You're evaluating whether sending a coupon to dormant users brings them back profitably. The treatment may increase completed orders, but it also adds coupon cost and could pull forward behavior from users who would have returned anyway.
How would you design an experiment to test whether offering a coupon to dormant users is profitable?
Choosing the right primary metric for an experimentMeasuring incremental profit rather than engagement aloneSeparating causal lift from subsidized baseline behaviorUsing guardrails to avoid harmful launches