Dave McClure stood on a stage in Seattle in 2007 — an Ignite talk, five minutes, twenty slides that auto-advanced every fifteen seconds — and introduced a framework that would reshape how a generation of startups measured growth. He called it "Startup Metrics for Pirates" because the acronym spelled AARRR: Acquisition, Activation, Retention, Revenue, Referral. Five stages. Five questions. How do users find you? Do they have a great first experience? Do they come back? Do you make money from them? Do they tell others? McClure, who would go on to co-found 500 Startups and invest in over 2,000 companies, built the framework because the startups he was advising were drowning in dashboards that measured everything and explained nothing.
Before AARRR, the default startup metrics were vanity metrics — page views, total registered users, app downloads, social media followers. Numbers that went up and to the right and told the founder absolutely nothing about the health of the business. A company could report 500,000 registered users and be dying: if 2% of those users ever completed onboarding (Activation), 0.5% came back after day seven (Retention), and 0.1% ever paid for anything (Revenue), the 500,000 number was a tombstone decorated as a trophy. AARRR forced founders to stop celebrating the top of the funnel and start interrogating every stage of the customer lifecycle.
Each stage represents a conversion point — a gate where users either advance to the next stage or drop out of the lifecycle entirely. The framework's power is not in any individual metric. It is in the decomposition itself — breaking a single aggregate number ("total users") into five separate conversion rates that reveal where the system is failing. Dropbox discovered that its Referral stage was the highest-leverage point: offering 500MB of free storage for every referred friend created a viral loop that drove Acquisition costs toward zero and fuelled one of the fastest growth trajectories in SaaS history. Slack's growth team identified that teams reaching 2,000 messages exchanged had crossed the Activation threshold — they were almost certain to retain. Once Slack understood this, they focused obsessively on getting new teams past that message count rather than spending on top-of-funnel advertising.
AARRR endures because it is stage-agnostic advice. A pre-launch startup focuses on Activation: can you get the first hundred users to experience the core value? A growth-stage company focuses on Retention: are users coming back without being prompted? A company preparing for an IPO focuses on Revenue and unit economics. The framework scales with the company because it is not a single metric — it is a diagnostic system for identifying where the growth engine is broken at any given moment. The five stages do not change. The stage that demands attention changes as the company matures.
The framework also embeds a critical insight that many growth teams miss: the stages are not equally important at any given time. Optimising all five simultaneously is a recipe for scattered resources and marginal progress on each. The discipline is identifying which single stage is the binding constraint — the bottleneck that limits the entire system — and concentrating resources there until it is no longer the weakest link.
Section 2
How to See It
AARRR is operating whenever a growth team diagnoses a specific stage of the customer lifecycle rather than chasing aggregate numbers. The diagnostic signature is decomposition: breaking the question "are we growing?" into five distinct, measurable questions — one for each stage.
You're seeing AARRR when a founder responds to "how's growth?" not with a single number but with a stage-by-stage breakdown of where customers are advancing and where they are dropping off. The vocabulary itself is diagnostic: a team that talks in AARRR stages is thinking about growth differently than a team that talks in aggregate user counts.
Product
You're seeing AARRR when a product team discovers that 60% of new sign-ups never complete onboarding. The team stops investing in Acquisition and redirects engineering resources to the Activation stage — redesigning the first-run experience, adding progress indicators, reducing friction in the setup flow. Fixing the leaky Activation stage doubles the effective output of every Acquisition dollar already being spent.
Growth
You're seeing AARRR when a growth team builds a dashboard that tracks each stage as a separate conversion rate: visitor-to-sign-up (Acquisition), sign-up-to-first-value-moment (Activation), day-7-return-rate (Retention), free-to-paid conversion (Revenue), and invite-sent-per-user (Referral). The dashboard makes the leaky stage visible in a way that a single "total users" number never could.
Marketing
You're seeing AARRR when a marketing team realises that doubling ad spend on Acquisition will not solve a Retention problem. If users churn after two weeks regardless of how they were acquired, the marketing budget is filling a bucket with a hole in the bottom. AARRR forces the team to ask whether the problem is at the top of the funnel or further down.
Leadership
You're seeing AARRR when a CEO structures the weekly growth meeting around five metrics — one for each AARRR stage — rather than a single north star metric. Each metric has an owner, a target, and a current trajectory. The conversation moves from "are we growing?" to "which stage is underperforming and what are we doing about it?"
Section 3
How to Use It
The primary application of AARRR is diagnostic: identify the weakest stage of the customer lifecycle and concentrate resources there before investing elsewhere. The framework's discipline is in the sequencing — fix the leaky bucket before pouring more water in.
Decision filter
"Before investing in growth, ask: which stage of the customer lifecycle is the constraint? Fix the leakiest stage first. Spending on Acquisition when Retention is broken is pouring water into a sieve."
As a founder
Map your customer journey to the five AARRR stages within the first month of having real users. For each stage, define one metric and one threshold. Acquisition: cost per acquired user. Activation: percentage of sign-ups who reach the "aha moment" — the action that correlates with long-term retention. Retention: day-7 or day-30 return rate. Revenue: conversion from free to paid, or average revenue per user. Referral: percentage of users who invite others.
The framework's value is not in the metrics themselves — it is in the discipline of measuring all five stages simultaneously. A company that tracks only Acquisition will not see the Retention crisis until it is too late. A company that tracks only Revenue will miss the Activation problem that is silently strangling its growth engine.
As an investor
Use AARRR as a diagnostic framework during due diligence. Ask the founder to walk you through each stage with real numbers. A founder who can tell you their day-30 retention rate, their activation threshold, and their viral coefficient understands their business at a mechanistic level. A founder who can only tell you total sign-ups and monthly revenue is flying with two instruments when they need five.
The most revealing question in early-stage diligence is often the simplest: "Which AARRR stage is your biggest constraint right now, and what are you doing about it?" The quality of the answer tells you more about the team's analytical sophistication than any pitch deck.
As a decision-maker
Apply AARRR to resource allocation decisions. If Retention is your weakest stage, redirecting engineering resources from Acquisition features to Retention features will compound more aggressively. Each retained user is a user you do not need to re-acquire — which means fixing Retention reduces the effective cost of Acquisition.
The framework also prevents a common trap: optimising a strong stage while ignoring a weak one. A company with excellent Acquisition and terrible Activation is spending efficiently to bring users to a product that fails to deliver value. The strong Acquisition metric masks the Activation crisis. AARRR exposes both simultaneously.
Common misapplication: Treating the stages as strictly sequential. In practice, the stages interact. Referral feeds back into Acquisition. Activation quality affects Retention. Revenue model design influences which users you acquire. AARRR is a diagnostic decomposition, not a waterfall process — the stages are interdependent, and optimising one often affects others in unexpected ways.
Second misapplication: Using AARRR as a permanent reporting framework rather than a diagnostic tool. The framework is most powerful in the early and growth stages of a company's lifecycle. At scale, the customer journey becomes too complex for five stages to capture meaningfully — a company like Amazon has dozens of activation moments, multiple retention loops, and revenue streams that do not map cleanly to a single funnel.
Third misapplication: Defining stages with the wrong metrics. "Activation" is not the same as "sign-up." Activation is the moment the user experiences core product value — a distinction that requires qualitative understanding of why users retain. Slack's 2,000-message threshold was discovered through data analysis, not assumed from a template. Importing another company's activation metric into your own product is a common and expensive error.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The leaders below share a common discipline: each decomposed growth into measurable stages and concentrated resources on the stage that constrained the entire system. They understood that aggregate growth numbers obscure more than they reveal — and that the highest-leverage intervention is almost always fixing the weakest conversion point rather than pouring resources into the strongest one.
One created the framework. The other built the most famous example of it in practice — even without using the name.
Mark ZuckerbergCo-founder & CEO, Facebook / Meta, 2004–present
Facebook's growth team, assembled in 2008 under Chamath Palihapitiya, became the canonical example of AARRR thinking in practice — even if they never used the acronym. The team's legendary insight was identifying the Activation threshold: users who connected with seven friends within their first ten days retained at dramatically higher rates. That single discovery reframed the entire growth strategy. Instead of pouring resources into Acquisition, Facebook invested in Activation — suggested friends, contact importers, "people you may know" algorithms — all designed to push new users past the seven-friend threshold.
By 2012, Facebook had crossed one billion monthly active users. The discipline of identifying which stage of the lifecycle leaked most, then concentrating resources on that specific stage, was the mechanism behind one of the fastest growth trajectories in technology history.
Dave McClureCo-founder, 500 Startups, 2010–2017
McClure built AARRR because the startups he advised kept drowning in meaningless dashboards. Founders would proudly report page views, total registered users, app downloads — numbers that rose reliably and explained nothing about business health. At 500 Startups, he institutionalised the framework as the standard diagnostic for every portfolio company. During the first board meeting, founders were expected to present metrics stage by stage: Acquisition cost, Activation rate, Retention curves, Revenue per user, Referral coefficient.
The framework exposed problems that vanity metrics concealed. A startup reporting 100,000 sign-ups with a 2% day-30 retention rate had a Retention crisis that no amount of Acquisition spending would fix. McClure's contribution was giving founders a shared language for diagnosing growth problems — a common vocabulary that made the difference between "we need more users" and "we need users who actually stay."
Section 6
Visual Explanation
AARRR (Pirate Metrics) — Five stages of the customer lifecycle, each narrowing as users drop off. The referral loop feeds satisfied users back into Acquisition, creating compounding growth.
The funnel shape is the framework's core visual metaphor: users enter at the top and drop off at every stage. The narrowing is not a design flaw — it is the natural shape of any customer lifecycle. The referral loop on the right side is what transforms the funnel from a draining process into a potentially self-sustaining engine. When the referral loop feeds enough new users back into Acquisition to offset the drop-off at each stage, the company has built a growth machine that compounds without proportional increases in spending.
Section 7
Connected Models
AARRR connects to models that describe the mechanics of growth, the dangers of measurement failure, and the strategic conditions required for each stage to function. Some reinforce the framework's logic by operating on the same decomposition principle. Others create tension by exposing the framework's limitations or by describing forces that work against its stages. The connections below map how AARRR interacts with the broader ecosystem of growth and measurement models.
Reinforces
Flywheel
AARRR's Referral stage is the connection point where a customer lifecycle becomes a flywheel. When satisfied users refer new users — who then activate, retain, pay, and refer others — the funnel feeds itself. Dropbox's referral programme turned the AARRR funnel into a self-reinforcing loop where each satisfied user reduced the cost of acquiring the next one.
Reinforces
Funnel
AARRR is a specific, five-stage instantiation of the general funnel concept applied to the customer lifecycle. The funnel metaphor captures the essential dynamic: users enter at the top and drop off at every stage, so the population narrows as it progresses. AARRR gives the generic funnel shape a fixed vocabulary that startup teams can use to communicate precisely about where the narrowing is steepest.
Tension
Churn
Churn is the force that attacks the Retention stage of AARRR. High churn means users are advancing through Acquisition and Activation but failing to persist — the funnel leaks at its most expensive point, after the company has already paid to acquire and activate each lost user. AARRR makes churn visible as a specific stage failure rather than burying it in aggregate growth numbers.
Leads-to
Network Effects
A strong Referral stage, compounded over multiple cycles, can produce network effects — the point where each additional user makes the product more valuable for all existing users. AARRR identifies when the Referral loop is functioning. Network effects describe what happens when that loop reaches sufficient scale and the product becomes self-reinforcing in a way that transcends the original funnel dynamics.
Section 8
One Key Quote
"The only way to win is to learn faster than anyone else."
— Eric Ries, The Lean Startup (2011)
Ries captured the philosophical foundation beneath AARRR. The framework is not about the metrics themselves — it is about the learning velocity they enable. A startup tracking all five AARRR stages generates structured feedback about what is working and what is broken at every point in the customer lifecycle. Each week of data produces five signals, not one. Each signal points to a specific lever. The company that diagnoses its growth problems fastest — and redirects resources to the leakiest stage before competitors even identify theirs — compounds learning into an advantage that widens with every cycle.
The connection between learning speed and AARRR is structural: a startup with a single growth metric learns one thing per experiment. A startup with five-stage measurement learns five things per experiment — or more precisely, it learns where in the lifecycle each change has its effect. A feature that boosts Acquisition but damages Retention produces a net negative that a single-metric company would not detect until the damage became existential. The five-stage framework catches it in the first data review.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
AARRR's lasting contribution is not the five stages — any reasonable decomposition of the customer lifecycle would produce something similar. The contribution is the shared language. Before McClure's framework, growth conversations in startups were a babel of competing metrics, misaligned definitions, and arguments about which single number mattered most. AARRR gave founders, investors, and growth teams a common vocabulary for diagnosing specific problems. "Our Activation rate is 23% and dropping" communicates more actionable information in eight words than a thirty-slide deck on "growth strategy." The vocabulary forces precision. Precision enables diagnosis. Diagnosis enables action. Most growth frameworks sound obvious in retrospect. AARRR's power was in creating a standard that hundreds of thousands of startups could adopt without customisation.
The failure mode I see most often is measurement theatre. A founder builds a beautiful AARRR dashboard, presents it at board meetings, and continues making decisions based on gut instinct. The dashboard exists as proof of analytical sophistication, not as an input to decision-making. The tell is resource allocation: if the dashboard shows Retention as the weakest stage but the engineering team is building Acquisition features, the metrics are decorative. The companies that get genuine value from AARRR are the ones where the leakiest stage determines the next sprint's priorities — not the CEO's latest idea or the most recent customer complaint.
The second pattern I notice is stage-ordering bias. Founders naturally start at the top of the funnel — Acquisition — because it is the most visible and emotionally satisfying stage. Watching user numbers grow feels like progress. But the highest-leverage intervention is almost never at the top. It is usually Activation or Retention, where small improvements multiply the value of every user already being acquired. The companies that grow fastest are the ones that resist the gravitational pull of Acquisition and work the funnel from the inside out.
The framework's biggest limitation is its silence on the quality of each stage. AARRR tells you that Acquisition is happening but not whether you are acquiring the right users. A company spending heavily on paid acquisition may show strong Acquisition numbers and terrible Retention — not because the product fails, but because the Acquisition channel is attracting users who were never a good fit. The most sophisticated growth teams layer AARRR with cohort analysis and channel-level decomposition: they track each stage separately for each acquisition source, revealing which channels produce users that activate and retain and which channels produce users that inflate the top of the funnel and vanish.
Section 10
Test Yourself
The scenarios below test whether you can identify AARRR thinking — stage-specific diagnosis and action — versus the default mode of chasing aggregate numbers without decomposing the customer lifecycle. The key diagnostic: is the team measuring and acting on specific stages of the funnel, or treating growth as a single undifferentiated number?
The distinction matters because the remedy for a growth problem depends entirely on which stage is failing. A company with weak Acquisition needs marketing investment. A company with weak Activation needs product improvement. A company with weak Retention has a deeper value-delivery problem that no amount of marketing or onboarding redesign will fix. AARRR's power is in forcing the correct diagnosis before prescribing the treatment.
Is this mental model at work here?
Scenario 1
A startup proudly announces at a board meeting that it has reached 200,000 registered users. The CEO presents a slide showing the user count growing 15% month-over-month. No other metrics are discussed. The board applauds the momentum.
Scenario 2
A growth team discovers that their sign-up-to-first-action conversion rate (Activation) is 18%, while industry benchmarks suggest 40–50% is achievable. They pause all Acquisition spending, redesign the onboarding flow, and within six weeks improve Activation to 41%. Total user growth accelerates even though Acquisition spending has not resumed.
Scenario 3
A mobile app company tracks daily active users (DAU) as its sole growth metric. DAU rises from 50,000 to 120,000 over six months. The team celebrates. An investor asks about day-30 retention and learns it has declined from 25% to 11% during the same period. The company has been spending aggressively on paid Acquisition to mask the Retention decline.
Section 11
Top Resources
The AARRR literature spans startup methodology, growth engineering, and product analytics. Start with McClure's original presentation for the conceptual foundation, move to Croll and Yoskovitz for the most practical implementation guide, and use Ellis for the operational playbook of building a growth team around the framework. Graham's essay provides the strategic context for why stage-specific growth metrics matter more than aggregate numbers.
The original presentation that introduced AARRR to the startup world. McClure's slides — profane, dense with examples, and relentlessly practical — lay out the five stages, define the key metric for each, and demonstrate how to identify which stage is the binding constraint. Dated in its examples but timeless in its framework. The long version includes detailed guidance on metrics selection for each stage.
The most practical guide to implementing AARRR-style metrics in a startup. Croll and Yoskovitz map different business models (SaaS, marketplace, e-commerce, media, UGC) to specific metrics at each stage, solving the customisation problem that generic AARRR descriptions leave open. Their concept of "the One Metric That Matters" — a single metric selected from the AARRR stage that is currently the binding constraint — adds operational focus to the framework.
Ellis coined "growth hacking" and built the growth team methodology that operationalises AARRR in cross-functional teams. The book documents how companies like Dropbox, Airbnb, and LinkedIn built growth engines by running rapid experiments at each stage of the customer lifecycle. Particularly strong on the organisational structure required to make AARRR actionable — dedicated growth teams with engineering, data, and marketing embedded.
Ries's build-measure-learn framework provides the operating system within which AARRR functions as the measurement layer. The book's treatment of "innovation accounting" — tracking progress against learning milestones rather than vanity metrics — is the philosophical foundation for why stage-specific metrics like AARRR matter more than aggregate growth numbers. Essential context for understanding why McClure's framework gained traction when it did.
Graham's essay defines a startup as "a company designed to grow fast" and argues that growth rate is the single most important metric for early-stage companies. The essay provides the strategic context for AARRR: if growth is the defining characteristic of a startup, then understanding which stage of the growth engine is constrained — which is precisely what AARRR measures — is the most important analytical capability a founder can develop.
Tension
Vanity Metrics
AARRR was built specifically to combat vanity metrics — numbers that look impressive but reveal nothing about business health. Total sign-ups is a vanity metric. Sign-up-to-activation conversion rate is an AARRR metric. The tension is productive: every time a founder defaults to reporting vanity numbers, the AARRR framework asks which stage those numbers actually reflect and what the conversion rates between stages reveal.
Reinforces
Product/Market Fit
Product/market fit manifests in the AARRR framework as strong Activation and Retention metrics. If users activate quickly and retain without being prompted, the product is delivering genuine value. Sean Ellis's canonical test — "How would you feel if you could no longer use this product?" — is essentially measuring whether Activation has translated into the kind of Retention that signals product/market fit.
