Rewarded opt-in rate decay across a 14-day cohort (labelled synthetic)
Method lab: rewarded opt-in rate by cohort day with illustrative inputs plus AdMob’s rewarded-ad definition. Not an industry average.
Fatigue in hybrid and IAA-heavy titles often shows first as a falling rewarded opt-in rate, before CPI moves. AdMob rewarded ads define the format as ads that reward users with in-app items for interacting with video, playable or survey ads, with an OnUserEarnedRewardListener when the reward is earned. That is a product definition, not a retention benchmark.
This lab uses illustrative cohort inputs. It is a method worksheet, not a claim about your live title.
Define the metric before using it
- Eligible session: session where a rewarded placement was offered (studio event).
- Opt-in: user started the rewarded experience (studio event; not the same as reward grant).
- Opt-in rate (day d): opt-ins on cohort day d ÷ eligible sessions on day d.
- Decay (d0→d14): 1 − (rate_d14 ÷ rate_d0), when rate_d0 > 0.
Map reward grants and impression revenue separately via your mediation/MMP path (for example AppsFlyer’s af_ad_revenue event in the ad revenue attribution guide). Do not treat reward grant as IAP.
Synthetic worked example
| Cohort day | Eligible sessions | Opt-ins | Opt-in rate |
|---|---|---|---|
| 0 | 1000 | 420 | 0.420 |
| 3 | 700 | 245 | 0.350 |
| 7 | 500 | 150 | 0.300 |
| 14 | 350 | 84 | 0.240 |
Decay d0→d14 = 1 − (0.240 / 0.420) = 0.4286 (42.86%).
| Output | Result | Equation |
|---|---|---|
| rate_d0 | 0.4200 | 420/1000 |
| rate_d14 | 0.2400 | 84/350 |
| decay_d0_d14 | 0.4286 | 1 − rate_d14/rate_d0 |
Facts versus limits
| Fact | Limit |
|---|---|
| AdMob documents rewarded ads and reward callbacks | Docs do not publish opt-in decay norms |
| MMP ad-revenue events exist for attributed IAA | Event presence ≠ fatigue diagnosis |
| Synthetic arithmetic is reproducible | Not a genre benchmark or CPI forecast |
Download the CSV. Replace synthetic rows with your export; keep the equations fixed.
