Activation metric

Activation is the percentage of new users who reach a defined first-value moment within a set time window. It sits at stage two of the AARRR framework, between acquisition and retention. The most common mistake is treating “completed onboarding” as activation. A user can finish every onboarding step and still churn in week one because the product never solved their problem. Interviewers hear this conflation constantly and it is the fastest way to signal a surface-level answer.

How to derive the metric from scratch

You do not guess the activation event. The correct method:

1. List 5-10 candidate events that plausibly represent first value delivery. For a project management tool: first project created, first task assigned, first teammate invited.
2. Run a cohort query for each event. Compare Day-30 retention for users who completed each event vs. those who did not, within your activation window.
3. Pick the event with the strongest retention delta. That is your activation event.
4. Validate with a holdout before declaring it canonical. Correlation runs in both directions.

The examples you already know are precise because of this method. Slack’s threshold is teams that sent 2,000 total messages together, which correlated with 93% likelihood of continued use. Not “sent a first message”: the volume threshold. Facebook’s is 7 friends connected within 10 days. Twitter’s is following 30 accounts. Pinterest’s is saving content weekly for 4 consecutive weeks post-signup. None were guesses. PostHog’s session replay activation is compound: 5 recordings analyzed AND 1 recording list filter changed within 14 days. Their surveys product is simpler: 1 survey launched within 30 days, because value is more immediate.

SaaS baseline: 30-40% activation rate for self-serve signups in 2026. Day-30 retention drops roughly 40% when users skip early steps like team invitations. Around 80-90% of users churn if they do not experience product value in the first week.

Activation windows vary by product type

The window is part of the definition. Consumer apps: 24-72 hours. PLG B2B: 7-14 days, because collaboration features require teammates. Analytics and enterprise tools: up to 30 days. Using the wrong window misreports your rate and pushes you toward the wrong interventions.

Leading vs lagging events

A leading activation event predicts retention. A lagging one tracks what already-retained users did. “Account fully configured” often describes users who would have stayed anyway. If removing the event from your onboarding flow would not change retention, it is lagging, not a signal.

AI-native products in 2026

For AI-native products, the activation event is often experiential rather than behavioral. Perplexity’s activation is tied to model quality: each model upgrade can re-trigger the aha moment, not just initial onboarding. You cannot directly observe a user feeling surprised by output quality, so you instrument proxies: copy events, shares, return visits within 48 hours of first use.

Two deliberate activation designs worth knowing: Viggle (AI video) required structured onboarding before granting full access, forcing the aha moment before unbounded use. Lapse (disposable camera app) required inviting 5 or more friends before use, embedding the network effect into activation itself. Teams replacing generic onboarding intake with AI customer interviews are seeing 2-4x activation gains.

Activation is the first honest test of lovability. With feasibility nearly free and basic usability having a high floor, the aha moment is no longer about completing a flow. It is about whether the product did something genuinely valuable enough that the user chose to return. The activation metric is how you track whether you kept that promise.

Weak vs strong interview answer

For how activation fits into the full growth model, see AARRR and cohort analysis.