Byrne's Blog

Metcalfe Was a Pessimist

The value of a network grows at roughly the square of the growth rate of the users. That makes sense for telephone networks—add one more user to a network, and the number of new connections available goes up by the number of existing users.

Classically, this breaks down because the first people to add it get the most out of it. Maybe a phone was crucial for the first few people to use it—but the next phone sold today is probably going to replace an old phone, replace borrowing somebody else’s phone, or complement an existing Skype connection.

But Metcalfe’s law can break in the opposite direction, and I think we’re seing that in social media. In fact, I think we’ll see a lot more of it.

The best way to consider this breakdown is to look at Twitter. Twitter grows fast, and the bigger it gets, the more valuable it is to the next potential user. Classic Metcalfe. But services that use Twitter get a double pull: the more users each service has, the faster it grows. The more users Twitter has, the faster the service grows. It’s Metcalfe squared! It gets better: some twitter services are themselves network-based. Stocktwits, for example, uses the Twitter discussion platform to create its own investment discussion platform. There’s (theoretically) nothing stopping someone from creating a new service on Stocktwits.

(Or, rather, the only thing stopping someone from doing that is that Stocktwits might buy them.)

There are two caveats to this: first, a company whose growth is hyper-exponential will spend more of its time in the boring part of the “hockey stick” chart. Compare this to this; in the second one, the growth rate ends up being higher, but it stays at basically zero for longer. The other problem: a service on a platform (or a platform on a platform) will reach the limits of its growth faster, simply because it’s in a smaller market.

(That doesn’t mean the growth stops there. Bit.ly has a dominant market share on Twitter. But I’ve started using it in emails, just because it’s a more convenient way to track clicks.)

What all this rapid, compounding growth means is that a service with a percieved advantage will get an almost instant monopoly. The barriers to entry on Twitter are low, but the cost of attention is high—anybody can start a service that their ten best friends will use, but the vast majority of those services will never crack the hundred-user mark. Once someone is established, the Metcalfe-squared formula takes over, and they start growing like crazy (and helping Twitter to keep growing like crazy).

In the future, we’ll probably see all but one of the URL shorteners die: if they’re worth the time it takes to make them, they’re not worth the money it takes to maintain them, unless they’re huge. There will probably be one company (maybe the same company) that packages and sells the data from them. There will only be one sponsored tweets company, one “mainstream” Twitter client (and plenty of smaller ones that focus on, e.g., tracking emerging trends, following the discussion about a billion-dollar brand, or trading stocks). The energy put into starting up competitors and knockoffs will go into starting completely new businesses, most of which will fail, some of which will (surprisingly quickly) grow to dominate the markets they’ve created, until those markets are rendered obsolete by technology changes.

The future of the Internet is million tiny, temporary monopolies.