Offer-wall CPI quality: when cheap installs fail D7 payback (labelled synthetic)
Synthetic worksheet comparing offer-wall CPI against D7 net contribution using labelled inputs and MMP ad-revenue event definitions. Not a partner ranking.
Offer-wall volume still tempts UA leads when CPI targets tighten. Cheap installs fail when D7 net contribution (IAP proceeds + attributed IAA − refunds − offer cost) stays below acquisition cost. This sheet uses illustrative inputs. It does not invent partner quality scores.
Metric definitions
Use your MMP’s revenue events rather than network screenshots alone. AppsFlyer’s ad revenue attribution guide documents af_ad_revenue as the event name for attributed ad revenue. IAP should use your purchase event with a clear gross-versus-net rule. Offer-wall “CPI” here means cost per attributed install from the offer-wall partner report—confirm the install definition matches the MMP.
Equations (reader inputs):
d7_net = d7_iap_net + d7_iaa − d7_refundspayback_gap = d7_net − cpipasses = payback_gap >= 0
Synthetic example
| Input | Value | Status |
|---|---|---|
| Offer-wall CPI | 0.40 | illustrative |
| D7 IAP net / install | 0.12 | illustrative |
| D7 IAA / install | 0.18 | illustrative |
| D7 refunds / install | 0.02 | illustrative |
Then d7_net = 0.12 + 0.18 − 0.02 = 0.28, payback_gap = 0.28 − 0.40 = −0.12 (fails D7 payback despite “cheap” CPI).
| Output | Result |
|---|---|
| d7_net | 0.2800 |
| payback_gap | −0.1200 |
| passes_d7 | false |
Facts versus limits
| Fact | Limit |
|---|---|
| MMP docs define ad-revenue event naming | Docs do not publish offer-wall quality bands |
| Arithmetic is reproducible from disclosed inputs | Not a recommendation to ban or buy offer-walls |
| Partner reporting fields vary | research_needed remains for partner-specific schemas |
Download the CSV. Replace inputs with your cohort export.
