What we ask for in the first document pack
A health review cannot start on a slide. This is the pack we request, why each item is there, and what happens if a piece is missing.
The pack is due before we start the ten-day clock. We send a checklist after the intake call. It is not a fishing exercise; each item maps to a page in the brief.
Crash extracts for each live title, two quarters, with the sampling note attached. If your vendor samples, we need the sampling rate in writing. Retention tables at day 1, 7, and 30, split by the same version windows as the crash extracts. A release calendar with storefront submission dates, not just ‘we shipped in May’. Incident records that mention an app, a version, or a shared library, including the ones closed as ‘user error’.
We also ask for the list of shared libraries and webviews, even if it is a messy spreadsheet. And a note on identity: do the titles share a login, a customer number, a device id, or nothing? ‘Nothing’ is an acceptable answer. Pretending they share a journey when they do not is not.
Storefront listings, ratings comments, and marketing calendars are optional. We read them when a retention cliff lines up with a campaign or a forced update. We do not need design files, brand decks, or the original business case for each app.
If crash extracts are missing for one title, that title is marked unscored on the heat table. We do not interpolate. If the release calendar is a set of Slack messages, we will still work, but the brief will say the dates are reconstructed. Incomplete packs pause the clock; we email once, then wait. Guessing a version number to keep a steering date is how reviews become theatre.