# Growth Funnels vs. Growth Loops

> Funnels leak. Loops compound. A practical guide to designing growth that feeds itself.

- Author: Rishikesh Ranjan · Published: Feb 18, 2026 · Updated: Jun 12, 2026
- Type: Essay
- Tags: Frameworks, Growth Loops, Product-Led Growth
- Growth levers: Acquisition (primary), also Referral
- ~1339 words

---

Dropbox's first growth engine was a funnel. Paid search at the top, signups at the bottom. It cost them [$233 to $388 to acquire one customer](https://www.slideshare.net/gueste94e4c/dropbox-startup-lessons-learned-3836587?utm_source=productgrowth.blog) for a product that sold for $99 a year.

Do the math. Every customer was a loss.

So they killed the ads and built a loop instead: refer a friend, you both get free space. Result: 100,000 registered users became 4 million in 15 months. **That's 3,900% growth**, with 35% of daily signups coming from the referral loop itself. Same product. Different machine.

Here's the one-line difference: a **growth funnel** is linear, you pour users in the top and some convert while the rest leak out. A **growth loop** is circular, each cohort of users takes an action that recruits the next cohort, so growth compounds without new spend.

- **Funnels are additive**: 1,000 users a month is 12,000 a year. Stop spending, growth stops.
- **Loops are multiplicative**: users bring users. The output of this month is the input of next month.
- **You need both**: a funnel to seed the loop, a loop to stop renting your growth from ad platforms.

## What is a growth funnel?

The funnel is the model most of us learned first: AARRR. Acquisition, activation, retention, referral, revenue. Traffic enters at the top, a percentage survives each stage, and customers drop out the bottom. Pirate metrics, [Hacking Growth](https://www.productgrowth.blog/p/hacking-growth), every dashboard you've ever built.

Funnels are genuinely useful. They tell you where users leak: a 60% drop between signup and [activation](https://www.productgrowth.blog/calculators/activation-rate) is a finding you can act on this sprint. They map cleanly to team ownership. And when you have zero users on day one, a funnel is the only machine that works, because there's nobody inside the product to power a loop yet.

The problem is what the funnel hides. Nothing flows back to the top. Every new user must be bought all over again, so your [CAC](https://www.productgrowth.blog/calculators/customer-acquisition-cost-cac) sets the ceiling on your growth. The channel gets more expensive every year, and the funnel never fights back.

## What is a growth loop?

A growth loop closes the circuit. A user joins, does something inside the product, and that something lands in front of people who aren't users yet. Some of them join and repeat the cycle. [Reforge](https://www.reforge.com/blog/growth-loops?utm_source=productgrowth.blog) made the framing famous in 2018: funnels answer "how do we get more users into stage X?", loops answer "how does the product grow itself?"

![Growth loop diagram with four clockwise steps: a new user joins, uses the product and creates output, that output reaches non-users, and some of them sign up, feeding the next cycle.](https://www.productgrowth.blog/media/posts/growth-funnels-vs-growth-loops/02-anatomy-of-a-growth-loop.webp)

The loop's fuel can be almost anything users produce:

- **Incentives**: Dropbox's give-space-get-space referral, the canonical example.
- **Content**: every Pinterest board and Notion template is a public page Google indexes; searchers land on it and some sign up to make their own.
- **Collaboration**: a Figma file or Loom link only works when you send it to someone, and that someone needs an account to reply in kind.
- **Output watermarks**: every AI answer Perplexity users shared carried the brand; the company stacked six loops like this on its way to a $9B valuation in two years. I broke them down in [the Perplexity teardown](https://www.productgrowth.blog/p/how-perplexity-hacked-its-growth).

Loops don't have to be acquisition loops either. [Duolingo's streak](https://www.productgrowth.blog/p/how-duolingo-hooks-users) is a retention loop: yesterday's practice is the reason you show up today. [Lovable](https://www.productgrowth.blog/p/how-lovable-dev-hacked-their-growth) runs five engines that feed each other. Once you see loops, you stop seeing acquisition, product, and monetisation as separate silos. They're stations on the same circuit.

|  | Growth funnel | Growth loop |
| --- | --- | --- |
| Shape | Linear, top to bottom | Circular, output feeds input |
| Growth type | Additive (X users/month) | Compounding (X% on the base) |
| Cost over time | Rises as channels saturate | Falls as the base grows |
| Who does the work | Your budget | Your users |
| Stops when | You stop spending | The loop's incentive decays |

## The math: why loops beat funnels over time

Take two identical products with identical budgets. Both buy 1,000 signups a month. Product A runs a pure funnel. Product B also ships a referral loop where every month, users equal to 15% of the existing base join through invites, shared links, and indexed content.

| Month | Funnel users | Loop users |
| --- | --- | --- |
| 1 | 1,000 | 1,000 |
| 6 | 6,000 | 8,753 |
| 12 | 12,000 | 29,000 |
| 18 | 18,000 | 75,833 |
| 24 | 24,000 | 184,161 |

![Line chart, same acquisition spend over 24 months: a funnel adding 1,000 users per month reaches 12,000 users at month 12 and 24,000 at month 24. The same spend plus a 15% monthly referral loop reaches 29,000 users at month 12 and 184,161 at month 24.](https://www.productgrowth.blog/media/posts/growth-funnels-vs-growth-loops/01-funnel-vs-loop-24-months.webp)

At month 6 the gap looks boring: 8,753 vs 6,000. By month 12 the loop is 2.4x ahead. By month 24 it's **7.7x ahead on the exact same spend**. That's the whole argument for loops in one chart. Compounding is invisible early and unbeatable late.

And notice what the early months imply: if you judge a loop in its first quarter, you will kill it. Funnels demo well. Loops need patience.

## Where funnels still win

Now the part the "funnels are dead" posts skip. Plenty of products should run a funnel, unapologetically:

- **Day zero**: loops need users to power them. With an empty product, paid and outbound funnels are how you seed the flywheel.
- **One-off, high-ticket purchases**: nobody refers their friends to the company that sold them an ERP migration. Sales-led B2B lives and dies by funnel discipline.
- **Private-by-nature products**: if usage produces nothing a non-user ever sees (think tax software), the loop has no surface to travel on.

And a warning from someone who has audited a lot of "loops": **most growth loops are funnels in disguise**. A share button is not a loop. A referral program nobody uses is not a loop. It's only a loop if the cycle actually completes at a meaningful rate, and you can see it in the numbers.

## How to tell which one you have

> **Rule of thumb:** If turning off paid spend stops growth entirely, you have a funnel. If growth keeps moving, you have a loop.

You don't have to literally turn off the ads (your CFO will thank you). Instrument it instead. Tag every new signup with its source and watch one number: **the share of new users attributable to existing users**. Invites sent, shared links clicked, public content indexed, team members added. If that share is meaningful and stable cohort after cohort, a loop is turning. If it rounds to zero, you're renting all of your growth.

Then work the loop like a funnel optimizer would: measure each arc of the circle (what % of users create output? what % of output gets seen? what % of viewers convert?) and fix the weakest arc first. Loop thinking and funnel metrics are friends, not rivals. More on the measurement side in [measuring success in product-led growth](https://www.productgrowth.blog/p/measuring-success-in-product-led-growth).

> **Steal this:** Map your funnel, then ask of every stage: what could users at this stage do that brings in the next user? The first answer you can instrument is your loop. Ship it small, measure the share of signups it drives, and only scale it once the cycle completes on its own.

## FAQ: growth funnels vs growth loops

#### Can a product use both a growth funnel and a growth loop?

Yes, and the best companies do exactly that. Paid and outbound funnels seed the user base; loops compound it. Dropbox kept acquisition channels running while the referral loop drove 35% of daily signups. The funnel buys the first cohort, the loop makes every later cohort cheaper.

#### What is an example of a growth loop?

Dropbox's referral program is the classic: refer a friend, both get free storage, new users repeat the cycle. Content loops are just as common: Pinterest boards and Notion templates are public pages that rank on Google, and some visitors sign up and create more pages. Perplexity stacked six loops, including shared AI answers that carried the brand to new users.

#### When is a funnel better than a loop?

Three cases: you have zero users (loops need users for fuel), you sell one-off high-ticket deals where referrals rarely happen, or your product's usage produces nothing a non-user ever sees. In those cases, optimize the funnel and don't force a loop that has no surface to travel on.

#### How do I know if my growth loop is working?

Track the share of new signups attributable to existing users: invites redeemed, shared links converted, public content signups, teammates added. A working loop shows a meaningful, stable share cohort after cohort. If growth flatlines whenever paid spend pauses, the loop is decorative.

Net-net: funnels get you your first thousand users. Loops decide whether you ever stop paying full price for the next thousand.

Map your product this week. If you find a loop hiding in it (or a share button cosplaying as one), hit reply and tell me. I'll help you pressure-test it.

---

All posts: https://www.productgrowth.blog/archive · Site: https://www.productgrowth.blog