ANALYTICS & REPORTING
Acquisition payback reading
Use when you need to know how long each channel takes to pay back and how solid that number is.
acquisition-payback-reading.md
You are reading acquisition economics. You are not producing a valuation and you are not projecting beyond the data.
Spend and new customers by channel and month: {{SPEND_AND_ACQUISITION}}
Revenue or margin per customer by month since acquisition: {{REVENUE_CURVE}}
What the cost figure currently includes: {{COST_INCLUSIONS}}
Gross margin and any variable servicing cost: {{MARGIN_ASSUMPTIONS}}
Output:
1. A table: Channel | Acquisition cost per customer | Months to payback on margin | Months of data actually observed | Payback observed or extrapolated.
2. The full arithmetic for one channel, so the rest can be checked by hand.
3. What changes when {{COST_INCLUSIONS}} widens to cover the costs currently left out. Name those costs and show the second set of numbers beside the first.
4. Where {{REVENUE_CURVE}} runs out before payback, how many months are observed and what assumption anything beyond that rests on.
5. Channels whose ranking would flip under a plausible change to {{MARGIN_ASSUMPTIONS}}, and how large that change would need to be.
Rules:
- Payback on revenue and payback on margin are different numbers. Report margin, and label any figure that is revenue only.
- Do not report a lifetime value multiple. If asked for one, state what the curve would have to do beyond the observed window.
- Cohorts in {{SPEND_AND_ACQUISITION}} with fewer than 30 customers are directional. Label them and do not rank on them.
Vor dem Start ausfüllen
Ersetzen Sie jeden Platzhalter durch Ihre eigenen Angaben. Je genauer Sie sind, desto weniger erfindet das Modell.
- {{SPEND_AND_ACQUISITION}}
- {{REVENUE_CURVE}}
- {{COST_INCLUSIONS}}
- {{MARGIN_ASSUMPTIONS}}
So wird das Ergebnis besser
-
1
Put agency fees, creative production and discounts into the cost figure, or section 3 becomes the real answer.
-
2
Use monthly cohorts rather than a blended average; a good month and a bad month cancel out and hide both.
-
3
Recheck the extrapolated rows each quarter as the curve matures rather than restating the original figure.