Network Effect

A network effect is a property of a good whose value to any one user increases as more people use it. A direct network effect runs between users of the same kind — a telephone line, a WebRTC connection — where each additional participant is directly reachable by every other. An indirect network effect runs across two distinct groups, as in a marketplace where more buyers attract more sellers and more sellers attract more buyers, so the value compounds sideways rather than in a straight line.

New entrants face a chicken-and-egg problem: a network with no users offers no value, so the hardest work is getting the first cohort to join a thing that is not yet worth joining, the same threshold logic behind Collective Action more broadly. Past some point adoption can tip — growth becomes self-reinforcing and a market consolidates around one or two winners — and once tipped, the outcome is sticky: Path Dependence keeps a network dominant well past the point where a late entrant might otherwise be better on the merits.

A popular version of the idea, its Naming borrowed from Robert Metcalfe, claims a network's value grows with the square of its user count. Treated as Big-O Notation-grade analysis this overstates the case: most added users interact with only a few others, not with everyone, so the real curve is far flatter than n² even where the effect is genuine.

Network effects are often confused with switching costs, and the two are related but distinct sources of lock-in. A network effect is a positive Externality — each new user makes the good more valuable to everyone already there. A switching cost is private: the effort of moving your own data, habits, or relationships elsewhere, which would exist even in a network of one.