AARRR Funnel: Acquisition, Activation, Retention, Referral and Revenue
Define the event. Name the denominator. Ask a better question.
One cohort. Five questions.
Build your measurement tableAcquisition
1,000First-time qualified visitors
Activation
400Send their first invoice
40% of acquisition usersRetention
220Send another invoice
55% of activation usersTwo independent outcomes among retained users
Referral
33Send a referral invitation
15% of retention usersRevenue
88Start a paid plan
40% of retention usersOn this page
The AARRR funnel is a growth-measurement framework that organizes the customer lifecycle into five areas: Acquisition, Activation, Retention, Referral and Revenue. It helps a business replace a long dashboard of disconnected numbers with a small set of measurable behaviors.
Also known as Pirate Metrics, the model was introduced by Dave McClure in 2007 in his Startup Metrics for Pirates presentations. The framework is especially associated with startups and digital products, but its basic logic is broader: define the important behaviors in a customer journey, measure them consistently and investigate where progress is weaker than expected.
What is the AARRR funnel?
AARRR asks five basic questions:
- Acquisition: How do people reach us?
- Activation: Do new users experience meaningful value?
- Retention: Do they come back and continue using the product?
- Referral: Do satisfied users bring or recommend other users?
- Revenue: Does the relationship generate money?
The framework does not prescribe one universal metric for each stage. A useful AARRR model begins by defining one observable event for each stage in your specific business.
Why does it matter?
Growth can look healthy at the top while failing further down. A campaign may produce thousands of visits, but those visits matter less if few users reach value, return or pay.
AARRR helps a team:
- separate traffic from meaningful usage;
- see where users stop progressing;
- avoid treating total registrations or page views as sufficient evidence of growth;
- connect marketing activity with product behavior and monetization; and
- choose a specific problem to investigate next.
The five stages
Acquisition
Acquisition records how new users arrive. Depending on the business, the event might be a qualified website visit, app install, store visit, lead or account creation.
Track the source as well as the volume. Search, referrals, paid campaigns and partnerships can produce users with very different behavior after arrival. The Bullseye Framework can help choose which traction channels to test; AARRR measures what happens across the lifecycle once users start arriving.
Activation
Activation is the first behavior that shows the user has experienced meaningful value. It should be more specific than simply opening a page or creating an account.
For an invoicing product, activation might be creating and sending the first invoice. For a booking service, it might be completing the first booking. The correct event depends on the value the product is supposed to deliver.
Retention
Retention asks whether activated users return and repeat a valuable behavior after a defined period.
The period matters. "Returned" could mean the next day, within a week or during the next billing cycle. Choose a window that matches normal usage and keep it consistent when comparing cohorts.
Referral
Referral measures whether existing users help bring in other users. A measurable event might be sending an invitation, sharing a referral link or generating a successful referred signup.
Do not confuse sharing with successful referral. Ten invitations and ten new referred users are different outcomes.
Revenue
Revenue measures monetization behavior: a purchase, paid subscription, transaction fee, upgrade or another event that generates income.
Revenue volume alone does not show whether growth is economically attractive. Unit economics provides the next layer: compare customer value with the costs of acquiring and serving customers.
How to calculate stage conversion
For a simple funnel step, use:
Stage conversion rate = users who reach the stage / eligible users in the previous group x 100
The denominator must be explicit. Referral and Revenue are not always sequential in real life, so do not force every business into one rigid path. Measure each rate against the relevant cohort.
A simple hypothetical example
Hypothetical example: A small subscription invoicing product reviews 1,000 first-time visitors acquired during April. Before looking at the results, the team defines these events and internal targets:
| Stage | Defined event | Result | Rate | Internal target |
|---|---|---|---|---|
| Acquisition | First-time qualified visitor | 1,000 | Not applicable | Not applicable |
| Activation | Sends first invoice within 24 hours | 400 | 40% of acquired | 50% |
| Retention | Sends another invoice within 30 days | 220 | 55% of activated | 55% |
| Referral | Retained user sends a referral invitation | 33 | 15% of retained | 10% |
| Revenue | Retained user starts a paid plan | 88 | 40% of retained | 30% |
The example does not specify a final observation cutoff or the Referral and Revenue windows. Confirm these definitions and allow the cohort enough time to complete them before interpreting real data.
The table gives the team a specific diagnosis. Retention, Referral and Revenue meet or exceed the hypothetical internal targets. Activation does not: only 40% of acquired users reach the first-value event against a 50% target.
That does not prove what is wrong. It tells the team where to investigate. The next question could be: Are visitors poorly matched to the offer, or are suitable users failing to complete the first-invoice flow?
Buying more traffic before answering that question could simply send more people into the same weak activation step.
When should you use AARRR?
Use AARRR when you can observe repeated customer behaviors across acquisition, first value, continued use and monetization. It is particularly useful for subscription products, apps, online services, marketplaces and other businesses with measurable user journeys.
For a long enterprise sale, a sales pipeline may be better for managing individual deals. AARRR can still provide a higher-level growth view, but it does not replace deal management.
How should you interpret the result?
Do not automatically optimize the smallest percentage. A 10% referral rate and a 40% activation rate measure different behaviors and may have different targets.
Compare each stage with:
- its own historical performance;
- a target defined before reviewing the result;
- differences between acquisition channels or customer segments; and
- the economic importance of improving that stage.
AARRR identifies where to investigate. It does not prove why a metric changed.
Common confusion: AARRR and AIDA
AARRR measures behaviors across a customer lifecycle. The AIDA Model helps structure one persuasive message through Attention, Interest, Desire and Action.
A message built with AIDA could support Acquisition or Activation, but AIDA is not a growth analytics framework and AARRR is not a copywriting formula.
Common mistakes
- Defining Activation as signup. Choose an event that represents first meaningful value.
- Mixing cohorts. Do not compare this month's acquisition with last year's retention without a clear reason.
- Using unclear denominators. Every percentage should state which users were eligible to reach the next event.
- Treating the model as strictly linear. Users may pay before referring, refer before paying or return many times.
- Optimizing Acquisition first by default. More traffic does not repair weak Activation or Retention.
- Using universal benchmarks as verdicts. Your product, market, price and measurement definitions affect what a useful rate looks like.
Limitations
AARRR simplifies a customer journey that may contain several channels, devices, stakeholders and repeated loops. It shows patterns in behavior but does not establish causation or explain customer motivation. Interviews, experiments, cohort analysis and financial analysis may be needed before choosing an action.
Quick checklist
Put it to work
Build a first AARRR table in under 30 minutes
Create six columns: Stage, Event, Count, Denominator, Rate, Next question.
- Define one event for each AARRR stage.
- Choose one recent customer cohort and one appropriate time window.
- Add the observed count for every event.
- Calculate each relevant conversion rate and write the denominator beside it.
- Compare the rates with previous performance or pre-defined internal targets.
- Select one stage that needs investigation and write one concrete question about it.
Do not finish with "improve growth." Finish with a question that can be tested with additional evidence.
Eligible users: Unknown. Count users who meet the event at least once, not repeated events. Blank counts stay unknown.
Key takeaways
- AARRR stands for Acquisition, Activation, Retention, Referral and Revenue.
- Each stage needs an observable business-specific event, not a vague label.
- Conversion rates are useful only when cohorts, time windows and denominators are clear.
- The lowest percentage is not automatically the biggest problem.
- AARRR identifies where to investigate; it does not explain the cause by itself.
Final takeaway
The AARRR funnel gives a business a compact view of how people arrive, experience value, return, recommend and generate revenue. Its usefulness comes from precise definitions and consistent measurement, not from the five labels alone.
Start with one cohort, define the events, calculate the rates and use the result to ask a better next question.