Code Review Velocity – Need for Speed

Interesting stuff! I feel somewhat validated to see that there's not an empirical relationship between change size and time to merge because I've always disliked the "small PRs police".

I'd personally argue that the most important metric is "time to consensus". There are sometimes good reasons to hold off on merging an approved PR (external effects) but that should not reflect poorly on how your team does code review. I define consensus as when the author and the reviewer have resolved all blocking conversations.

That's the model I've used on CodeApprove (https://codeapprove.com) which focuses somewhat obsessively on unresolved conversations as the key element of a code review.