Network effects. Perhaps no other phrase can get a VC's pulse higher.
They're the holy grail of online business.
What is a network effect? It's what happens when the value of a product to one user depends on how many other users there are, as economists Carl Shapiro and Hal Varian put it.
Examples include Microsoft Windows and the phone network. Windows is valuable because most other software is made for Windows, which makes more people buy Windows, which makes more developers build their apps for Windows, and so on and so forth in a virtuous circle.
Network economics online get people's hearts racing so much for two reasons: first of all, because the internet is at its base a communications network and so network effects tend to happen more there, and second of all, it's one of the few strong barriers to entry in a market where there are so few.
Or is it?
It's taken as a given that network effects online are a magical barrier to entry, but are they?
One reason Facebook is so valuable, we're told, is because of its huge network effects which make it unstoppable and undefeatable.
But Friendster had network effects. So did MySpace. Founding Facebook President Sean Parker says MySpace lost to Facebook because of its gross incompetence. Fair enough.
Except that's not the only way network effects businesses can lose. One way that network effects can be defeated is through what we'll call "verticalization."
Craigslist is perhaps one of the best network effects businesses: the reason why everyone goes there is because everyone is already there. Plenty of people have pointed out how awful Craigslist's design can be, how many things are wrong with it, and yet plenty of well-funded startups that have tried to take Craigslist on frontally with slicker offerings have foundered.
And yet... And yet, Craigslist's traffic seems to be plateau-ing. Why? This graphic by VC Andrew Parker shows why:
While no service has been able to defeat Craigslist head-on, plenty have built "niches" in specific verticals, with a more tailored offering, and now Craigslist seems to be stalling.
And some of these "niches" are big: Etsy, AirBnB and Ashley Madison are huge businesses.
Could the same thing happen to Facebook? We would argue it already is.
Plenty of apps are taking specific use-cases of Facebook and turning them into full-blown services.
The best example might be Twitter, which really takes the "status update" feature on Facebook and turns it into its own service, with its own identity and use case. (That's not how Twitter was invented, but that's what it largely does in practice.) Now most of the people we want to know about are on Twitter, and we barely ever use Facebook. We're a tech blogger so we're not especially indicative of the Normals who decide what services get huge or not, but Twitter's huge growth seems to indicate we're not alone.
The other big usage of Facebook is sharing and looking at photos, and there are also plenty of startups tackling that opportunity, with the fastest-growing being Instagram. With cameras and internet connections in our phones, it's possible to make apps for photo sharing that are much more tailored to the vast majority of Facebook photo sharing use cases than Facebook itself.
If you can see your friend's pictures on Instagram and your friends' status updates on Twitter, why would you visit Facebook?
The point here isn't to say that Facebook is going to crash tomorrow or that Twitter will "kill Facebook" or any of that crap. The point is to say that online network effects are probably overhyped.
It's worth pointing out here that the two most successful online companies--Google and Amazon--don't benefit from network effects, but from economies of scale.
Google search gets better the more people use it because it gives Google more data about what people search for and how, which helps them refine their algorithm. But Google search isn't a network. Amazon also benefits from economies of scale like any other retailer: their size allows them to get better prices from suppliers and operate more efficiently, which allows them to pass on savings to consumers which makes them even bigger. These are scale externalities, not network externalities.
The answer is that it's not black or white. Facebook was helped by superior execution AND network effects in highly dense and active social networks called college campuses.
But it shows the limits of network effects as well. What the story of Craigslist and Facebook shows is that online network effects are strong barriers to entry to FRONTAL competition but not to LATERAL competition. And lateral competition can be just as dangerous--perhaps even more, because it's easier for the incumbent to miss.
If Facebook had focused on bands and Los Angeles party goers in 2005, they would have hit a brick wall because MySpace had them locked up with strong network effects. Instead Facebook targeted a population that was less into MySpace, attacking laterally--and won.
Paradoxically, to pump up the story of their invincibility, Facebook has to play down their superior execution to make people believe they won thanks to the magical power of network effects. In a talk at the Web 2.0 Summit, Sean Parker says colleges were the last demographic MySpace hadn't touched, but that's not true in this writer's experience. In 2007, the Ivy League college students we knew were on MySpace and Facebook (and had been on Friendster before that). They just used Facebook more because they liked it better; in other words, Facebook won through superior execution and lateral attack at least as much as network effects.
This overblown faith in network effects can lead investors and analysts to make mistakes.
In a now-paywalled PEHub interview, a parnter at Founders Fund, a big Facebook investor, said they had the opportunity to invest in Zynga but didn't because they thought social games were a hit-based business without strong network effects. They were right. It just happens that social games are a very profitable business and that Zynga was able to parlay its first-mover advantage into economies of scale by spending more than anyone to acquire players for its social games. To be fair, hindsight is 20/20 and every top VC has turned down at least one startup that went on to become a huge success. The point here isn't to pick anyone, it's to show that people may pay too much attention to network effects.
Groupon, we're told, has no barriers to entry and therefore can't be worth so many billions of dollars. It turns out that building a business with 50 million subscribers in dozens of countries selling to thousands of merchants is really really hard and that the daily deals market, like e-commerce, has low barriers to entry but high barriers to scale. Groupon is still huge and growing fast, and so is the second-comer LivingSocial, but well-funded well-staffed followers like BuyWithMe are nowhere. First prize is a Cadillac, second prize is a set of steak knives, third prize is you're fired. It's no coincidence that Groupon came out of Chicago and not Silicon Valley, where the belief in network effects is the strongest.
Gilt Groupe, which should do $500 million in revenue this year, also doesn't have network effects, and that doesn't seem to stop them.
The three startups we've mentioned happen to be the three fastest-growing startups (both in terms of revenue and userbase) we know of since 2007. None of them have network effects. All of them have economies of scale.
Again, the point here isn't to say that there is no such thing as network effects or that they're not great. There is, and they are. But they may be overhyped.
Sean Parker says: "Companies that harness the power of networks will dominate the future of the internet."
That may be true.
We're just not so sure.
Don't Miss: The 10 Huge Tech IPOs Coming This Year →
For the latest tech news, visit SAI: Silicon Alley Insider. Follow us on Twitter and Facebook.
Join the conversation about this story »
See Also:
- Guess Who The Second Biggest Search Engine In Latin America Is
- Google's Perennial Hiring "Problem" -- The SAIcast Discuss
- This Is Who Gilt Is Up Against
How Strong Are Network Effects Online, REALLY? (EBAY, GOOG, AMZN)
