Prediction Studio

Import from RevenueCat, Adapty or Apphud, or upload a cohort CSV, to start.

These results use previous data. Import data again before you export the report.

1. Import data

Provider Saved setup

No saved setup selected.

RevenueCat API key Project ID App name Start date End date

Pull full history Use 7-day LTV as sub price

Project

Or upload a cohort CSV in the template format.

No file selected.

2. Select training window

Import full history first, then choose where stable forecasting starts.

Imported cohorts: —
Training starts: —
Training ends: —
Used cohorts: —
Excluded early cohorts: —
Excluded recent cohorts: —

3. Model options

Duration split modelWeekly, monthly and yearly plans are modelled separately, then summed. RevenueCat historical tailUses RevenueCat's Prediction Explorer for cohorts that already paid. Alpha still drives future new cohorts. Gradual decayForecasts new paid subs from the app's own fitted decline, slowing with app age. Never grows. Full retention historyBuilds the retention curve from every cohort except the incomplete last one. The training window still drives the new-subs forecast. Skip spikesLeaves cohorts with Subs above 3x the median of their 8 neighbours out of the retention curve. Needs full retention history.

4. Economics

Trend blend: 0 follows the recent average, 1 follows the trend line; the scenarios use 0.3, 0.5 and 0.7. Acquisition price is your offer and is only needed for annual ROI.

Weighted Retention Curve

Future Monthly Net Revenue

Forecast Cohorts

#	Cohort date	Forecast new subs	Baseline active subs	Week-0 net revenue

Future Weekly Projection

Period	Date	Historical active	Forecast active	Total active	Net revenue	Cumulative profit