Calculating LTV

The most widely used LTV formula multiplies average revenue per user by gross margin, then divides by churn rate.

Dividing by churn is mathematically equivalent to multiplying by average lifetime, so this single line captures revenue, profitability, and retention together.

Suppose a streaming service charges 12 per month, runs a 70% gross margin, and loses 4% of subscribers each month.

Plugging these into the formula gives a clean, defensible LTV that you can compare directly against acquisition cost.

ARPU and churn must use the same time unit. Mixing monthly ARPU with annual churn produces wildly wrong numbers.

For subscription products, monthly is natural. For retail, a yearly period often fits better because purchases are sporadic. Pick one unit and apply it consistently across every input.

Non-subscription businesses calculate LTV from purchase behavior: average order value, purchase frequency, and customer lifespan.

This version suits ecommerce and retail where there is no recurring fee, only repeated discretionary purchases.

For long-lived customers, future margin should be discounted to present value. The adjusted formula divides ongoing margin by churn plus the discount rate.

This produces a more conservative LTV that finance teams trust, especially when lifetimes stretch beyond two or three years.

Rather than one formula, cohort analysis tracks groups of customers acquired in the same month and plots cumulative margin over time.

The curve typically rises steeply then flattens. Extrapolating the flattening tail gives an empirical LTV grounded in real behavior instead of a single churn assumption.

Real churn is rarely constant. New customers churn fastest, while survivors become increasingly loyal, so a flat churn rate underestimates long-term value.

Survival curves or retention-by-month tables capture this shape. Multiplying period margin by surviving customers in each period yields a more accurate sum.

Advanced teams use probabilistic models such as BG/NBD for purchase frequency and Gamma-Gamma for spend value.

These statistical models estimate, per customer, the probability of being alive and their expected future transactions, producing individual LTV predictions rather than a single average.

Always cross-check a formula-based LTV against historical cohort data. If the formula predicts 400 but no cohort has ever exceeded 250, the inputs are wrong.

Common culprits are optimistic churn assumptions and using revenue instead of margin. A believable LTV survives this reconciliation.

Once calculated, LTV sets target CAC. Dividing LTV by your desired LTV:CAC ratio gives the maximum profitable acquisition cost.

If LTV is 210 and you want a 3:1 ratio, your CAC ceiling is 70. Bids, budgets, and channel choices all flow from this single derived limit.

Every LTV figure is only as good as its assumptions. Record the time period, margin source, churn definition, and discount rate alongside the number.

This transparency lets teams compare LTV across products and time, and prevents silent input changes from quietly inflating the metric.

Apply the standard subscription LTV formula.

The core formula is ARPU x margin x (1 / churn), kept in consistent time units. Transaction businesses use AOV, frequency, and lifespan instead.

Add a discount rate for long lifetimes, validate against cohort curves, and document every assumption. Probabilistic models extend this to per-customer LTV, which the next lesson uses for segmentation.