Customer Lifetime Value (LTV): A Practical Guide

Customer lifetime value (LTV) is the total contribution a customer is expected to generate over the time they remain a customer, and it tells you the most you can afford to spend to win them. It is one of the most used and most abused numbers in business, because it combines a margin, a behavior and a forecast, and each one can be made to look better than it is. This guide explains what LTV measures, three ways to calculate it, a cohort example with checked arithmetic, how to discount it, how to use it by segment, and how to avoid inflating it. It is part of unit economics and budgeting and forecasting, follows customer acquisition cost and blended vs. paid CAC, and leads to how to calculate LTV and the LTV to CAC ratio.

What LTV is and is not

LTV is a forecast of the contribution from a customer relationship. The word "contribution" matters: use margin after variable costs, not revenue. A customer who spends $1,000 over their life but whose purchases leave 20 percent margin is worth $200, not $1,000.

LTV is not a fact about a customer. It is an estimate for a group of similar customers, built from past behavior of similar groups. The further it looks ahead, the less reliable it is.

Three ways to calculate LTV

For a subscription, the formula-based method is monthly margin per account divided by monthly churn.

Method 1: historic LTV from a cohort

A hypothetical cohort of 1,000 customers acquired in the same month. Contribution per order is $22 (illustrative) and orders per customer accumulate over time:

Check: 1.25 × 22 = 27.50; 1.50 × 22 = 33.00; 1.85 × 22 = 40.70; 2.40 × 22 = 52.80.

The observed 24-month LTV is $52.80. If CAC is $35, the cohort recovers it between months 6 ($33.00) and 12 ($40.70). Interpolating, payback comes around 6 + (35 - 33) / (40.70 - 33.00) × 6 = 6 + 1.56 = month 7.6. At 24 months, LTV : CAC = 52.80 / 35 = 1.5. That is the real record of this cohort, and no formula should be allowed to replace it without a reason.

Method 2: a formula for subscriptions

A subscription with average revenue per account of $40 a month, gross margin of 70 percent and monthly churn of 4 percent.

• Monthly margin per account = 40 × 0.70 = $28
• Expected lifetime = 1 / 0.04 = 25 months
• LTV = 28 × 25 = $700

Method 3: a formula for repeat purchase

A retailer with average order value of $60, 2.4 orders per customer per year, a 38 percent margin after variable costs and an expected customer lifespan of 3 years.

LTV = 60 × 2.4 × 3 × 0.38 = $164.16. Check: 60 × 2.4 = 144; × 3 = 432; × 0.38 = 164.16.

This works only if each of the averages is itself reliable and if the three-year lifespan can be defended. Treat it as a first estimate, then replace it with a cohort view as soon as you have data.

Discounting

A dollar of contribution in two years is worth less than a dollar today, because of the time value of money and because the future is uncertain. For longer horizons, discount.

Example: the cohort above earns $52.80 - $40.70 = $12.10 of contribution between months 12 and 24, on average about 18 months from the start. At an annual discount rate of 10 percent, the present value is 12.10 / 1.10^1.5 = 12.10 / 1.1537 = $10.49, about 13 percent lower than the nominal figure.

For subscriptions, a quick form is LTV = monthly margin / (monthly churn + monthly discount rate). With a monthly discount rate of 0.8 percent: 28 / (0.04 + 0.008) = 28 / 0.048 = $583.33, compared with $700 undiscounted, a drop of 17 percent. Discounting matters more the longer the lifetime and the higher the rate.

What to use LTV for

LTV by segment

The average hides large differences. Calculate LTV for at least:

• Acquisition channel. Customers from different channels behave differently.
• First product purchased. Entry products often predict repeat purchase.
• Customer size or plan in B2B or subscription.
• Geography where shipping or costs differ.
• Acquisition period. Compare cohorts to see whether quality is changing.

Illustration: suppose two segments each make up half of new customers. Segment X has a 12-month contribution of $60 and Segment Y has $20. The blended figure is $40, but the right CAC limit for X is up to three times that for Y. Spending to the blended limit overpays for Y and underinvests in X.

What inflates LTV

A practical routine

1. Calculate observed LTV for each cohort at 3, 6, 12 and 24 months.
2. Compare cohorts. Are newer ones better or worse at the same age?
3. Add a projection for months beyond the data, with a stated tail assumption and cap.
4. Discount if the horizon is beyond two years.
5. Pair LTV with CAC, payback and cash needs.
6. Review quarterly and after any major change in product, pricing or channel mix. See cohort analysis.

Operations and LTV

Customer lifetime value depends on experience. Orders that arrive late, items that are out of stock and returns that take weeks all shorten the relationship. Consumer product companies improve LTV through reliable availability and delivery as much as through marketing. See demand planning and inventory optimization.

Choosing the horizon

Every LTV figure depends on its horizon. A twelve-month figure is the most defensible for most consumer businesses, because it reflects behavior you have actually observed and is short enough for the company to act on. A 24-month figure suits businesses with longer repeat cycles. Beyond that, LTV is a forecast and should be labeled one. Whatever you choose, use it consistently when comparing channels and cohorts, and use the same horizon for the payback conversation. A common failure is to calculate acquisition limits from a three-year LTV and judge results at twelve months.

Two entry products: a worked comparison

Entry product matters because it often predicts repeat behavior. Suppose a brand has two products customers can buy first, and CAC is $30 for both.

Customers who start with Product X place more repeat orders, and their twelve-month contribution averages $45. Customers who start with Product Y buy less often, and their twelve-month contribution averages $28. The result per customer is 45 - 30 = +$15 for X and 28 - 30 = -$2 for Y. A blended figure of, say, $36.50 for a half and half mix would suggest a profit of $6.50 per customer and hide the fact that the Y customers do not repay. The practical response is to feature Product X in acquisition campaigns, redesign the Y experience to encourage a second purchase, or reduce what you are willing to pay to acquire Y buyers.

Presenting LTV to a board or investor

State five things next to the number: the margin basis, the horizon, the discounting, the customer group and whether it is observed or projected. A clear presentation looks like this: "Twelve-month observed contribution per customer is $40.70 for the January cohort, up from $38.10 for the October cohort. Our 24-month projection is $52.80 with a capped tail." That sentence can be audited. A line saying "LTV is $164" cannot.

How small changes in retention show up in LTV

Retention improvements are valuable because they compound. Suppose that of every 1,000 new customers, 300 place a second order. A program that lifts this to 350 adds 50 second orders. At $22 of contribution per order, that is 50 × 22 = $1,100 more contribution per 1,000 customers, or $1.10 per customer within the first period alone, before any effect on third and later orders. Programs aimed at the second purchase are often the best retention investment, because the second purchase is the strongest predictor of later ones.

FAQs

What is customer lifetime value?

The expected total contribution from a customer over the relationship, used to decide how much to spend acquiring and keeping customers.

What is the formula for LTV?

For subscriptions: monthly margin per account divided by monthly churn. For repeat purchase: average order value × purchase frequency × margin × expected lifespan. Cohort methods use observed data.

Should LTV use revenue or profit?

Use contribution or gross margin after variable costs. Revenue overstates the value.

How long should the LTV horizon be?

Long enough to reflect behavior, short enough to be reliable. Many teams show 12 and 24 months observed and cap projections at three years or fewer.

Why do my LTV numbers differ between teams?

Different definitions of margin, lifespan, discounting and customer segment. Agree on one definition and document it.