“Network effects are the most important force in technology. They explain why so many tech markets end up as monopolies or near-monopolies.” — Andrew Chen
How does a thirty-person startup defeat Google? In 2006, Google launched Google Video and Google Base — products built with enormous resources, engineering talent, and distribution power. Yet YouTube (video) and Craigslist (classifieds) — scrappy startups with tiny teams — won those markets decisively.
This paradox sits at the heart of The Cold Start Problem. It seems to violate common sense. Resources should win. Brand should win. Distribution should win. But in networked markets, they often don’t — and understanding why is the beginning of understanding network effects.
The answer is simple in theory and profound in practice: the value of networked products grows with usage. YouTube had more videos. Craigslist had more listings. Google’s superior technology didn’t matter because the network was already elsewhere. Google was building an empty stadium; the other guys already had the crowd.
In traditional markets, resources compound into advantage. The company with the most engineers builds the best product. The company with the most capital outspends rivals on distribution. But in networked markets, connections compound into advantage — and connections are hard to buy.
A social network with 10 million users isn’t just twice as valuable as one with 5 million — it may be ten times more valuable, because it includes more of the people you know. This non-linear scaling is what economists call “network effects,” and it explains why technology markets so often end up with one dominant player.
A product exhibits network effects when its value to each user increases as more users join. The classic example is the telephone: one telephone is useless, two telephones enable one conversation, and a million telephones enable an essentially unlimited number of conversations.
But not all network effects are equal. Chen distinguishes between several types:
Robert Metcalfe, co-inventor of Ethernet, observed that the value of a network is proportional to the square of the number of its users. If a network has N users, it can support N² possible connections. Add one more user and the value jumps dramatically.
This mathematical reality is what creates winner-take-all dynamics in technology. Small advantages in network size become self-reinforcing. The larger network attracts more users, which makes it larger still. Rivals can’t just match features — they need to match the network, which requires matching the users first.
Google had resources and David had a sling. In networked markets, the sling works because:
Networks have to start somewhere. Even Google’s giant ad network started as a handful of advertisers and publishers. The cold start is universal.
Focus beats resources in early markets. A 30-person team fully focused on one market can outmaneuver a 30,000-person company with attention spread across hundreds of products.
Early network quality matters more than size. The right 1,000 users are worth more than any 100,000 users. A dense, engaged, relevant network is worth more than a sparse, disengaged large one.
Networks are hard to displace once established. When you’re already inside someone’s network — their friends, their professional contacts, their marketplace listings — switching costs are enormous.
Chen draws on dozens of examples throughout the book, but several pattern-matching themes emerge:
The very thing that makes networked products powerful — that they’re more valuable with more users — also makes them fragile at the beginning. An empty network delivers no value. This is the cold start problem: how do you create value for early users before the network exists?
Chen’s entire book is an answer to this question. But first, you need to understand the starting point: the atomic network.