Growth Marketing: The AARRR Funnel, Growth Loops and Experiments That Work

Growth marketing borrows its method from the growth teams of software start-ups and applies it to marketing as a whole: set a measurable goal, test ideas quickly, keep what works and drop the rest. Its classic framework is the AARRR model presented by investor Dave McClure in 2007. Its more recent evolution is the growth loop, described by the Reforge team in Growth Loops are the New Funnels (2018). This guide covers the funnel, the experimentation process, loops, channels, metrics, team organisation and examples of experiments you can run next week.

What is growth marketing, and how is it different from traditional marketing?

Traditional marketing is often organised around campaigns: a launch, a budget, a reach goal, then a report. Growth marketing is organised around a metric that matters for the business (active customers, recurring revenue, qualified leads) and works continuously on every step that influences it. Three differences stand out.

• Scope. The whole journey, including activation and retention, which are usually left to the product or customer success teams.
• Method. Hypotheses written down, experiments with a success criterion fixed in advance, decisions based on data rather than opinions.
• Pace. Many small tests instead of a few big bets, so the team learns faster than its competitors.

Growth marketing does not replace brand building. A known brand converts better in almost every test, so the two reinforce each other.

What is the AARRR funnel?

The AARRR model, nicknamed the pirate metrics, splits the customer journey into five stages. Its value lies in showing where people drop off, so you can work on the weakest stage first instead of buying more traffic.

Some teams reorder the model to start with retention: there is little point in pouring visitors into a product that does not keep them. A retention curve that flattens instead of falling to zero is the first sign that acquisition spending will pay off.

How does the growth experimentation loop work?

The engine of growth marketing is a repeated cycle, usually weekly or every two weeks.

1. Analyse. Look at the funnel data and find the stage with the largest drop.
2. Generate ideas. Collect ideas from the whole company, sales and support included, in a single backlog.
3. Prioritise. Score each idea, for example with the ICE grid (impact, confidence, ease) on a scale of 1 to 10.
4. Test. Write the hypothesis, the metric, the target and the duration before launching. One variable per test.
5. Learn. Record the result, positive or negative, in a shared log, roll out the winners and feed the next cycle.

For conversion tests on pages, our comparison of A/B testing tools lists the main options, and the basics of conversion rate optimisation help you choose what to test. A failed test is useful as long as it is recorded: it stops the team from running it again six months later.

Growth loops versus funnels: what is the difference?

A funnel is linear: you pour visitors in at the top, a share converts at the bottom, and growth stops as soon as you stop pouring. A loop is circular: the output of one cycle becomes the input of the next. Reforge describes loops as closed systems in which part of the output is reinvested in the input, which produces compounding growth.

The most common loops are the following.

• Content loop. An article attracts visitors from Google and AI assistants, some become customers, the revenue finances more articles, which attract more visitors.
• Viral loop. A user invites colleagues or shares a document that carries the brand, and the new users invite others in turn.
• User-generated content loop. Users create pages (reviews, profiles, templates) that rank in search and bring new users who create more pages.
• Paid loop. Revenue from customers acquired through ads is reinvested in ads, as long as customer value stays above acquisition cost.

A healthy company usually has one or two dominant loops. Identifying yours tells you where to invest first.

Why are SEO and content a compounding growth loop?

Paid acquisition stops the day the budget stops. An article that ranks keeps bringing visitors for months or years, with no extra cost per click. Each new article adds to the traffic of the previous ones, and internal links between them strengthen the topical authority of the whole site. Google's guidance on helpful content rewards pages written for people, which are also the pages that convert best.

The loop also feeds AI assistants: ChatGPT, Gemini or Perplexity cite pages that answer questions clearly, which brings referral visits and brand searches. To make the loop turn you need three things: a steady publishing pace, distribution so that each article reaches its first readers (see our guide to content distribution strategies), and a way to estimate what the loop will produce (our SEO forecasting guide explains how). SEO results usually appear after two to four months, so this loop rewards patience more than any other.

Which channels should a growth marketer test?

There is no universal best channel. The right one depends on your audience, your price and your sales cycle. A useful rule is to test several channels quickly, then concentrate on the one or two that work. Our list of 50 customer acquisition channels gives a broad starting point.

Which metrics should a growth team follow?

• A North Star metric: one number that reflects the value customers receive, for example weekly active accounts or orders delivered.
• Customer acquisition cost (CAC) per channel, and the time it takes to recover it.
• Customer lifetime value (LTV), and the LTV to CAC ratio.
• Activation rate: the share of new users who reach the first moment of value.
• Retention by cohort, at 7, 30 and 90 days.
• Experiment velocity: the number of tests completed per month, a good indicator of how fast the team learns.

How do you organise a growth team?

Small companies often start with one growth marketer who owns the backlog and borrows time from a developer and a designer. As the company grows, two models dominate: a central growth team that works across the whole funnel, or small teams dedicated to one stage (acquisition, activation, retention). In both cases the team needs direct access to data, the right to change pages and emails quickly, and a weekly meeting where results are reviewed and the next tests are chosen.

Typical skills include analytics, copywriting, SEO, basic design, marketing automation and enough technical knowledge to set up tracking. Nobody masters everything, so the strength of the team lies in its process more than in individual talent.

What are examples of growth experiments?

• Replace a generic home page headline with the main benefit stated in the customer's own words, measured on the sign-up rate.
• Add a comparison table and a short FAQ to the pricing page, measured on clicks to checkout.
• Send a "you are halfway there" email on day two of a trial, measured on activation.
• Publish ten articles answering the exact questions the sales team hears, measured on organic sign-ups after three months.
• Add an "invite your team" step at the end of onboarding, measured on invitations per account.

Conclusion: is growth marketing worth it?

Growth marketing is a disciplined way to grow: look at the whole journey, fix the weakest stage first, test quickly and invest in loops that compound. Among those loops, SEO and content are the most accessible for most companies, because each article keeps working after it is published. If your bottleneck is producing and publishing enough optimised content, Sorank runs your SEO and GEO on autopilot: it finds keywords, writes and publishes articles on your CMS and tracks your visibility in AI assistants, so your content loop keeps turning.