Framework
Term

Activation

The point a new user has experienced enough of the product's core value that they're likely to return — usually defined by a specific in-product behavior.

Activation is the moment a new user crosses from "signed up" to "got value". It's distinct from sign-up (which is just creating an account) and from retention (which is whether they come back). Activation is the bridge.

Canonical examples

  • Slack: a team sends 2,000+ messages (then retention jumps sharply)
  • Facebook (historical): friend 7+ people in 10 days
  • Twitter (historical): follow 30+ accounts
  • Notion: create 3+ pages in the first session
  • Stripe: process a first live payment

These are not arbitrary. Each was discovered by analyzing cohorts: among users who did action X in the first N days, retention was meaningfully higher. The activation event is the earliest predictor of long-term retention.

Why it matters

Most products have a leaky funnel between sign-up and retention. Activation defines where the leak actually happens. Without an activation metric, the team treats sign-up growth as success — but sign-ups that don't activate are wasted acquisition spend.

The right thing to optimize for is percent of new users who activate within N days — that single metric concentrates the team on the bridge that determines retention.

Finding your activation metric

Standard method:

  1. Pull cohorts of users from 60–90 days ago
  2. Compare those who retained (still active week 4) to those who didn't
  3. Find the in-product behavior in week 1 that most distinguishes the two groups
  4. That behavior is your candidate activation event
  5. Validate by checking other cohorts and forward-testing

This is the work most product teams skip. Without it, "improve onboarding" is unfocused and produces little.

Related

  • AHA moment — the qualitative cousin: when the user feels the value (activation is the operational proxy)
  • PMF — pre-PMF, activation rates are usually unreliable
  • Retention cohort

Nearby terms

All terms →