Growth loops vs funnels: Andrew Chen's lesson for SMEs designing one loop

Who Andrew Chen is and what he argues

Andrew Chen is a general partner at Andreessen Horowitz (a16z) and was previously head of rider growth at Uber, where, according to his publisher's author page, he oversaw growth from 15 million to 100 million active users. His book The Cold Start Problem (2021) is about network effects: products that become more valuable as more people use them, and the difficulty of getting such a product off the ground.

Chen also co-authored, with Brian Balfour, Casey Winters and Kevin Kwok, the 2018 Reforge essay "Growth Loops are the New Funnels". Its central claim is that the familiar acquisition–activation–retention–referral–revenue funnel describes only one slice of how fast-growing products actually grow.

"The fastest-growing products are better represented as a system of loops, not funnels." — Balfour, Winters, Kwok and Chen, Reforge

The essay names three problems with funnel thinking. It creates strategic silos (product, channels and monetisation are planned separately, even though they constrain each other), functional silos (marketing optimises sign-ups while product owns retention, so low-quality acquisition quietly damages retention), and it is one-directional: nothing is reinvested, so every extra unit of growth needs more budget, more people or more channels.

The network-effect lens from The Cold Start Problem

In the book, Chen describes five stages that networked products go through: the Cold Start, the Tipping Point, Escape Velocity, Hitting the Ceiling and the Moat. He outlines these in his essay on hiring a head of growth, where he also notes that the largest internet products are driven by only a few growth loops, and that founders who chase many ideas at once rarely build a repeatable system.

Two of his ideas are useful far beyond Silicon Valley:

• The atomic network. The smallest network that can sustain itself: two people for a Zoom call, a team for Slack, hundreds of listings in one city for Airbnb. In a book excerpt published by Lenny's Newsletter, Chen advises starting with density in a narrow niche rather than chasing market size.
• Three effects that compound. At Escape Velocity, Chen separates an acquisition effect (the network brings in users cheaply), an engagement effect (usage rises as the network fills out) and an economic effect (monetisation improves as the network grows).

"Your product's first atomic network is probably smaller and more specific than you think." — Andrew Chen, The Cold Start Problem

Why this matters for a European SME

A funnel-only growth plan is a plan to keep paying for attention. When the budget stops, growth stops. A loop, even a modest one, means each customer cohort contributes something to the next.

The Reforge essay illustrates the trade-off with two hypothetical initiatives: one that brings in 500 users once, and one that brings in 20 users in its first week and grows 10% a week. By our own arithmetic, the second overtakes the first in cumulative users around week 14 and keeps going. That is the difference between a campaign and a system.

You do not need a marketplace or a social app to benefit. The Reforge essay's Pinterest example is instructive: users save content, Pinterest makes that content findable in search engines, and searchers arrive, sign up and save more content. Replace "pins" with case studies, product reviews, recipes or technical answers and the same mechanics apply to many B2B and e-commerce businesses.

A playbook: design one loop in eight steps

1. Ask the cohort question. The Reforge essay's starting point: how does one group of customers lead to the next group? Write the honest answer down, even if it is "we buy more ads".
2. Run the whiteboard test. Ask five people across the business to sketch, separately, how the company grows. Mismatched answers show where your growth model is unclear.
3. Choose one loop type. Typical options for SMEs: a content loop (customer work or questions become published content that attracts the next customers), a referral loop (happy customers invite peers), a user-generated loop (reviews, photos or projects shared publicly), or a paid loop (revenue from new customers is reinvested in acquisition).
4. Map each step as a verb. For a content loop: client project completed → anonymised insight published → found via search or LinkedIn → enquiry → new project. Each arrow is a conversion you can measure.
5. Find your atomic network. Pick the narrowest segment where the loop can sustain itself: one region, one industry, one product line. Density beats breadth.
6. Measure the loop as a ratio. In a 2025 essay on viral loops, Chen describes the viral factor as a ratio between generations of users and suggests tagging shared links so each sign-up can be traced to the person who brought it in. Referral codes and UTM parameters do the same job for a small business.
7. Remove friction at the weakest step. Fix the lowest-converting step before adding tactics.
8. Review quarterly. Loops degrade as channels saturate, the stage Chen calls Hitting the Ceiling. Plan the next loop before the current one stalls.

Common mistakes and limits

• Drawing too many loops. The Reforge authors warn that a diagram full of loops usually means many weak ones. Fast-growing products typically rely on one or two major loops.
• Ignoring retention. In an essay on virality, Chen argues that viral growth depends on retention: aggressive invite tactics from the early social-app era spiked and then collapsed because users did not stay.
• Assuming network effects you do not have. Most service businesses and shops do not get more valuable to each customer as they add customers. A loop can still compound, but do not plan for a winner-takes-all moat.
• Expecting instant results. Loops start slowly. Judge them on their trajectory, not their first month.

FAQ

Sources

• The Cold Start Problem (book page) — Andrew Chen / Penguin Books, 2021
• Growth Loops are the New Funnels — Brian Balfour, Casey Winters, Kevin Kwok, Andrew Chen / Reforge, 2018
• What to look for when you're hiring a Head of Growth — Andrew Chen
• Essay on viral growth and retention — Andrew Chen
• The atomic network (book excerpt) — Andrew Chen / Lenny's Newsletter, 2021
• The lost art of designing viral loops — Andrew Chen / a16z speedrun, 2025