No single customer lifetime value model fits every business. The appropriate model depends on the decision, available history, purchase pattern and whether the output represents revenue, gross profit or contribution.
Model selection summary
| Model | Minimum inputs | Best use | Main limitation |
|---|---|---|---|
| Simple purchase model | AOV, frequency, lifespan, margin | Transparent planning | Assumes stable averages |
| Churn-based model | Periodic value and churn | Stable subscriptions | Assumes a stable cancellation pattern |
| Historical cohort model | Customer-level orders and costs | Observed channel or cohort comparison | Recent cohorts are incomplete |
| Retention-curve model | Retention by age and contribution | Forecasting repeat behavior | Sensitive to curve choice |
| Discounted cash-flow model | Periodic cash flow, survival and discount rate | Capital and long-horizon decisions | Requires more assumptions |
Always label the model, observation date, customer cohort and value layer.
Simple purchase model
Average order value × purchase frequency × lifespan × margin is transparent and useful for planning. Its weakness is the assumption that averages remain stable.
CLV = AOV × purchases per year × lifespan × margin rate
With $80 AOV, four annual purchases, a three-year lifespan and 45% gross margin:
$80 × 4 × 3 × 45% = $432 gross-profit CLV
The Customer Lifetime Value Calculator reproduces this model and also shows lifetime revenue of $80 × 4 × 3 = $960.
This approach is useful for scenario planning. It should not be described as observed lifetime value unless the inputs come from completed customer histories.
Churn-based model
Subscription businesses sometimes approximate lifespan as 1 ÷ churn rate. This is sensitive to period choice and assumes a stable hazard of cancellation.
If monthly contribution per customer is $30 and monthly customer churn is 5%:
Estimated lifespan = 1 ÷ 0.05 = 20 months
Simplified contribution CLV = $30 × 20 = $600
Do not insert an annual churn percentage into a monthly contribution formula. The period must match. The model also becomes unstable at very low churn and undefined at zero churn, which is evidence that the shortcut is not a literal lifetime forecast.
The retention and churn guide explains denominator and cohort timing choices.
Cohort model
Track cumulative contribution by acquisition cohort. This reveals channel and seasonal differences without forcing one retention curve onto every customer.
Suppose 100 customers acquired in January produce the following cumulative contribution after refunds and variable service costs:
| Cohort age | Cumulative contribution | Contribution per original customer |
|---|---|---|
| First order | $4,000 | $40 |
| 90 days | $6,500 | $65 |
| 180 days | $8,200 | $82 |
| 365 days | $10,400 | $104 |
The observed one-year contribution value is $104 per original customer. This is not a lifetime forecast. It is a comparable cohort milestone. Compare different acquisition cohorts at the same age.
Historical cohort analysis avoids assuming a lifespan but cannot reveal the eventual value of young cohorts without a forecast layer.
Retention-curve model
A retention model estimates contribution in each future period from the probability that a customer remains active.
Expected CLV = sum of expected contribution in each period
If expected contribution per active customer is $25 per quarter and modeled retention is 100%, 70%, 50% and 35% across four quarters, undiscounted expected contribution is:
$25 × (1.00 + 0.70 + 0.50 + 0.35) = $63.75
This structure makes changing retention visible. It also exposes the assumptions that a single average lifespan hides.
Discounted model
Finance-oriented models discount future cash flows and may include survival probabilities. They are more precise but require reliable historical data and explicit assumptions.
Discounted CLV = sum of expected period contribution ÷ (1 + discount rate)^period
Discounting gives less weight to cash expected further in the future. A more complex formula is not automatically more accurate. Weak retention estimates, missing costs or an arbitrary discount rate can create precise-looking output with poor decision value.
Keep the value layer consistent
| Output label | Revenue and costs included |
|---|---|
| Revenue CLV | Customer revenue only |
| Gross-profit CLV | Revenue less product or direct service cost |
| Contribution CLV | Revenue less the defined variable costs |
| Net customer value | Contribution less allocated acquisition and other costs |
Use contribution-based value when comparing with fully loaded CAC where the data supports it. The LTV:CAC Calculator does not correct incompatible definitions automatically.
Decision sequence
- State the decision and required time horizon.
- Choose revenue, gross profit or contribution as the value layer.
- Define acquisition cohort and observation date.
- Start with historical cohort evidence.
- Add the simplest forecast that resolves the missing horizon.
- Test retention, margin and discount-rate assumptions separately.
- Compare predicted value with later observed cohort outcomes.
Common mistakes
- Mixing monthly churn with annual contribution.
- Comparing mature and recent cohorts at different ages.
- Calling revenue CLV profit.
- Subtracting CAC inside CLV and then subtracting it again in LTV:CAC.
- Treating
1 ÷ churnas a guaranteed lifespan. - Adding model complexity without better underlying data.
Use the simplest model that supports the decision, and label whether the output represents revenue, gross profit or contribution.
Sources
- Shopify: customer lifetime value; Shopify; accessed Jul 13, 2026
- Stripe: retention rate versus churn rate; Stripe; accessed Jul 13, 2026
This guide is educational and does not provide financial, accounting, tax or legal advice.