App analytics
Retention Cohort Rebuild
Rebuild who returns after first value—and whether the chart moved because of product, or clocks and identity drift.
£7,400 · 3-week study
New releases look successful until someone asks about week-four return among people who actually reached first value. UTC warehouses mixed with UK product calendars, SDK upgrades, and identity merges invent retention lifts that never happened in the app.
This study rebuilds cohorts from raw history, checks identity continuity across surfaces, and tests early behaviours against later return. Sample limits stay visible—we refuse to dress pipeline artefacts as insights.
You leave with behaviours worth protecting, artefacts to stop celebrating, and a cohort outline your analysts can maintain after we leave.
Typically included
- Retention cohort rebuild from raw events
- Identity continuity checks
- Behaviour shortlist for the next release
- Analyst handoff outline