Skip to content
Back to Growth & Acquisition

AARRR Funnel (Pirate Metrics)

Model the user journey as Acquisition → Activation → Retention → Referral → Revenue and instrument each stage's conversion rate.

Why

AARRR gives a shared vocabulary for the five stages every user passes through, so a growth team can say 'retention dropped 3%' and everyone knows exactly which funnel step and which dashboard. It exposes the bottleneck: if 100k sign up but 2k activate, no amount of ad spend fixes the leak — you must fix activation first. Each stage has different owners, tools, and experiments, so naming the stage tells you who acts and with what. Without a funnel model, teams chase the metric closest to their comfort zone (engineers fix retention, marketers buy ads) and the actual bottleneck goes unaddressed for quarters.

When

When you have a working product and need to identify which stage leaks the most users. Build the funnel once, instrument it permanently.

How

1. Define the concrete event for each stage: Acquisition = first visit; Activation = performed core action within 24h; Retention = returned in week 2; Referral = invited someone; Revenue = paid
2. Instrument every stage event in your analytics tool with a user_id and timestamp
3. Build a funnel report: count and conversion rate for each stage transition
4. Identify the weakest transition (lowest conversion rate) — that is your current bottleneck
5. Form a hypothesis for why users drop at that stage (interview 10 who dropped)
6. Run 3-5 experiments targeting only that bottleneck stage
7. Re-measure the funnel weekly; when the bottleneck moves, move the team with it
8. Report the full funnel in every growth review so the team sees the whole system, not one stage

Example

# AARRR funnel for a SaaS note app (weekly cohort)
Acquisition:  10,000 signups        (100%)
Activation:    2,500 did core action (25%)  ← bottleneck
Retention:     1,750 returned W2     (70% of activated)
Referral:        350 invited someone  (20% of retained)
Revenue:         105 paid             (30% of referrers)

# Weakest transition: Acquisition → Activation (25%)
# Action: move 2 engineers + 1 designer to onboarding experiments
# Target: raise activation to 35% →  +1,000 activated users/week

Common Problems

  • ⚠Defining 'activation' as 'logged in' instead of 'performed the core action' — a login that doesn't experience value predicts churn, not retention.
  • ⚠Measuring the funnel once and never again — funnels drift as the product changes, and a stale funnel misdirects the whole team.
  • ⚠Attacking the stage you're comfortable with instead of the actual bottleneck — engineers over-invest in retention while activation leaks.
  • ⚠Using different definitions of 'active user' across teams, so the funnel numbers never reconcile and debates replace action.

Best Practices

  • ✓Define each stage as a concrete user-performed event, not a time-window label — 'created first document within 24h' beats 'activated user'.
  • ✓Always attack the lowest-conversion transition first; fixing a non-bottleneck stage does not move the final number.
  • ✓Re-measure the full funnel every week and move the team to wherever the bottleneck has moved; funnels are dynamic.
  • ✓Use one shared definition of each stage across all teams so the numbers reconcile and debate ends.