AARRR: how to run the five stages in practice

“Aarrr, aarr!” Perhaps you have seen these interjections in an adventure film whose central theme is pirates. Inside the marketing universe they took on a very real connotation: they stand for five conversion metrics guided by the analysis of digital user behavior.

AARRR is one of the few models that survived more than a decade without losing usefulness, and the reason is simple: it forces you to separate numbers that usually travel together in the report. Here you will see what each letter measures, how to tie that to conversion rate and where to start when the operation is already running.

What is the AARRR framework?

The AARRR framework is a model of five metrics that follows the user life cycle: Acquisition, Activation, Retention, Revenue and Referral. Each stage has its own number, and the useful reading lies in comparing them, not in any single value.

Caption: AARRR stops being an acronym when each stage has a number of its own and a person responsible for it

The difference from a traditional marketing funnel is that AARRR does not end at the sale. It keeps measuring afterwards, which changes the growth conversation: recurring revenue and referral enter the account, not only lead conversion at the entrance.

That is why the AARRR pirate funnel tends to be adopted by operations that already have volume and want to understand why volume is not turning into results. It does not grow the top; it shows where the top is leaking.

Seen from a distance, AARRR is a marketing funnel with memory: it remembers what happened after the conversion. That is why it serves both teams measuring qualified leads and teams measuring contract renewal.

How did the AARRR pirate funnel come about?

The model was created by Dave McClure, co-founder of the accelerator 500 Startups, today 500 Global. He realized that many startups failed because they focused their attention on superficial metrics that did not necessarily provide information that could convert into results for their businesses.

And convert, in this case, is the keyword behind the emergence of this model. It is an architecture designed to optimize conversion rate, and it works very well alongside a solid CRO strategy.

McClure established a new way of looking at the funnel, or flywheel according to current parameters, using just five metrics based on user behavior and their life cycle as a consumer. The term flywheel gained traction from 2018 as an alternative to the linear funnel, emphasizing momentum and customer-driven growth, and it helps align marketing, sales and customer success under a continuous cycle.

One spelling is worth correcting, because it circulates widely: the acronym has four Rs, AARRR, one letter for each stage after acquisition. The three-A version does not exist in the original model.

What does each AARRR stage measure?

As with other marketing theories broken into acronyms, such as Kotler's 4 Ps, McClure divided his model following the same playful and functional pattern: five letters, one for each phase.

A for Acquisition

The first type of measurement corresponds to the acquisition channels used to capture leads or, later, customers. The question here is: how are users getting to my website? Content marketing strategies, such as landing pages, newsletters and blog posts, are essential, as are social media, SEO and SEM.

• Actions that perform best
• Actions that generate the greatest volume of traffic
• Lower cost actions

A for Activation

At this stage, some user behaviors related to the quality of their experience on your website are measured. The main observation to make is: are these visitors taking the actions I want within my website? The focus should be on giving the user a pleasant first experience as an initial way to influence their purchasing decision and to ease onboarding.

• Time spent and navigation depth
• Number of clicks on links and buttons
• Newsletter subscriptions and material downloads

R for Retention

Once you have managed to attract visitors and give them a good first experience, it is time for the next step: making them keep their interest in your product or service and interact with your offers again. For that, the oldest and most traditional digital relationship strategy still holds up: email marketing strategies. When you acquire the lead's contact, make sure they enter an automation flow that nurtures them with content that keeps them warm enough to move along the buying journey. There is no universal cadence: define it by objective, by user behavior and by A/B testing.

R for Revenue

No strategy can be considered efficient when there is no return. The focus of this phase is to measure whether users have generated revenue for your company, and the way to do it depends on your business model. One way to measure it is the LTV/CAC ratio (Lifetime Value over customer acquisition cost), tracked over time rather than in an isolated month.

R for Referral

There are certain strategies we can call incomparable, and the biggest example is referral marketing, in which a user becomes an advocate for your business and makes a point of recommending it to other consumers. Word of mouth remains the oldest and most efficient way of marketing products and services, with social proof as its flagship. Here it is worth monitoring backlinks, mentions on social networks, reviews and testimonials. To measure the real propensity to recommend, track the Net Promoter Score (NPS).

How does AARRR connect to CRO and conversion rate?

AARRR provides the ruler and CRO provides the method. Without the five stages separated, any test measures the wrong thing: improving the form fixes activation, not acquisition, and neither of them fixes retention.

In practice, CRO marketing works on the conversion rate from one stage to the next, and the pirate funnel says which one deserves the next test. The priority rule is to look at the largest relative drop, not the smallest absolute number.

A marketing strategy that invests in acquisition while activation is low buys an audience that does not stay. It is the most expensive way to grow, and the most common.

That fit also changes what you expect from content. A content strategy designed stage by stage feeds acquisition and activation at once, and a mature content marketing strategy still supports retention, because it gives the contact a reason to come back.

In a CRO marketing routine this becomes an agenda: one hypothesis per stage, one test at a time, and the next stage's number as the judge. A digital marketing strategy without that separation tests the whole page and learns nothing from the result.

Video: why fixing the right stage pays off more than reworking the whole page. In Portuguese, on the mkt4edu channel

Which metrics should you track at each stage?

The table below sums up what to measure at each letter and which decision that number supports.

Table: What to measure at each AARRR stage and which decision each number supports

Note that qualified leads do not appear as a stage of their own: they are the result of acquisition and activation read together. When the two are measured separately, it becomes visible whether the problem is the channel or the experience.

In operations with a sales team, it pays to cross these numbers with the traditional sales funnel. AARRR explains behavior; the sales funnel explains negotiation. Together they prevent the easy conclusion that “the lead was bad”.

When should you use RARRA instead of AARRR?

In some contexts, teams adopt RARRA, which puts retention first, to prove value early and reduce churn before scaling acquisition. It does not replace the AARRR framework; it is a useful variation when channels are saturated or the customer acquisition cost is rising.

• Saturated channel: when every extra unit of media spend brings a worse lead than the last.
• High churn: when the base enters and leaves in the same quarter and LTV never forms.
• New product: when you still do not know whether the people who arrive find value.

The choice between the two is about order, not content: the five metrics stay the same. An inbound marketing strategy with a small base usually gains more by starting with retention, and an inbound sales strategy with a long cycle usually gains more by looking at activation first.

The same reasoning applies to production: the content strategy shifts weight depending on the chosen order, and a content marketing strategy aimed at retention speaks to existing customers, not to people still deciding.

How do you apply AARRR at a university?

In education, the five stages get concrete names: acquisition is the candidate arriving from the blog or the campaign; activation is the application started; retention is the candidate who comes back in the next intake window; revenue is enrollment and re-enrollment; referral is the student who brings a classmate. It is the same logic as student recruitment, only measured stage by stage.

1. Measure by intake window, not by calendar month: the education cycle is seasonal and monthly comparison misleads.
2. Separate application from enrollment: they are different stages and usually have different owners.
3. Count re-enrollment as recurring revenue: it is what decides whether LTV exists at all.
4. Ask for the source on the application form: it is the cheapest way to measure referral without a new tool.

An inbound marketing strategy applied to this cut feeds acquisition and activation with the same material, while an inbound sales strategy decides the timing of the call from the qualified leads activation revealed.

A digital marketing strategy that applies AARRR this way stops arguing about lead volume and starts arguing about where the candidate gives up. It is a less comfortable conversation and a far more useful one.

Frequently asked questions about AARRR

AARRR in practice: where do you start?

Start by measuring all five stages over a single window, even roughly. One bad but comparable number is worth more than five perfect isolated ones, because it is the comparison between stages that points to the bottleneck.

Then pick one stage, the one with the largest relative drop, and run a single test at a time. The compounding gain of the pirate funnel comes from fixing the right stage, not from touching all of them.

If the chosen stage is activation or revenue, the next step is to get into the mechanics of conversion rate optimization, which is where the model stops being a diagnosis and becomes a testing routine.