Reading Activation Without Vanity Metrics

Login completed is an authentication milestone, not activation. Look for the action that correlates with week-four retention in your own data.

Interview recent retained users. Ask what they finished in the first session that made the app feel useful. Translate those answers into discrete, instrumentable events.

Protect activation from experiment noise. If onboarding changes weekly, keep a stable activation definition and track onboarding variants as separate properties.

Report activation as a rate among newly registered users within a fixed window—commonly seven days—so growth and product share one denominator.