Field note
How to read crash-free session rates before a store release
A practical guide to interpreting crash-free session percentages against your own release history, not industry averages.
Crash-free session rate is useful only when you compare it to your own past releases. A number that looks healthy in isolation can still be a regression if last month’s build ran quieter on the same device mix.
Start with the window that matches how users actually adopt updates. For many consumer apps in Thailand, the first 48 hours after a store push concentrate enough traffic to judge whether a new crash family is spreading. Stretch that window only when your audience updates slowly.
Separate foreground crashes from background failures when your telemetry allows it. Background noise can inflate volume without matching the frustration users feel during active use. Your release meeting should weight interactive sessions more heavily.
Pair the rate with a short list of top signatures. A stable overall percentage can hide one severe crash that hits a smaller but critical flow—checkout, login, or onboarding. Rank by affected unique users when deciding blockers.
Document the threshold you used for go or hold. Next month’s team will thank you for a paper trail that explains why a borderline rate shipped or waited.