AARRR funnel (pirate metrics)

Also called: AARRR, pirate metrics, AARRR framework, AAARRR, growth funnel stages

A five-stage framework for the customer lifecycle: Acquisition, Activation, Retention, Referral and Revenue, with one or two metrics per stage to find where growth leaks.

AARRR, nicknamed pirate metrics because the acronym sounds like a pirate's growl, splits the customer lifecycle into five stages and asks for one or two metrics at each:

Stage	Question	Typical metric
Acquisition	Do people find us and sign up?	New sign-ups per channel, cost per sign-up
Activation	Do they reach the first real value?	Share of new users who hit the activation action within N days
Retention	Do they come back?	Day-30 retention, weekly retention by cohort
Referral	Do they bring others?	Share of users whose invite was accepted, viral coefficient
Revenue	Do they pay?	Conversion to paid, ARPU, LTV

Its value is as a checklist and a shared vocabulary: it forces a team to define each stage concretely for its product and to measure all five, instead of staring at acquisition because that's where the budget goes. Read as a funnel on one signup cohort, it shows the stage with the biggest leak, which is usually where the next investment belongs.

Some caveats matter in practice. The stages aren't strictly sequential: in a collaborative product, inviting others is part of activation, and people can pay before they refer anyone (some versions put Revenue before Referral, and some add Awareness at the start, AAARRR). Retention is the base: pouring acquisition into a product that doesn't retain only buys churn. And each stage needs a precise definition (action, window, denominator); "activation" means nothing until you write it down.

Example

Halves reads AARRR on its June signup cohort (20,000 new users). Each stage is defined in the tracking plan:

Stage	Definition	Users	Share of cohort	Share of previous stage
Acquisition	Signed up in June (staff excluded)	20,000	100%	—
Activation	In a group with at least one other member and added a first expense within 7 days	8,000	8,000 / 20,000 = 40%	40%
Retention	Activated and active in their 4th week	3,200	3,200 / 20,000 = 16%	3,200 / 8,000 = 40%
Referral	Invited someone who joined within 30 days	2,400	2,400 / 20,000 = 12%	— (not a subset)
Revenue	Started paying for Halves Plus within 30 days	400	400 / 20,000 = 2%	400 / 8,000 = 5% of activated

Reading it:

The biggest leak is activation: 12,000 of 20,000 new users (60%) never share a single expense with anyone.
Referral isn't a later stage here: for a group app, inviting a flatmate is usually how people activate, so the referral and activation numbers overlap.
Revenue per cohort is small in absolute terms, but 5% of activated users pay within a month, so each extra activated user is worth far more than each extra sign-up.

Maya's call: the next quarter goes to activation (making the first invite and first shared expense easier), not to more acquisition spend.

Common mistakes

Undefined stages. "Activation" and "retention" need an exact action, window and denominator for your product before the numbers mean anything.
Treating it as a strict sequence. In collaborative products referral is part of activation, and revenue can come before referral; read overlaps, don't force a line.
Mixing cohorts across stages. June's sign-ups divided by May's payers is not a conversion rate; follow one cohort through every stage.
Optimizing acquisition first. If activation or retention leaks, more sign-ups mostly buy churn.
One vanity metric per stage. "Invitations sent out" or "app opens" can fill the table without saying anything; pick actionable metrics.