The Cold Start Problem
by Andrew Chen · 2021
Every network product looks impossible right up until it isn't -- Chen maps exactly how it tips.
Worth reading? The Cold Start Problem is Crossing the Chasm updated for network products. Geoffrey Moore's chasm is about getting from early adopters to the mainstream for any tech product; Chen's cold start theory is about the harder version of that problem when your product is worthless until other people show up too. Read Chen's book if you're building a marketplace, social app, or anything with network effects -- the atomic network concept alone is worth it. Skip it if you're not building a network product; most of the framework, tipping points and moats included, simply doesn't apply to a single-player or non-network business.
| Full Title | The Cold Start Problem: How to Start and Scale Network Effects |
|---|---|
| Author | Andrew Chen |
| Published | 2021 |
| Publisher | Harper Business |
| Category | Business & Money |
| Favorite quote | “Network effects are the closest thing to magic in the business world.” |
The Verdict
Chen spent years inside Uber watching a real network product fight its way from zero to global scale, then spent more years at Andreessen Horowitz watching dozens of founders try to repeat it. The five-stage framework here is the closest thing to a repeatable playbook anyone’s written for a problem that looks different in every case study but somehow always plays out the same way.
you're building or growing anything with network effects -- a marketplace, social app, or communication tool
you're not building a network product -- most of this framework doesn't transfer to a business without network effects
Book Summary
Chen, a former Uber growth executive and Andreessen Horowitz partner, tackles the cold start problem head-on: how does a network product get anyone to use it when it's worthless with zero other users on it? His answer is the Cold Start Theory, a five-stage framework -- the cold start problem, the tipping point, escape velocity, hitting the ceiling, and the moat -- that every successful network product moves through.
The key insight is that you don't launch a network product to a whole market. You find and ignite a single "atomic network," the smallest group of users who get enough value from each other to stick around even at almost no scale -- a single apartment building for a marketplace, a single friend group for a social app. Only once that atomic network is stable do you copy-paste the same formula into the next one.
Later stages cover why growth stalls even after early success -- the ceiling, whether it's market saturation or the core loop breaking under scale -- and how network effects eventually become a genuine moat competitors can't easily replicate, unlike a feature advantage, which erodes fast. Chen leans heavily on case studies from Uber, Airbnb, Tinder, Slack, and Zoom.
Top 10 Lessons from The Cold Start Problem
- A network product is worthless with zero users -- solve that cold start first.
- Find the smallest atomic network that can sustain itself before scaling anything.
- Don't launch to a whole market; ignite one tight cluster of users first.
- Hard-side supply, like drivers or hosts, usually matters more early than raw demand.
- Growth eventually hits a ceiling -- market saturation or a broken loop, not bad luck.
- Network effects only become a real moat after they clear a tipping point.
- Copy the atomic network formula city by city, campus by campus, building by building.
- A feature advantage erodes fast; a genuine network effect compounds and defends itself.
- Escape velocity looks organic, but it's usually engineered deliberately.
- The same five-stage arc shows up across marketplaces, social apps, and B2B tools alike.
Top 3 Quotes from The Cold Start Problem
"Network effects are the closest thing to magic in the business world."
Andrew Chen, The Cold Start Problem
"The cold start problem is the central problem that all networks must overcome."
Andrew Chen, The Cold Start Problem
"An atomic network is the smallest possible network that can survive on its own."
Andrew Chen, The Cold Start Problem
Frequently Asked Questions
Is The Cold Start Problem worth reading?
Yes, if you're building or growing a marketplace, social app, or any product where value depends on other users showing up. Chen's five-stage framework is the clearest map of how these products actually tip from zero to unstoppable.
What is the main idea of The Cold Start Problem?
Network products are worthless with no users, so you don't launch to a whole market. You find and stabilize the smallest 'atomic network' that works on its own, then copy that formula outward until you clear a tipping point and build a real moat.
What is an atomic network?
The smallest group of users -- one apartment building, one friend group, one college campus -- who get enough value from each other to keep using the product even at tiny scale, before it has any broader network to lean on.
Does The Cold Start Problem apply to non-network businesses?
Not much. If your product's value doesn't increase with more users interacting with each other, the atomic network, tipping point, and moat concepts mostly don't transfer -- this book is written specifically for network-effect businesses.
Ready to read it?
Get The Cold Start Problem on Amazon
