Wall your product into one tightly bounded population and win it so completely that the density of real connections inside pulls each adjacent group in on its own.
Facebook is a social network that launched on February 4, 2004 as "Thefacebook," a Harvard-only directory built by sophomore Mark Zuckerberg. Its early story is a study in restraint: rather than chase the whole internet the way Myspace and Friendster did, it deliberately walled itself into one campus, then let the density of real connections inside that wall pull each new campus in. The two patterns below are really two halves of one move, winning a narrow market so completely that the network became the product.
Beachhead: winning Harvard before anywhere else
The problem. In early 2004 the social web was already crowded: Friendster had millions of users and Myspace was surging. A new entrant that opened to everyone would be a thin, empty product competing against incumbents with huge head starts and no reason for anyone to switch. Zuckerberg needed a market small enough to dominate on day one.
The approach. He restricted Thefacebook to a single, tightly bounded population: Harvard undergraduates, gated by a harvard.edu email address. The product mapped directly onto a need students already had (Harvard had been slow to produce an official online "facebook" of the class), and the .edu gate meant every profile was a real, verified classmate rather than an anonymous handle.
How it solved it. Adoption inside the beachhead was near total and fast. Within roughly 24 hours between 1,200 and 1,500 students had registered, and within about two weeks close to two-thirds of Harvard's roughly 6,600 undergraduates had profiles. Owning one campus that completely gave Facebook something the incumbents could not claim: a market where it was not the challenger but the default.
Network Effect: exporting campus density, one school at a time
The problem. A social network is worthless if your friends are not on it, and total user count is the wrong metric: a million scattered strangers is less useful than a few thousand people you actually know. Facebook had to grow without diluting the very density that made Harvard work.
The approach. Instead of opening the floodgates, Facebook expanded school by school, launching a fresh closed network for each new campus. In March 2004 it opened to Stanford, Columbia, and Yale, then rolled out across the Ivy League and other colleges, each one a self-contained, .edu-gated community that repeated the Harvard pattern before the next was added.
How it solved it. Because every new campus arrived pre-loaded with a real friend graph, each rollout hit critical mass almost immediately, and the growth compounded: by June 2004 more than 250,000 students across 34 schools had joined, and by December 2004 the service passed 1 million users. Each new member made the network more valuable to everyone already inside their school, which is exactly why the campus-by-campus method beat opening to everyone at once.