The AAARRR Funnel: What It Is and How It Works for Games

AAARRR is a growth framework that breaks the journey into six stages — Awareness, Acquisition, Activation, Retention, Referral, Revenue — each with one metric that gates the next. Its whole value is diagnostic: instead of optimising everything at once, you find the single stage where growth is leaking and fix that.

It's also called Pirate Metrics, because the acronym reads as a growl. Dave McClure's original model had five stages and began at Acquisition. The extra A came later.

The Six Stages at a Glance

The order matters. A leak at stage three makes every improvement at stages four through six mathematically smaller. Fix downward from the top.

Why the Extra A Exists

AARRR begins the moment somebody lands on your store page. That was a reasonable starting line for the SaaS products the model was built for, where demand arrives through search.

Games don't work that way. A player finds your game because a creator they watch played it, because a friend mentioned it, or because a trailer caught them mid-scroll. All of that happens before any measurable "acquisition" event — and if your funnel starts at the install, that entire stage is invisible to you while consuming most of your budget.

Stage by Stage, for Games

Awareness

Measure: creator reach, video views, share of voice against competing titles, branded search volume.

The trap here is counting impressions. Reach is cheap; being remembered is not. A single creator whose audience trusts them moves more installs than ten times the impressions from an audience that doesn't. That's the case for treating creator-led work as a measured growth stage rather than a brand line item.

Acquisition

Measure: installs, CPI, store page conversion rate.

2026 benchmarks: CPI averages around $4.22 on iOS and $2.97 on Android, with iOS consistently 40–60% higher.

The store page is the quietest leak in this stage. Traffic that arrives warm from a creator video and then meets a store listing built for a different audience converts badly, and the loss looks like a CPI problem rather than a creative mismatch.

Activation

Measure: tutorial completion rate, first-session length, time to first meaningful reward.

This is the stage that breaks most often, and the cheapest one to fix. If tutorial completion sits below ~70%, you are paying full price for installs and losing a third of them in the first ninety seconds. No bidding strategy recovers that.

Activation is also where "aha" has to be defined concretely. For a match-3 it might be finishing level five; for a strategy title, completing a first base build. If nobody on the team can name the moment, nobody is optimising for it.

Retention

Measure: D1, D7, D30.

2026 benchmarks: industry medians run roughly D1 22–26%, D7 4–10%, D30 under 4%, and D1 has fallen from about 27% a year earlier. Genre spread is wide — match games retain near 32.7% at D1, strategy near 25.4%.

Read them as three different diagnoses: D1 is your onboarding, D7 is your core loop, D30 is whether you have a live game. Full breakdown in our 13 mobile game KPIs.

Referral

Measure: organic uplift (organic installs ÷ paid installs), K-factor, invite acceptance.

The most neglected stage in mobile games, and the one that decides whether UA compounds or just rents. A channel with a mediocre CPI and a 2x organic halo beats a channel with a great CPI and none — but only if you measure the halo, which last-click attribution structurally cannot.

Revenue

Measure: payer conversion, ARPDAU, ARPPU, LTV against CPI.

2026 benchmarks: payer conversion above 2% is strong, top free-to-play titles reach 3–5%. ARPDAU health floors sit around $0.15 for ad-monetised and $0.30 for IAP-led games. LTV should clear CPI by at least 1.5x to survive platform fees and overhead.

Revenue sits last for a reason. Almost every revenue problem in a free-to-play game is a retention problem wearing a costume.

Where the Funnel Model Breaks

Worth saying plainly, because the diagram is tidier than reality.

Players don't move in one direction. They churn at D3, come back for a seasonal event, play free for four months and then spend forty dollars in a weekend. A funnel implies a single pass; games run loops.

Referral often precedes Revenue. Players recommend games they love long before they pay for them. Ordering Referral above Revenue makes the acronym work, not the behaviour.

The stages share causes. Weak Activation shows up as bad Retention, bad Retention shows up as bad Revenue, and a team that reads only the last number spends months fixing monetisation when the tutorial was the problem.

Use AAARRR as a diagnostic ladder — climb down it until you find the broken rung. Then stop treating what follows Activation as a line and start treating it as a live-ops loop.

How to Actually Use It

One metric per stage. Six numbers, reviewed weekly. If your dashboard has forty, nobody is watching any of them.

Segment before concluding. A blended figure across geos, channels and platforms averages together things that behave nothing alike. The insight is always one level down.

Find the binding constraint. At any moment exactly one stage is limiting growth. Improving the other five produces work, not results.

Measure Awareness and Referral properly or accept you're flying blind on both. They're the two stages last-click attribution cannot see, and the two where creator campaigns do most of their work — which is why we hold ourselves to incrementality across our case studies and performance work.

Six stages, six numbers, one question each. The framework's value isn't the acronym — it's that it forces you to name where you're actually losing people, instead of optimising the stage you happen to enjoy optimising.