AARRR vs HEART vs Growth Loops vs Flywheel: The Four Models, Compared

What is AARRR (Pirate Metrics) and is it still relevant in 2026?

AARRR is a five-stage funnel framework for measuring startup growth: Acquisition, Activation, Retention, Referral, Revenue. Dave McClure introduced it at Seattle Ignite in 2007 while running 500 Startups. The nickname comes from the pirate-shout abbreviation.

In 2026, AARRR is half-dead and half-alive. It is dead as a planning tool. It assumes growth is linear, treats users as a one-way flow, and cannot describe compounding loops, network effects, or product-led motions. Reforge's 2018 essay on Growth Loops made this critique definitive.

But AARRR is still alive as a scoreboard. It gives marketing, product, finance, and execs a shared vocabulary. When the CFO asks 'where are users dropping off?' you don't answer with a loop diagram. You answer with AARRR.

The failure mode: using AARRR to plan a viral product. If your real growth mechanism is a UGC loop (Pinterest, Notion templates) or a network loop (Slack Connect), AARRR will hide the actual lever and let you optimize the wrong stage for years.

What does each AARRR stage measure?

• Acquisition: how users find you (channels, CAC, traffic sources)
• Activation: first quality experience (signup -> aha moment)
• Retention: do users come back (DAU/MAU, cohort curves)
• Referral: do users invite others (k-factor, NPS, viral coefficient)
• Revenue: do users pay (ARPU, LTV, conversion to paid)

Thomas Petit and Gabor Papp proposed RARRA in 2017, reordering retention to the front. Same stages, different priority order.

What is the HEART framework and when should you use it?

HEART is a user-experience metrics framework developed by Kerry Rodden, Hilary Hutchinson, and Xin Fu at Google in 2010. It measures five dimensions: Happiness, Engagement, Adoption, Retention, Task Success.

HEART exists to solve a specific problem: UX teams drown in raw usage data and can't pick which metrics matter. HEART forces a small number of meaningful metrics per feature. Rodden pairs it with a Goals-Signals-Metrics process, so each metric maps to an actual user goal.

Use HEART when you need to diagnose engagement at the feature or surface level. Onboarding flow not converting? Run HEART against it. Notification volume hurting Happiness scores? HEART tells you. New feature adoption flat? HEART has an Adoption row.

Don't use HEART when you need to plan acquisition, distribution, or compounding growth. HEART is a UX lens. It assumes users are already in the product. The failure mode is treating HEART as a growth model -- you'll optimize for engagement while your acquisition channel quietly dies.

How HEART overlaps with AARRR

HEART's Adoption and Retention rows overlap with AARRR's Activation and Retention stages. The difference: AARRR counts users (how many activated, how many retained). HEART measures quality (how happy, how engaged, how successful). They're complements, not substitutes. Pre-PMF teams often pair AARRR (the count) with HEART (the quality).

What are growth loops and how do they differ from funnels?

Growth loops are closed systems where the output of one cycle becomes the input of the next. Brian Balfour, Casey Winters, Kevin Kwok, and Andrew Chen formalized the concept at Reforge in 2018: 'The fastest-growing products are better represented as a system of loops, not funnels.'

Growth loops differ from funnels in three ways:

1. Direction: Funnels are one-way (users in, users out). Loops feed back (output reinvests as input).
2. Compounding: Funnels grow linearly with input. Loops grow exponentially when the loop coefficient exceeds 1.
3. Strategy integration: Funnels silo product, marketing, and monetization. Loops force you to design them as one system.

The canonical examples are Pinterest's UGC loop (users save pins -> pins get indexed by Google -> SEO drives new users -> new users save more pins), Slack's network loop (one team adopts -> invites external collaborators via Slack Connect -> those collaborators bring Slack to their teams), and Notion's template loop (users build templates -> share publicly -> templates rank in search -> search drives signups).

The failure mode: drawing a loop that doesn't actually loop. If you can't write the math (X new users -> Y actions -> Z outputs -> X' new users where X' is a function of X), you've drawn a flywheel diagram, not a growth loop.

What are the main loop archetypes?

Most B2B SaaS growth runs on one of four archetypes:

• UGC / SEO loops: users create content -> content ranks in search -> search drives users (Notion, Canva)
• Viral / network loops: users invite others as a core feature of value (Slack, Figma, Loom)
• Paid loops: revenue from existing users funds paid acquisition of new users (most B2C subscription)
• Sales-assisted loops: champions move companies, bringing the product to new accounts (Gong, Linear)

More on this in growth loop archetypes.

What is the HubSpot Flywheel and is it a metric or a metaphor?

The Flywheel is a narrative, not a metric. HubSpot CEO Brian Halligan introduced it at Inbound 2018 as a replacement for the marketing funnel. The model has three stages -- Attract, Engage, Delight -- arranged as a wheel around the word 'Growth.'

Halligan's argument: funnels lose energy at the bottom; flywheels store and release it. Customers aren't an output; they're an input that powers more growth via referrals, reviews, and word of mouth. As a piece of executive storytelling, this is excellent. As a planning tool, it's almost useless.

The flywheel cannot be instrumented. There is no Flywheel KPI. 'Delight' is not a number. The model is qualitative by design. That's a feature when you're addressing a board or writing a brand narrative. It's a bug when you're trying to allocate engineering capacity.

The failure mode: treating the flywheel as a metric. Teams build 'flywheel dashboards' that are really just renamed AARRR funnels arranged in a circle. The math doesn't change because you drew it round.

The right pairing: use the flywheel for the all-hands deck and the homepage. Use Growth Loops to plan what you're actually building.

Where the Flywheel and Growth Loops actually agree

Both reject the funnel as a complete model. Both treat existing customers as inputs to future growth. The disagreement is rigor: Halligan kept it qualitative for HubSpot's enterprise sales narrative; Balfour insisted on the math because Reforge is teaching operators to ship. If you're an operator, this is not a coin flip.

How do AARRR, HEART, Growth Loops, and Flywheel actually compare?

Below is the comparison spine of this article. Read across the rows to see when each framework breaks and what to pair it with.

Framework	Origin	What It Measures	When It Breaks	Best Used For	Pairs Well With
AARRR	Dave McClure, 2007	Funnel conversion across 5 stages	Treats growth as linear; misses compounding	Org alignment, financial reporting	Growth Loops
HEART	Rodden, Hutchinson, Fu at Google, 2010	UX quality at feature level	Doesn't describe acquisition or compounding	Feature/UX diagnostics	AARRR
Growth Loops	Balfour et al, Reforge, 2018	Closed-system compounding	Hard to communicate cross-functionally	Long-term strategy	AARRR
Flywheel	Halligan, HubSpot, 2018	Customer-centric narrative	Cannot be instrumented as KPI	Executive narrative, brand	Growth Loops

The pattern: measurement frameworks (AARRR, HEART) need to pair with strategy frameworks (Growth Loops) or narrative frameworks (Flywheel). One alone is incomplete.

Which growth framework should a startup pick in 2026?

Pick two, not one. Almost every team needs a measurement framework plus a strategy framework. Here's the matrix:

• Pre-product-market-fit: AARRR + HEART. AARRR gives you the funnel scoreboard. HEART tells you whether activation is working at the UX level. Skip Growth Loops until you have signal.
• Post-PMF, scaling: AARRR + Growth Loops. AARRR for cross-functional reporting and finance. Growth Loops for actual planning. This is the modern default for product-led companies.
• Enterprise sales-led: AARRR + Flywheel. The flywheel narrative aligns sales, CS, and marketing around customer expansion. AARRR keeps the pipeline honest.
• Mature product-led with brand ambitions: Growth Loops + Flywheel. Loops for the engineering plan, flywheel for the keynote.

The three failure modes to avoid:

1. Using AARRR to plan a viral product. You'll optimize Activation while ignoring the loop coefficient that actually matters.
2. Using Flywheel as a metric. You'll build a dashboard that looks circular but measures nothing.
3. Using Growth Loops as your only language. Finance and the board don't speak loops. They speak funnels. AARRR is the translator.

If you don't know which loop you're running, start with build your first growth loop and map your growth model step by step.

What are the wrong-tool failure modes for each framework?

Every framework has a specific way it goes wrong when applied to the wrong problem. Knowing the failure mode is faster than reading the manual.

• AARRR misused as strategy: A consumer social app team plans 12 months around AARRR. They optimize each stage in isolation. Acquisition costs creep up because they never built a referral loop. Retention plateaus because activation was tuned for the wrong cohort. The team realizes too late that growth requires a loop, not a funnel.
• HEART misused as growth model: A B2B SaaS team adopts HEART company-wide. Engagement dashboards look great. Six months later, signups have flatlined because nobody is responsible for the acquisition channel. HEART has no Acquisition row.
• Growth Loops misused for finance reporting: A growth team presents a three-loop diagram to the CFO. The CFO needs cohort revenue and CAC payback. The diagram doesn't help. The growth team loses budget for not speaking the finance dialect.
• Flywheel misused as a KPI: A marketing team builds a 'flywheel velocity score.' It is an unweighted average of NPS, retention, and referral rate. It moves around, but no decision is ever made because of it. The dashboard is theater.

If the framework you picked is producing one of these failure modes, you picked the wrong tool. Add or swap, don't double down.

Framework	Origin	What It Measures	When It Breaks	Best Used For	Pairs Well With
AARRR (Pirate Metrics)	Dave McClure, 500 Startups, 2007	Funnel conversion at 5 stages: Acquisition, Activation, Retention, Referral, Revenue	Treats growth as linear and one-directional. Misses compounding, ignores re-acquisition of churned users.	Cross-functional org alignment. Giving marketing, product, and finance a shared vocabulary.	Growth Loops (for strategy on top of measurement)
HEART	Kerry Rodden, Hilary Hutchinson, Xin Fu at Google, 2010	User experience quality: Happiness, Engagement, Adoption, Retention, Task Success	Doesn't tell you where to acquire users or how the product compounds. UX-only lens.	Diagnosing engagement and product quality at feature or surface level.	AARRR (HEART fills in the Activation/Retention diagnostics)
Growth Loops	Brian Balfour, Casey Winters, Kevin Kwok, Andrew Chen, Reforge, 2018	Closed systems where outputs reinvest as inputs (UGC loops, viral loops, content loops, sales loops)	Hard to map to org structure. Doesn't give finance a P&L view. Requires deeper data instrumentation.	Long-term compounding strategy. Answering 'how does our product actually grow?'	AARRR (loops as strategy, AARRR as scoreboard)
Flywheel	HubSpot, Brian Halligan, Inbound 2018	Customer-centric momentum: Attract, Engage, Delight (qualitative narrative)	Not a metric. Cannot be instrumented as a KPI. Often used as an org chart, not a model.	Executive narrative and customer-experience storytelling.	Growth Loops (flywheel as story, loops as mechanics)