CLTV (Customer Lifetime Value)

Estimation of the total value that a customer is expected to generate for the company over the entire commercial relationship, taking into account purchase frequency, average basket, retention time and margin. In a CRO logic, CLV makes it possible to assess not only the immediate impact of an action, but also its long-term profitability.

Relevance in A/B testing :

CLV is an advanced KPI, to be used when the hypothesis being tested is likely to influence future customer behavior, such as :

• anupsell or cross-sell test (e.g. offering a premium service),
• modification of the customer loyalty path (e.g. onboarding, post-purchase relaunch),
• a new pricing logic (e.g., packaged offers, subscriptions),
• adjusting thepost-purchase experience to encourage repeat purchases.

💡 If the experiment tested only impacts initial conversion with no expected effect on retention, it is irrelevant to rely on CLV as the primary endpoint.

Example: GA4 and probabilistic CLV

Google Analytics 4 offers a native probabilistic CLV model, based on user behavior within a defined time window (e.g. 90 days). This model enables :

• compare the value generated and expected according to source, channel, or exposure to a test variation,
• initiate CLV analysis even without CRM data, to identify segments with high potential value.

This is an accessible first approach to integrate into post-test analysis, notably via BigQuery to cross-reference test data with future conversions.

Limits :

• The reliability of CLV is limited in login-free environments, as cookies gradually disappear, making long-term tracking more difficult.
• The observation window must be consistent with the customer's life cycle (e.g. 90, 180 or 365 days, depending on the business model).
• An isolated CLV reading can mask a deterioration in conversion or user experience if not coupled with other KPIs.

‍