A simple system for finishing what you start

The projects that stall rarely stall at the start. They stall at 70%, where the interesting part is done and the tedious part is all that is left.

Scope to a demo, not to a vision

Before you begin, write down the thing you will show someone. Not the finished product — the demo. "A page that loads with three real products on it." That sentence is your scope. Everything not required to reach it is version two.

Keep one list, and keep it embarrassingly short

A list of forty tasks is a mood board, not a plan. Keep three. When one is done, pull the next in. The friction of choosing what earns a slot is the useful part.

End every session mid-sentence

Stop while you still know exactly what comes next, and write it down. The cost of restarting a project is almost entirely the cost of rebuilding context, and a single line of "next: wire the coupon field to the API" removes most of it.

The weekly question

Once a week, ask: if I had to ship this on Friday, what would I cut? Then cut it now. You will be surprised how often the answer is the part you were dreading.