Network Effect

Strategy · 3 min read

A network effect is when each new user makes the product more valuable for every existing user. It is the rarest and most powerful moat in modern business. It is also the most over-claimed.

Each new node makes every existing node more valuable.

The first telephone, famously, was useless. There was nobody to call. The second telephone made the first one slightly valuable — now there was exactly one possible conversation. The third made it slightly more useful, the hundredth dramatically more, and by the time there were a million telephones in service, the value of being on the network was so high that staying off it was almost impossible.

This is a network effect. The product gets more valuable to each existing user as new users join. The relationship is not additive — it is multiplicative. A network of ten users has roughly 45 possible connections between them. A network of a hundred users has roughly 4,950. The mathematical name for this is Metcalfe's Law, and while the precise formula is debated, the underlying truth is not: the value of a network grows faster than the number of its members.

Network effects are widely considered the most powerful moat in modern business for one specific reason: they get stronger as the business gets bigger. Most moats stay the same width or slowly narrow as competitors find ways around them. Network effects do the opposite. Every new user widens the moat, because the gap between "what we offer" and "what a competitor without our network could offer" grows with every signup. By the time a competitor tries to launch a similar product, the gap is so large that even a technically superior product cannot attract users — because the network is the product.

This is why Facebook beat dozens of objectively better-designed social networks. Why eBay's seller and buyer base became so dense that competitors could not replicate it. Why LinkedIn became the only place professional networking actually happens online, even though several rivals tried. The product was not "the best social network," "the best auction site," or "the best networking platform." The product was access to everyone who was already there. And the only way to compete with "everyone is already there" is to get there first.

There are several distinct types of network effects, and they behave differently.

Direct network effects occur when each user makes the product directly more valuable for every other user. Telephones, email, messaging apps, social networks. The value comes from being connected to the other users.

Indirect network effects occur when one group of users attracts another group, who then attract more of the first group. Marketplace platforms like eBay, Etsy, or Airbnb have indirect network effects — more buyers attract more sellers, and more sellers attract more buyers. The two sides reinforce each other in a self-strengthening loop.

Data network effects occur when each user contributes data that improves the product for everyone. Google Search becomes more accurate as more people search. Maps applications become more useful as more drivers contribute traffic data. Recommendation systems become smarter as more users rate, click, and purchase. These effects are subtle but enormously powerful — the moat is invisible to users, but it is real.

Local network effects occur when the effect only matters within a defined geography or community. Restaurant review platforms like Yelp need density within each city — having a million reviews nationally is useless to a user in Tulsa if there are only three reviews of Tulsa restaurants. This is why local network effect businesses often grow city by city, building density before expanding.

The reason network effects are so over-claimed is that the term has become a shorthand for "we expect to get big." Many founders describe their business as having "network effects" when they really mean "we hope a lot of people will use this." Real network effects require a specific structural property: each new user must make the product objectively more valuable for existing users, not just contribute to the company's revenue. A subscription business with a million customers is not a network effect business if those customers gain nothing from each other's presence. It is just a large customer base.

The test is simple. Does adding a user create value for the existing users, automatically and without effort from the company? If yes, the business has network effects, and the strategic implications are enormous. If no, the business may still be excellent, but it is competing without one of the strongest moats available. Both are valid. Only one is what the term actually describes.