Growth Loops.
A practitioner-written reference on growth loops: the compounding alternative to linear funnels. Definition, four loop types, worked examples, and the metrics that matter.
By Daria Dovzhikova · Updated May 2026
TL;DR
- A growth loop feeds each activity's output back as the next iteration's input, so growth compounds instead of running linearly like a funnel.
- Four canonical types: viral/referral, content/SEO, paid acquisition, and sales-assisted — each with its own trigger, mechanism, and failure mode, and most companies run two or three at once.
- Loops and funnels answer different questions: loops show what compounds at scale, funnels show where a cohort breaks — run both diagnostics, not one.
What a growth loop is
A growth loop is a self-reinforcing system where the output of one growth activity feeds back as input to the next iteration. Unlike a funnel (which is linear: traffic → signup → activation → revenue), a loop produces compounding growth: each customer makes the next customer easier to acquire.
The framing was popularized by Brian Balfour's 2018 essay at Reforge, "Growth Loops are the New Funnels," and developed further across the Reforge Growth Series curriculum. Andrew Chen's writing on network effects and Lenny Rachitsky's loop teardowns on Airbnb, Pinterest, and Tinder are the standard worked examples.
Four canonical loop types
Viral / referral loop
Existing user invites new user → new user uses product → some % invite their own contacts. Slack, Dropbox, Calendly, Loom all run versions of this. Strongest when collaboration is the core value.
Content / SEO loop
User-generated content (questions, listings, profiles) gets indexed by Google → new visitors land on the content → some % convert. Pinterest, Stack Overflow, Glassdoor, GitHub all run this loop.
Paid acquisition loop
Paid ad → user signs up → user generates revenue → some % of revenue funds more ads. Works when LTV:CAC payback is short and the channel has scale (Facebook, Google). The most mechanical loop but also the most fragile to platform changes.
Sales-assisted loop
Customer success drives expansion within accounts → expansion funds more AE hiring → AEs source new logos → new logos expand. Slow but durable for enterprise SaaS. Standard at $10M+ ARR.
Most successful companies run two or three loops in parallel. Slack runs the viral loop + the SEO loop simultaneously; HubSpot runs the SEO loop + sales-assisted loop; Datadog runs paid + sales-assisted. The compounding effect comes from loops reinforcing each other.
The four loops, side by side
Each loop has a trigger, an action, the value it creates, the mechanism that closes output back to input, the failure mode that kills it, and a DevTools company running it at scale. Six axes that make the abstraction operational.
| Loop | Trigger | Action | Value created | Output → input mechanism | Where it fails | Example |
|---|---|---|---|---|---|---|
| Viral / referral | User experiences value worth sharing | User invites collaborators / shares an artifact | More users inside the same workspace, doc, or channel | Invite link → signup → invite more collaborators | Adding a referral program before the product is actually referable | Slack, Figma, Linear, Loom, Notion |
| Content / SEO | User generates content as a side effect of using the product | Content gets indexed by Google or surfaced inside a community | Organic discovery for new prospects searching the same problem | User content → search index → new visitor → signup → creates more content | Treating SEO as a marketing team output instead of a product output | GitHub (READMEs), Stack Overflow (questions), Vercel (deployment URLs) |
| Paid acquisition | Revenue from existing customers exceeds CAC payback threshold | Marketing reinvests a fixed % of revenue into paid channels | More customers funded by the unit economics of existing customers | Revenue → ad spend → new signups → new revenue → more ad spend | Channel saturation or platform policy changes that break the math overnight | Datadog, MongoDB, HubSpot (at scale, alongside other loops) |
| Sales-assisted | Self-serve account hits a usage tier worth a human conversation | AE expands within the account; CS drives multi-team adoption | Expansion revenue funds more AE hiring, which sources more enterprise logos | Expansion revenue → AE hire → new enterprise logo → expansion → next AE hire | Building the sales motion before the product can deliver standalone value at the team tier | Atlassian, Datadog, Snowflake (at $10M+ ARR) |
Loop taxonomy synthesized from Brian Balfour's Reforge essays, the Reforge Growth Series curriculum, and Lenny Rachitsky's loop teardowns.
Loops vs funnels: when to use which
The two frames answer different questions. Funnels describe what happens to a specific cohort moving through the product. Loops describe how the system compounds over time. The audit-ready growth org runs both diagnostics. Funnels surface where activation breaks for the current sign-up cohort; loops surface whether the business produces compounding growth at all.
For the operational view, see the funnel analyzer. For the systems-design view, see the growth engine framework. For the GTM motion that loops fit into, see go-to-market strategy.
By the numbers
Loops compound across measurable surfaces. One data point that frames what a compounding retention-and-expansion loop actually looks like in the numbers.
Net dollar retention top-quartile SaaS sustains per OpenView's benchmarks — the clearest evidence a retention-and-expansion loop is compounding. Existing customers generating more revenue over time is a growth loop, not a vanity metric.
OpenView SaaS Benchmarks · 2024→Each figure links to the primary source. If a number moves in a subsequent annual report, this page gets updated.
FAQ
What is a growth loop?
A growth loop is a self-reinforcing system where the output of one growth activity feeds back as input to the next iteration. Unlike a funnel (which is linear: traffic → signup → activation → revenue), a loop produces compounding growth: each customer makes the next customer easier to acquire. Brian Balfour at Reforge popularized the framing in his 2018 essay 'Growth Loops are the New Funnels'.
What's the difference between a growth loop and a funnel?
A funnel describes a one-way customer journey from awareness to conversion. A loop describes a system where the conversion's output (a new customer, a piece of content, more revenue) becomes the next iteration's input. Funnels measure where you're losing people; loops measure what compounds. Both are useful frameworks — the audit-ready growth org runs both diagnostics, not one or the other.
What are the most common growth loop types?
Four canonical types: (1) viral / referral loops where users invite users; (2) content / SEO loops where user-generated content drives organic traffic; (3) paid acquisition loops where revenue funds more ads; (4) sales-assisted loops where customer success expansion funds more enterprise AE hiring. Most successful companies run 2-3 loops in parallel. The Reforge Growth Series is the practitioner reference.
How do you design a growth loop?
Start with the question: what does each new user produce that helps the next user find the product? If the answer is nothing, you don't have a loop, you have a funnel with a viral coefficient near zero. Design the loop by identifying the user-generated output (content, invitations, revenue), the channel that exposes it (search, social, ads, sales), and the conversion mechanism. Iterate on the loop's friction points, not on top-of-funnel campaigns.
Why do funnels still matter if loops exist?
Loops describe how a system compounds over time; funnels describe what happens to a specific user. You need both: loops for strategy (what compounds at scale), funnels for tactics (where is the friction in the current cohort). The growth org that only thinks in funnels misses compounding; the growth org that only thinks in loops misses operational diagnostics.
What metrics measure a growth loop?
Loop closure rate (% of users who complete the loop), viral coefficient (new users brought in per existing user), and loop velocity (how fast the cycle runs). A viral coefficient above 1.0 means the loop is self-sustaining without paid acquisition. Below 1.0 means the loop amplifies acquisition but doesn't replace it. Lenny Rachitsky's analyses of Pinterest, Tinder, and Airbnb loops in his newsletter are the standard worked examples.
Designing your loops?
Take the diagnostic.
8 questions on your stage, motion, and current GTM gaps. Get matched to an engagement that designs the loops.
Start the diagnosticReady when you are.
Discovery calls are 20 minutes. First one's on me.