limited·by·sleep
$ cd ~/published

How quickly can a new starter commit code?

How quickly can a new starter commit code? That question tells you more about your project's health than most metrics on your dashboard.

I've used this as a benchmark for the last 10 years, whether joining a team or leading one. The range is wild. Some teams have people shipping on day one. Others take weeks before someone can even run the application locally.

A slow time to first commit usually points to deeper problems. Fragile setup processes held together by tribal knowledge. Missing tests that would help someone understand what the code actually does. An assumption that documentation will cover the gaps, when half of it is outdated and the other half assumes context nobody wrote down.

I stopped treating onboarding documentation as the safety net a while ago. The real documentation is the code, the tests, and whether a fresh install actually works. If someone can't clone a repo and get running without pulling three people into a call, that's not a new starter problem. That's a team problem.

One thing I always tried to keep available was a handful of small tidying tickets. Not throwaway work, real tasks that needed doing but were low risk and well contained. Nothing gets someone feeling part of a team faster than shipping something on their first day. It builds confidence, gives them a reason to touch the codebase, and quietly proves the whole setup actually works.

The time it takes a new person to contribute isn't just about them. It reflects the care a team puts into how they work together.

follow via rss