Pirate metrics, or AARRR, is a five-stage growth framework covering Acquisition, Activation, Retention, Referral, and Revenue. Coined by Dave McClure, it gives early-stage founders one funnel view of how strangers become paying, returning customers, so you instrument and fix the weakest stage instead of guessing.
Key Takeaways
- AARRR stands for Acquisition, Activation, Retention, Referral, Revenue - the five stages every customer passes through.
- It is a diagnostic funnel, not a dashboard. Its job is to tell you which single stage is leaking worst.
- Most pre-seed and seed startups are convinced they have an acquisition problem when they actually have an activation or retention problem.
- Each stage needs one primary rate metric plus the event instrumentation behind it, or the framework is decorative.
- AARRR complements rather than replaces a north star metric: the north star is where you steer, AARRR is where you look for the leak.
What Are Pirate Metrics (AARRR)?
Pirate metrics are five sequential customer-lifecycle stages, named for the sound of the acronym. Each stage answers one question about the customer journey and carries one dominant rate metric. The framework was designed for early-stage startups precisely because it works without a data team: you can run it on a spreadsheet and a product analytics tool.
| Stage | Question it answers | Primary metric | Typical owner |
|---|---|---|---|
| Acquisition | How do people find us? | New qualified signups or leads per channel | Growth or founder |
| Activation | Do they reach first value? | Activation rate (share hitting the first-value event) | Product |
| Retention | Do they come back? | Week 4 or month 3 cohort retention | Product |
| Referral | Do they bring others? | Invite or referral rate per active user | Growth |
| Revenue | Do they pay, and profitably? | Conversion to paid, ARPA, contribution margin | Founder or finance |
The ordering matters. McClure's original point was that founders obsess over the first stage because it is the most visible, while the compounding value sits in the middle three. A leaky activation step means every dollar of acquisition spend is being poured into a bucket with a hole in it.
How Do You Instrument Each AARRR Stage?
Instrumentation is the difference between a framework and a poster. Work through the stages in this order, defining one event per stage before you add any secondary metric.
- Define the first-value event. Name the single in-product action that means a user got what they came for - first project created, first message sent, first report generated. Everything upstream is acquisition, everything downstream is retention.
- Write the tracking plan before the code. One row per event, with name, trigger, properties, and owner. Our event tracking plan guide covers the naming conventions that keep this from rotting after two sprints.
- Wire acquisition sources correctly. Signup events need channel, campaign, and referrer properties attached at creation, or every downstream cut by channel is guesswork. See conversion tracking setup for startups.
- Build cohorts, not averages. Retention only means something by signup cohort. A blended retention number hides the fact that last month's cohort is worse than the one before it - see cohort retention analysis.
- Instrument referral as an event pair. Invite sent and invite accepted. Without both you cannot tell a sharing problem from a landing-page problem.
- Close the loop to revenue. Join product events to billing so conversion-to-paid and expansion sit in the same view as activation. Pair it with unit economics so growth is not celebrated at negative margin.
Which AARRR Stage Should an Early-Stage Startup Fix First?
Fix the earliest stage with a rate far below its plausible ceiling, and treat retention as the gate on everything else. A practical sequence for a pre-seed or seed team:
- Retention first, if it is flat-lining near zero. No amount of acquisition fixes a product nobody returns to. This is a product-market fit signal, not a marketing task.
- Activation next. Activation is usually the cheapest large win available, because the traffic is already arriving and the fix is onboarding copy, setup friction, or time-to-value.
- Acquisition third. Once a cohort activates and returns, spend is compounding rather than leaking. This is the point where paid channels become defensible.
- Revenue and referral last. Pricing and referral loops amplify a working funnel; they cannot create one.
This ordering is also why startups waste marketing budget: the spend gets approved before the middle of the funnel can hold water.
How Does AARRR Differ from a North Star Metric?
AARRR is a diagnostic map with five numbers; a north star metric is a single steering number for the whole company. They answer different questions, and healthy teams run both. The north star tells you whether you are winning this quarter. AARRR tells you which stage to send an engineer at on Monday.
In practice the north star usually lives inside the retention or revenue stage - weekly active teams, weekly value events delivered - and the AARRR stages become its input metrics. If you are choosing yours now, start with north star metric selection, then map the AARRR stages underneath it. For which numbers to report to a board rather than to your team, see growth marketing metrics that matter.
What Does an AARRR Dashboard Look Like for a Seed-Stage Startup?
Five tiles, one per stage, each showing a rate and its trend against the prior period, plus a channel breakdown on acquisition and a cohort grid on retention. That is the whole thing. Resist adding a sixth tile until the first five have owners.
Keep two rules. First, every tile shows a rate, not a raw count, because raw counts rise with spend and hide efficiency decay. Second, every tile has one named owner who reports on it in the weekly growth meeting. Our startup growth metrics dashboard guide covers layout, and growth KPIs by startup stage covers which targets are reasonable at pre-seed versus Series A.
Where Do Founders Get AARRR Wrong?
- Treating it as a reporting template. The output of an AARRR review is a decision about the next two weeks, not a slide.
- Measuring stages with blended numbers. Blended funnels average a great channel and a terrible one into a mediocre story.
- Defining activation as signup. If activation is measured at account creation, the framework loses the exact stage it exists to expose.
- Building referral loops pre-retention. Referral of a product people churn out of accelerates churn.
- Reporting AARRR to investors as traction. Investors want revenue, retention, and efficiency; keep the diagnostic internal and show outcomes externally, as in how to show traction to investors.
How Do You Run a Weekly AARRR Review Without a Data Team?
Book 30 minutes, pull the five rates, and answer three questions in order: which stage moved, why, and what one change we are shipping this week. A two-person team can do this in a spreadsheet fed by a product analytics export; you do not need a warehouse until the review is a habit.
The discipline that makes it work is a single owner per stage and a written decision at the end. Without a decision, the review degrades into number-reading. Teams that formalise this early tend to find their first real growth lever in weeks rather than quarters, because the review forces a comparison between stages instead of a defence of one channel. If the meeting keeps stalling on data trust, that is a tracking problem, not a metrics problem, and it belongs in the tracking plan before the next review.
How Does AARRR Change Between Pre-Seed and Series A?
The stages stay the same; the questions get harder. At pre-seed you are proving that activation and retention exist at all, usually with dozens of users and qualitative follow-ups rather than statistical confidence. At seed you are looking for one repeatable acquisition channel that feeds a funnel already known to hold. By Series A the same five stages get segment cuts - by plan, by ICP, by channel - and revenue-stage efficiency metrics like payback period become the constraint on how fast you can spend.
What changes most is tolerance for ambiguity. Early on, a directional read from 40 users is enough to justify shipping an onboarding fix. Later, the same decision needs a segment cut and a cost figure attached, because spend is larger and mistakes compound faster. Founders moving through this transition should also revisit the pre-seed to Series A marketing playbook so the funnel work and the fundraising narrative stay aligned.
Frequently Asked Questions
What Does AARRR Stand For?
AARRR stands for Acquisition, Activation, Retention, Referral, and Revenue. These are the five stages of the pirate metrics framework, each tracking one step in how a stranger becomes a paying, returning customer.
Who Created Pirate Metrics?
Investor Dave McClure introduced the AARRR framework for early-stage startups. The nickname comes from the acronym sounding like a pirate, and it stuck because the model is simple enough to run without a data team.
Is AARRR Still Relevant for AI and Product-Led Startups?
Yes, though the stage definitions shift. Product-led and AI-native products often see activation and retention decide the outcome, so those two stages carry more weight than acquisition, and cost-to-serve belongs inside the revenue stage.
How Many Metrics Should Each AARRR Stage Have?
One primary rate metric per stage, at least until the funnel is stable. Adding secondary metrics before each stage has a named owner and a clean event definition tends to produce dashboards nobody acts on.
Should AARRR Replace Our North Star Metric?
No. Use the north star metric to steer the company and AARRR to diagnose which stage is leaking. The north star usually sits in the retention or revenue stage, with the other AARRR stages acting as its input metrics.