Three steps: map where your customers sit, see what each route to a sale costs you, then move the sliders to estimate the profit each play could recover. Fill the amber boxes; everything else is calculated.
Two numbers turn every percentage below into money.
Split your customers by how often they buy (rows) and how recently they paid attention (columns). Enter C% = share of customers and R% = share of revenue in each cell. The None row has no revenue. Totals should reach 100%.
| Stronglistening · 0–30d | Weakeningfading · 30–90d | Lostgone quiet · 90d+ | Row total | |
|---|---|---|---|---|
| RepeatBest · 2+ | B C % R % = $0.00M | B− C % R % = $0.00M | R1 C % R % = $0.00M highest risk · drifting to adtech | C 0% · R 0% |
| OneTest · 1 | T C % R % = $0.00M | T− C % R % = $0.00M | R2 C % R % = $0.00M | C 0% · R 0% |
| NoneNext · 0 | N C % no revenue | N− C % no revenue | R3 C % no revenue | C 0% |
| Total | C 0% · R 0% | C 0% · R 0% | C 0% · R 0% | customers — revenue — |
Every sale arrives through one of five routes, each charging a different tax. Set the tax rate for each route and the revenue share (R%) in each bucket. New = a first-time buyer; Repeat = someone you already owned. Should total 100%.
Each play acts on the revenue you mapped above. Drag a slider to set how hard the play works — the profit updates live. These are estimates from your assumptions; a holdout test is what turns an estimate into proven profit. (Capture — getting anonymous and marketplace buyers onto the grid — is the on-ramp before these five and isn’t scored here.)