Gacha banner revenue forecast: pulls, pity, and refund risk
UA spend against banner weekends needs a pull-volume forecast with explicit store fee and refund inputs — illustrative sheet.
UA spend against a banner weekend needs a pull-volume forecast that still shows fees and refunds. Without those rows, “banner ROAS” is theatre.
Fee definition page
Google Play Service fees (source read 20 September 2026) is the definitional source for the fee-fraction input. Use your live programme rate; the prior-schedule 15%/30% split and the dated EEA/UK/US redesign both live on that page.
Illustrative forecast
| Input | Value | Class |
|---|---|---|
| Forecast paid pulls over banner weekend | 50,000 | synthetic |
| Average revenue per paid pull (list) | 1.60 | synthetic |
| Refund rate on banner gross | 3% | synthetic |
| Store fee fraction | 0.15 | synthetic input (see Google page) |
| Incremental UA spend for banner weekend | 20,000 | synthetic |
| Output | Result |
|---|---|
| Gross | 80,000 |
| After refunds | 77,600 |
| After 15% fee | 65,960 |
| Contribution after UA | 45,960 |
| Gross ROAS on UA | 4.00 |
| Net-after-fee ROAS on UA | 3.30 |
Pity affects pull volume and ARPU mix; model it as a separate volume scenario, not a hidden uplift. Refund rate is a finance input — not an industry constant.
Facts vs limits
| Fact | Limit |
|---|---|
| Fee schedule is documented by Google | Pull volume and refund % here are synthetic |
| Contribution after fee ≠ gross ROAS | Apple proceeds schedule not quoted as a universal rate |
| Teaching sheet is reproducible | No title’s live banner KPI is claimed |
