts. If you can't tolerate the boring parts in the first week, you'll never tolerate them in month three when there's a real bug.
The corollary: scope hard, ship ugly, refactor later. A v0.1 that works beats a v1.0 that's still in your head.