Is Facebook the AOL of the 2010s? A Skeptical Examination of Social Media Network Effects.November 28th, 2018 · 34 comments
In economics, a network effect is a positive benefit created by a new user buying a product or joining a service. In the context of computer networks, these benefits are commonly believed to scale quickly with the number of users.
In technology circles, perhaps the best known instantiation of network effects is Metcalfe’s Law, named for Ethernet co-inventor Bob Metcalfe, who was likely inspired by similar theories developed at Bell Telephone in the early 20th century.
This law concerned the value of the Ethernet network cards sold by Metcalfe’s company 3Com. It states that given a network with N users, buying one additional Ethernet card provides you with N new possible network connections (e.g., from the new card to each of the N existing users).
It then follows, roughly speaking, that the value of N network cards grows as N^2 instead of N. Once a network achieves a certain critical size, therefore, the value it returns will quickly begin to far exceed the cost of joining it, creating a powerful positive feedback loop.
Metcalfe’s Law is incredibly influential in Silicon Valley, where it’s often applied to justify the monopoly status of the social media conglomerates. If a network like Facebook has over a 1,000,000,000 users, the law tells us, then its value to users grows as (1,000,000,000)^2 — a quantity so vast that any attempt to compete with this giant must be futile.
It’s widely believed among many Silicon Valley types that this calculus helps explains the lack of venture capital investment in new social media start-ups in recent years. The power of network effects in this sector is unimpeachable.
But should they be?
I’ve long harbored suspicions about how network effects are referenced to justify massive social media conglomerates.
If you examine the canonical examples of these effects, such as 3Com’s Ethernet cards or Bell’s telephones, you’ll notice that joining the network in question is the only way to connect to other people in that general manner.
In 1908, if you didn’t own a Bell telephone, you couldn’t talk in real time to people over distance. In 1988, if your computer didn’t have an Ethernet card, it couldn’t connect to other devices in your office. In these scenarios, buying the relevant product shifted you from completely disconnected to massively connected.
Social media, however, is different.
In 2018, joining a network like Facebook enables you to connect with or monitor the status of people you know using digital networks. Unlike telephones or Ethernet cards, however, you don’t need a private network like Facebook for these benefits. Both the Internet and SMS, among other technologies, already provide many different tools, protocols, and services for connecting and disseminating information digitally.
Case in point: I’ve never had a social media account, and yet I constantly enjoy connecting to people, and posting and monitoring information using digital networks.
So what then exactly do massive social media platforms like Facebook provide? A more honest answer is that they offer a more convenient experience than the wilder, less centralized social internet, but not something fundamentally unique.
There’s value in convenience, but not Metcalfe’s Law level, dominating value. In some sense, Facebook is to the social internet today what AOL was to the world wide web in the 1990s — a walled garden that provides a gentle on-ramp to the capabilities of a more exuberant decentralized network roiling beyond its boundaries.
These are thoughts I’m in the early stages of developing, so I’m interested in any pointers you can share about people smarter than me exploring similar ideas. But it increasingly seems to me that social media giants like Facebook offer at best network enhancements to its users, not the mythical network effects that helped make the monopolies of past eras so inescapable.