Six weeks of optimisation, and the fix was two lines
Profiling found the hot function. The hot function was called a thousand times for one reason.
measurement
4 posts
We shipped it, measured it, and deleted it eleven weeks later
Deciding the success metric before building is what makes deletion possible afterwards.
Engagement went up, and it was because the app got slower
Time-in-app rose because loading took longer. A metric that improves under a regression is measuring the wrong thing.
I fixed 30 of 31 failures. Nineteen were left.
A checker that stops at the first problem per item reports a queue depth, not a defect count.