What Are the AARRR Metrics?
Flash back to August 8, 2007: The Plain White T’s’ “Hey There Delilah” was the No. 1 song in the country; the Republican primary was in full swing, with Mitt Romney in the lead; New York had just weathered a tornado.
But in the Pacific Northwest, the weather was mild, and Dave McClure, fresh off stints leading marketing at PayPal and Simplyhired, was giving a talk at a conference called Ignite Seattle.
The event’s tagline was “enlighten us, but quickly,” and McClure complied. He spoke at an auctioneer’s pace, and his talk, “Product Marketing Metrics for Pirates: AARRR!,” clocked in at just over five minutes. But it had an outsized impact on the growth space.
In the talk (above), McClure laid out five key moments in digital products’ customer life cycle. Let’s take a closer look at these metrics.
The AARRR Metrics
What It Is: The moment a user finds the company’s digital presence. They typically arrive from a wide variety of channels, including email, social media, SEO and more.
How to Measure It: Acquisition is usually measured by click-through, though McClure recommends ignoring users who click-through and “bounce,” or promptly leave the site — they probably arrived by accident. From there, he recommends breaking down bounce-controlled click-through by channel, and evaluating each channel based on its price point, the volume of users it brings in, the quality of those users. Do users stick around? Do they refer their friends?
What It Is: Sometimes referred to as the “aha! moment,” this is the point in the journey where the user sees the value in the product, and tips from a neutral to what McClure calls a “happy” user experience.
How to Measure It: Ultimately, the activation metric depends on the product. At Dropbox, activation means a new user uploads at least one file to a Dropbox folder; in the early days of Facebook, activation meant a new user made seven friends within their first 10 days on the platform. McClure suggests startups use something simple as their activation metric — a certain number of page views, say, or 30 seconds spent on the site.
What It Is: Unlike the other stages so far, this one is less a particular moment than a process — the process by which a new user becomes a long-term user.
How to Measure It: McClure recommends that startups send out regular, automated emails to new users — maybe three in the first month after they first arrive on the site — to boost retention. These emails then offer up natural retention metrics: email open rates and click-through rates over time.
What It Is: The user recommends the product to other people.
How to Measure It: Though McClure doesn’t touch on specific metrics for this, there are various ways to track and incentivize referrals. For instance, new apps and services often offer users referral bonuses, like in-app cash, for getting their contacts on board.
What It Is: The user takes a money-making action. What that action is, exactly, can vary widely from product to product. It could mean going from hosting free events to paid events on Eventbrite; for Netflix, it could mean continuing a subscription past the free trial stage.
How to Measure It: In dollars, unsurprisingly — though McClure also recommends comparing it with minimum viable revenue, and the revenue level at which the company would break even.
* * *
Pirates and hackers have a similar rogue energy, and many consider AARRR a growth hacking framework, but McClure’s talk actually predated any mainstream tech discussions about growth hacking. The term wasn’t coined until 2010. McClure gave his talk the same year Facebook created the tech industry’s first internal growth team. No one knew if it would work.
Spoiler: It did. By the end of 2019, the growth team had gotten Facebook to 2.5 billion active monthly users. Many other companies have followed Facebook’s lead; today growth teams are tech industry standard. “Growth hacking,” not so much. It became a ubiquitous buzzword, then petered out of use.
McClure, for his part, moved into the venture capital sphere shortly after his talk; he later resigned from a fund he cofounded, acknowledging he had a pattern of workplace sexual misconduct. Recently, he founded Practical Venture Capital.
As for his AARRR metrics: They’re still widely used in the tech world, but they’re no longer strictly his. Other tech leaders have adapted and modernized McClure’s initial framework.
The AARRR Metrics Grow Up
For one thing, McClure’s “startup metric” terminology has fallen away. “I think the term ‘startup’ is becoming a misnomer,” Ben Winter, head of growth and marketing at Fairmarkit, told Built In. He sees the “pirate metrics” as broadly applicable for tech companies — especially for SaaS companies striving for product-led growth.
Why Saas? SaaS products bring in revenue through monthly or annual subscription fees, and, in Winter’s view, AARRR metrics are most relevant when “your existing revenue impacts future revenues.” With subscriptions, current and future revenue are clearly linked, because payment recurs.
Back when McClure gave his talk, though, he probably wasn’t thinking of SaaS companies. His metrics tracked the customer journey chronologically, and he put retention before revenue, which implied a freemium business model. In the SaaS space, though, the same AARRR acronym works — retention and revenue just switch places, according to Winter.
What about product-led growth — what does that have to do with AARRR metrics? Well, it involves optimizing product features through constant small experiments. These metrics, Winter noted, can help growth teams better prioritize potential experiments.
Typically, growth teams prioritize based on a project’s ICE score. (ICE is an acronym for impact, confidence, and ease of implementation — the higher the score, the more important the test.) In Winter’s experience in his previous role at Appcues, AARRR metrics refine impact forecasts.
“Every experiment that I would run ... would involve measuring all of these metrics,” he said. He even built a public AARRR-based impact calculator while he was there.
Today, Winter works at Fairmarkit, a company whose enterprise software uses AI to shop around for small purchases — “small,” here, sometimes meaning “costing less than $50,000”— and optimize clients’ tail spend. It’s not a SaaS product, and growth isn’t particularly product-led, so the AARRR metrics don’t have the same salience in his day to day.
“I don’t think we have an activation moment,” Winter said. “If we did, it would be the first time you send out a requisition, but every single one of our customers does that because they’ve already bought the software and got it set up.”
The AARRR metrics are still relevant at Fairmarkit, though, when it comes to the vendors bidding on clients’ tail spend projects. “We’re a two-sided marketplace,” Winter explained — and on the vendor side, he and his team try to optimize acquisition and activation.
The ultra-streamlined bidding process, for instance, makes vendor acquisition easier. (“We don’t want to create friction,” Winter explained.) Meanwhile, signups for a vendor portal, which shows metrics for that vendor’s bids across Fairmarkit’s platform, arguably constitute activation.
So yes: AARRR metrics even have relevance in enterprise software.
Winter has actually found them consistently relevant since he first learned about them, back in 2013 at Reforge. The bootcamp-style growth course was created by Brian Balfour, fresh from leading growth at Hubspot, and Andrew Chen, who had led explosive growth at Uber.
Back then, Winter said, a lot of the coursework focused on viral loops: new customers generating more new customers. This happens through a variety of flywheel-like processes. Take Pinterest’s SEO strategy: It encouraged new users to discover Pinterest boards through search, and then create their own boards, which in turn attract more new users through search.
There are more than 20 similar growth loops, Chen and Balfour argue — McClure’s referrals are just one of many ways users can sustainably beget users.
This is one of various reasons they’ve been critical of McClure’s framework, which offers “too micro a view,” they argue. In general, it focuses more on the individual customer’s journey than the big-picture process of growing a customer base.
The framework also encourages siloing, Chen and Balfour worry, by presenting the five stages as separate.
“You shouldn’t test on one [AARRR metric] at a time,” Winter agreed. “In the last five or so years, we’ve figured out how interdependent they are.”
The Rise of Retention
For his part, Winter has begun shifting his focus away from referral, toward acquisition and retention — two potential bottlenecks in the customer lifecycle. Issues with these two metrics can hamstring huge improvements in acquisition, referrals and even revenue, he has argued.
Retention, especially, has emerged as a key metric — even to Balfour as an AARRR skeptic. “Any metric that is intended to be an indicator of overall authentic growth must have a retention built in,” he wrote on his personal blog.
One retention problem, in particular, has been on many growth leaders’ mind lately. “The area that’s really being honed in on now is this leaky bucket concept,” Winter said.
That’s when weak retention — also known as high churn — undercuts otherwise strong pirate metrics, because customers are “leaking” out of the user base “bucket.” Even if a company with low retention can somehow juice revenue, for instance, that growth may not be sustainable for even a month.
Worse, low retention “means people aren’t happy with your software,” Winter said.
Retention is so key that it’s made Winter reevaluate what he considers growth in the first place. It’s not just about “growing” the overall customer base, he said — which can be done through a loop of referral, acquisition and activation — but also about “growing” individual established customers.
If a company can retain a core group of users and continually upgrade their subscriptions, he pointed out, that company can grow revenues without much acquisition or referral at all — and with net-negative churn like that, investments in customer acquisition quite literally pay off.
Of course, activation remains important too. If customers don’t activate, newly acquired users don’t engage more deeply with the product, and it can’t amass a core group of loyal users.
As a metric, though, activation poses challenges. “I’ve worked at companies where I’ve changed the definition of activation every three months,” Winter said. “Activation’s an emotional feeling, which I think is really hard to measure.”
He predicts, though, that many companies will begin to measure activation based on long-term user’s feedback. What hooked them?
In other words, “you’ll start to see activation be dictated by retention,” Winter said. “That retention metric’s more and more important.”