See predicted 12-month LTV winners in Experiments

Experiments with a revenue primary metric now show a predicted long-term winner. When an experiment has enough conversion data, RevenueCat calculates the expected 12-month LTV for each variant and surfaces which one is predicted to generate more revenue over time.

The prediction appears alongside your standard experiment results. It uses each variant's observed conversion and retention data to model future revenue. It shows automatically once the guardrails are met. No configuration needed.

This helps when short-term conversion rates are close and you're not sure which variant to ship. Instead of waiting months for revenue to fully accumulate, you get a forward-looking signal while the experiment is still running.