[disposedtrolley]

> Ask them for a narrower review — e.g., a design review or an architecture review

I like this technique for PRs of any size (except very trivial changes). Code reviews can be a great learning opportunity, made even better if the author has specific aspects of the code they'd like feedback on.

For example, I opened a PR the other week with a bunch of refactoring work. One chunk of the refactor changed how a complex data transformation was being done by replacing some unnecessary local variables with a chain of map, filter, and forEach functions. I highlighted this as a particular point of interest, which spawned a good discussion on testability, how small our tests should be, and if there's any value testing specific bits of the transform vs. the chain as a whole.