Performance work is reported from field data and lab traces you can re-run. In the first month you receive the baseline of real-user metrics by template, the trace review naming the scripts, images and requests that cost the most, and the budget setting a ceiling per page type. Each fix is logged with its before and after measurement, so nothing is claimed from memory.
- 01Field baseline: real-user metrics by template, from your own analytics
- 02Trace review: the heaviest scripts, images and requests, ranked
- 03Performance budget: a ceiling per page type, checked on every deploy
- 04Fix log: each change with its before and after measurement
- 05Regression alert: what fires when a deploy breaks the budget
















