When Activation Is Not Login

Treating first login as activation flatters onboarding charts and hides weak product value. Look for the earliest action that correlates with retained cohorts in your own data.

Interview users who stayed past week four. Ask what they finished in the first session that made the app feel useful, then translate those answers into discrete events you can verify in staging.

Hold activation definitions stable while onboarding experiments change. Track experiment variants as properties so the activation rate stays comparable.

Report activation among newly registered users inside a fixed window—often seven days—so growth and product share one denominator.