✨Activation — The Ultimate Guide
The single highest-leverage point in your funnel. Activation done right makes everything downstream easier.
Improving activation by 10 points often compounds into 30-50% revenue growth, because every downstream metric (retention, monetization, expansion) is anchored to it. PMs who master activation move company-level metrics with team-level investments.
Activation is the moment a user hits the 'aha' — they experience the product's value clearly enough to commit to using it. The PM job: identify the specific behavior that defines activation, instrument it, redesign onboarding to maximize the path to it, and measure relentlessly.
Defining the activation event
A good activation event is:
- A specific user behavior (not 'they signed up')
- Highly correlated with long-term retention
- Achievable in the first session ideally
Examples:
- Slack: send 2,000 messages in a workspace (org-level activation)
- Dropbox: install client + add at least one file
- Notion: create at least 3 pages in week 1
- Tinder: send 3 messages (not just swipes)
Finding your activation event
- Cohort analysis. Look at retained vs lapsed users. What behavior do retained users have in week 1 that lapsed users don't? That's your candidate.
- Test for causation. Correlation isn't enough. Run an experiment that nudges users toward the behavior; if retention improves, the behavior was causal.
- Pick the lowest-effort, highest-correlation behavior as your activation event.
Onboarding redesigned around activation
Once you have the event, redesign onboarding:
- Strip everything not necessary to get to the activation event.
- Make the path obvious. Tooltips, progress indicators, default values.
- Pre-populate context. Use signup data, OAuth scope, social graph to skip steps.
- Provide quick wins. Templates, samples, demo content.
Measuring time-to-value (TTV)
The single most useful activation metric. Median minutes from signup to activation event. Aim to cut TTV by 50% each year.
The activation curve
Plot activation rate by day-since-signup. Most products see:
- 30-50% activated by end of session 1
- 60-70% by day 7 (if at all)
- Past day 7, very few new users activate
If your day-7 activation rate is below 40%, you have a huge opportunity. If it's above 70%, focus your investment elsewhere.
Edge cases
- Two-sided marketplaces. Activation is per side. Often different events (host activation = first booking received; guest activation = first booking made).
- B2B team products. Activation can be individual or team-level. Team-level is harder but predicts retention much better.
- AI products. Activation often = first 'wow' moment (the user sees the AI produce something useful). Different from traditional SaaS.
Real-world examples
Slack discovered that workspaces that sent 2,000 messages had ~90% retention vs much lower below. The entire onboarding was redesigned to drive workspaces past this threshold — channels created automatically, invites prompted aggressively, onboarding tour around early team use cases.
Go deeper — recommended reading
Interview questions (1)
Q1How would you identify the activation event for a product you just joined?executionsenior▼
Three-step process:
- Cohort analysis on retained vs lapsed users. Pull data on users who were still active at day 30 and users who churned. Compare their behaviors in week 1. Look for behaviors that retained users do at much higher rates.
- Generate candidate activation events. Often 3-5 candidates emerge — e.g., 'created first project,' 'invited a teammate,' 'completed setup wizard.'
- Test for causation. Correlation isn't enough — maybe retained users do behavior X because they were already going to retain, not because X causes retention. Run a small experiment: nudge a treatment group toward behavior X (via onboarding or email). If treatment retention exceeds control, you've found a causal activation event.
Pick the candidate with the highest causal effect AND the lowest user effort. Then redesign onboarding around it.
Common mistake: declaring activation based on correlation alone. Most teams' first 'activation event' is wrong; the causal test corrects it.