Customer lifetime value

Customer lifetime value (LTV, sometimes CLV or CLTV) is the total revenue a customer generates over their entire relationship with the business. For a one-off purchase product it’s the AOV plus whatever repeat purchases happen. For subscription it’s the monthly revenue multiplied by expected retention. For a B2B contract it’s the annual value times retention plus any expansion.

The single most useful derived number is LTV:CAC, the ratio of lifetime value to customer acquisition cost. 3:1 is the rough industry benchmark for a healthy business - you spend £1 to acquire a customer who generates £3 over their relationship. Below 1:1 you’re paying to acquire customers. Above 5:1 you’re probably underspending on acquisition and could grow faster.

Why this matters for CRO

CRO prioritisation usually focuses on first-order conversion rate. Lift conversion 10%, revenue goes up 10%, ship it. That ignores which customers you’re winning. Two variants might lift conversion equally but bring in customers with very different LTVs.

Examples:

• A discount-heavy landing page converts 15% better but the discounted customers churn faster and never buy at full price. Net LTV is lower despite the conversion lift.
• A subscription-first checkout flow converts worse than one-off purchase, but the customers who do convert have 3x the LTV. Worth shipping even at lower conversion.
• A free-trial flow on a SaaS product converts 4x more sign-ups than a credit-card-required trial, but those sign-ups convert to paid at half the rate. Same paid customer volume, different funnels.

The honest version of CRO accounts for downstream LTV, not just immediate conversion. Most programmes don’t because the LTV signal takes months to develop and the conversion signal is in the test report next week.

Calculating it

Two main approaches:

• Cohort-based - track actual revenue per customer cohort over time and project forward. More accurate, requires data and patience.
• Predictive - regression or machine-learning models that predict LTV from early signals (first-purchase value, source, product mix). Faster but only as good as the model.

For early-stage businesses, neither works well because you don’t have enough history. The honest answer there is to use a rough estimate (industry benchmark for your category) and revisit as you gather data.

Things people get wrong

• Using LTV as a single number when it varies massively by segment. LTV from paid social is usually much lower than LTV from organic. Averaging hides the variance.
• Conflating LTV with retention. LTV is total revenue. A customer who buys once at £200 has higher LTV than one who buys monthly at £10 for 18 months, but the latter is “retained” longer.
• Optimising for the offer that maximises first-order conversion without checking LTV impact. Discounting always wins short-term and often loses long-term.
• Ignoring how LTV interacts with the value ladder. The whole point of a ladder is to maximise LTV, so you should be testing against ascension rates, not just entry conversion.