First-purchase offer holdout: synthetic teaching set

A first-purchase offer without a holdout confuses discount response with baseline payer conversion. Use AppsFlyer in-app events to define the purchase event. This lab is a labelled synthetic teaching set.

Define the calculation

Payer conversion = payers / installs. Incremental payers = holdout gap × treatment installs. Incremental revenue = incremental payers × ARPPU assumption.

Synthetic example

Treatment payer rate 6.0%, holdout 4.0%, treatment installs 10,000 → incremental payers 200. ARPPU 12 → incremental revenue 2,400. Offer cost (discount depth × payers) is a separate finance input — not invented here as an industry average.

Limits

All figures synthetic. AppsFlyer defines event logging, not your holdout design ethics or sample-size needs.

Download the evidence table.

Research checked 20 September 2026. Local draft; human editorial review pending. No CMS write performed.