Ship the small version first
A note to myself about scope, written after cutting a three-week project down to an afternoon.
1 min read
- process
I had a three-week plan. It had a database, an auth flow, a settings page, and a billing integration. What I actually needed to learn was whether anyone wanted the thing at all.
The afternoon version was a single page with a text box and a button. It answered the question. The three-week version would have answered it too, three weeks later, with more code to delete.
The rule I'm trying to follow
Build the version that answers the question you actually have.
Not the version that would be embarrassing to show a senior engineer. Not the version that handles every edge case. The version that tells you what to build next.
When this doesn't apply
Some things have no small version. A payment system either handles money correctly or it doesn't. But most things I build aren't payment systems, and I keep planning like they are.