The smallest possible network that can stand on its own
Every platform company faces the same paralyzing question on day one: why would anyone join if nobody's there? Andrew Chen spent years studying how the best ones solved it.
Network effects are among the most powerful forces in business — once a network reaches critical mass, it becomes nearly impossible to dislodge. But every network starts with zero users, and zero users means zero value.
Andrew Chen, who studied hundreds of platforms while at Andreessen Horowitz, identified the key to escaping this trap in The Cold Start Problem: the atomic network. This is the smallest possible group of users that makes the product self-sustaining — not big, not global, just stable enough that people keep coming back even when the broader platform hasn't taken off yet.
For Zoom's earliest version, a single business with a few remote employees constituted a working atomic network. For Wikipedia, it was a small band of editors passionate enough about a single topic to maintain a complete, accurate article. For Tinder, it was one college campus where enough students on both sides of the matching equation joined in a single week.
The insight is counterintuitive: don't try to grow everywhere at once. Find the smallest territory where the product actually works, make it work perfectly there, and let density create the conditions for expansion. A thin layer of users spread across a continent is weaker than a dense cluster in a single neighborhood.