Fictional example: a cohort observed for twelve months
For 100 customers, observed net revenue is 30000 EUR and defined variable costs are 18000 EUR: production, delivery, payment and variable service costs. Contribution is 12000 EUR, or 120 EUR per original customer. Acquisition costs and fixed costs are excluded here.
| Fictional twelve-month measure | Value | Calculation or limitation |
|---|---|---|
| Original customers | 100 | Denominator retained. |
| Net revenue (EUR) | 30000 | Observed amount. |
| Defined variable costs (EUR) | 18000 | Scope specified in the text. |
| Contribution (EUR) | 12000 | 30000 − 18000. |
| Revenue/customer (EUR) | 300 | 30000 / 100. |
| Contribution/customer (EUR) | 120 | 12000 / 100. |
Make assumptions inspectable
- Keep all 100 original customers in the denominator, including those who stop buying. Document refunds, taxes, included costs and the horizon.
- Separate observations from projections. A forecast specifies retention, future revenue and costs, the horizon and any discounting; test several assumptions.
- Deduplicate customers across channels. An attribution convention allocates an outcome; it does not prove that a channel caused it.
Do not turn an estimate into net profit
The 120 EUR describes observed twelve-month contribution, not net profit or an already known lifetime value. When comparing with CAC, state both scopes and avoid counting acquisition costs twice. Show variation too: an average can hide a very uneven cohort.
Primary references
Separate historical value from estimated CLV
Cumulative revenue over a period is observed, whereas projected lifetime value depends on assumptions. Define customer, margin or revenue basis, costs, horizon and refunds. Compare cohorts across consistent durations before concluding that one channel acquires longer-lasting customers.
- Show observed value separately from projections.
- Document horizon and included costs.
- Examine sensitivity to repeat-purchase assumptions.
Can a simple formula be presented as a certain forecast?
No. A simplified model can illuminate a scenario when assumptions are visible. Check it against available history and present uncertainty rather than a guaranteed single value.
Primary reference: Google SRE — Data processing pipelines. Related method: read the practical guide.
Separate observed CLV from a projection
Fictional example: these inputs describe no customer or observed result.
A dashboard presents estimated future revenue for recently acquired customers as value already earned.
- Separate observed contribution and projection, including horizon, selected costs, returns and retention assumptions. Compare cohorts at equal age.
- Show alternative assumptions and a review date. Projected value is neither recognised revenue nor a guarantee of future retention.
Primary reference: Google Analytics — Cohort exploration. Related method: read the practical guide.