CLTV (Customer Lifetime Value)

Estimate of the total value a customer is expected to generate for the company over the entire business relationship, taking into account purchase frequency, average order value, retention duration, and margin. Within a CRO approach, CLV helps evaluate 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 tested hypothesis is likely to influence the customer's future behavior, such as:

• an upsell or cross-sell test (e.g., offering a premium service),
• a change to the loyalty journey (e.g., onboarding, post-purchase follow-up),
• a new pricing logic (e.g., bundled offers, subscription),
• an adjustment to the post-purchase experience likely to encourage repeat purchases.

💡 If the tested experience only impacts initial conversion with no expected effect on retention, it is not relevant to rely on CLV as the main evaluation criterion.

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 makes it possible to:

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

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

Limitations:

• The reliability of CLV is limited in login-free environments, due to the gradual disappearance of cookies, making long-term tracking more difficult.
• The observation window must be consistent with the customer lifecycle (e.g., 90, 180 or 365 days depending on the business model).
• An isolated reading of CLV can mask a decline in conversion or user experience if it is not combined with other KPIs.