The 5-stage framework used to evaluate a product: Acquisition, Activation, Retention, Referral, Revenue. Popularized by Dave McClure in 2007.
In practice
Dave McClure (500 Startups) introduced this framework in 2007 to solve a simple problem: founders were tracking too many metrics without knowing which ones mattered at which stage. The order isn't arbitrary — it's a funnel. A product spending on acquisition while activation leaks is filling a leaky bucket: every euro spent upstream loses value downstream. Most teams find that one single stage is dragging the whole score down; that's exactly what this test calculates, stage by stage, instead of giving one vague overall impression.
The formula
Paying, retained customers = acquired × activation rate × retention rate × paying rate — and referral feeds the first term with K × the last
- Each stage is a rate
Acquisition is the only absolute number; every other stage is the share of the previous one that made it through. That is why the framework is drawn as a funnel, and why a weak stage late in the funnel wastes everything spent before it.
- The stages multiply
Which is the practical point of the whole model: a 10% improvement at each of four stages is not +40% at the end but 1.1⁴ ≈ +46%, and a stage at zero makes every other stage worthless.
- Referral closes the loop
The one stage that feeds back into the top: each retained customer brings K more acquired ones. It is listed last but changes the economics of the first.
The order was named by Dave McClure in 2007 as "Startup Metrics for Pirates" — the acronym is the joke, the funnel is the point. Variants swap the last two (Revenue before Referral) or add Awareness in front (AAARRR); the arithmetic is the same.
Worked example
Where a founder's month actually goes
- 10,000 visitors acquired. 25% activate (2,500). 40% of those are still active at month three (1,000). 10% of those pay (100 customers).
- Option A, the reflex: double acquisition to 20,000 visitors. Result: 200 customers, at double the acquisition cost.
- Option B: leave acquisition alone, take activation from 25% to 35% and retention from 40% to 50% — two onboarding and product changes. 10,000 × 0.35 × 0.5 × 0.1 = 175 customers, at zero extra acquisition cost.
- Then double acquisition: 350 customers. The same spend that bought 200 in option A buys 350 after B — because the funnel was fixed before it was fed.
The framework's real instruction is an order of operations: measure all five, find the weakest, fix it, then pour. Which is also, stage by stage, what the Tour scores.
Orders of magnitude
- There is no benchmark for "AARRR" as a whole — each stage has its own (see CAC, activation, retention, viral coefficient, revenue). What is consistent across companies is the shape of the problem: one or two stages dragging the rest, rarely all five weak at once.
- The stage most often neglected in early-stage products is retention: it is the slowest to measure (cohorts have to age), the least visible at a launch, and the one every other stage depends on. Acquisition is the most often over-invested, for the opposite reasons.
- A useful sanity check: can you name one number per stage, today, from memory? Teams that can are rare; the Tour's fifteen questions are, in effect, that check in three minutes.
How to improve it
- Put one metric per stage on a single page and look at it weekly. Not fifteen dashboards — five numbers, one funnel.
- Work the weakest stage first, even when it is the least exciting. Fixing retention before acquisition feels slow; it is the only order in which the arithmetic works.
- Define each stage's metric in a sentence anyone in the team can repeat: what counts as activated, what counts as retained, what counts as referred. Most disagreements about growth are disagreements about definitions.
- Re-run the diagnosis every quarter. The weakest stage moves: fix activation and retention becomes the bottleneck, fix retention and acquisition is suddenly worth scaling.
In the Tour
Do you have a primary acquisition channel that's identified and measured?
- Yes, clearly identified and tracked20 pts
- We have one, but don't track it closely7 pts
- Not really / it's scattered0 pts
The Tour is this framework turned into fifteen questions: three per stage, each worth 20, 7 or 0, averaged into a stage score out of 20, the five summed into a score out of 100. The first question opens the Acquisition stage; the last closes Revenue. What the result page shows first is not the total but the weakest stage — because that, not the sum, is what the framework tells you to work on.
Questions people ask
What does AARRR stand for?
Acquisition, Activation, Retention, Referral, Revenue: the five stages a user goes through, from first hearing of a product to paying for it and bringing others. Dave McClure coined it in 2007 as "pirate metrics" because the acronym sounds like one. Some versions put Revenue before Referral; the order of the last two matters less than the idea that each stage is a share of the previous one.
Is AARRR still relevant, or has it been replaced by growth loops?
Both are true. Loops describe how the output of one cycle feeds the next (a shared result brings a new user who shares); AARRR describes the stages inside each cycle. You need the funnel to find where users drop and the loop to understand why growth compounds — or doesn't. Teams that dropped the funnel for loops usually rediscover it the first time they ask "where exactly are we losing people?".
How do I apply AARRR to a B2B product with a sales team?
The stages hold; the units change. Acquisition becomes qualified leads by channel, activation becomes the first real use inside the customer's team (not the signed contract), retention becomes renewal and usage depth, referral becomes references and expansion into sister teams, revenue becomes net revenue retention. The trap is to let the CRM's stages replace the user's: a closed deal is a revenue event, not an activation.