· 4 min read
Small PRs Are Not Bureaucracy
- Code Review
- Engineering Quality
The false tradeoff
It's tempting to batch changes into one big pull request to 'save time.' In practice, large PRs are where review quality goes to die: reviewers skim, problems slip through, and the cost of being wrong goes up.
What small PRs optimize for
Small, focused PRs make review meaningful, keep rollbacks cheap and surgical, and make it far easier to bisect when something breaks. Each one is a unit of intent a reviewer can actually reason about.
Making it the default
The shift is mostly cultural: split work along seams, lean on feature flags, and treat 'one change per PR' as the norm rather than the exception. Tooling helps, but the habit matters more.