Also read our counter-argument: Facebook is a goldmine in the making
It’s hard to explain Facebook’s current business model, but here’s a reasonable analogy: Facebook is a large, inefficient engine for transforming electricity and programmers into a down-market place to sell low-value advertising. Most of the numbers in its recent earnings report indicate that the company is in a race against various forces merely to maintain its current advertising revenue, and there is little to indicate that the company is going to come up with any fundamentally new sources of revenue anytime soon.
Running just to stay in place
Here’s the most important number in Facebook’s yearly report: The average amount of money it makes on every user.
Notice that this amount barely went up over the past year (compare fourth quarter of 2012 to 2011, and the whole year of 2012 to 2011). Meanwhile, Facebook’s expenses are way up—82% in the past year—and the company is projecting that they’re going to go even higher in 2013. As a result, Facebook only made $64 million in profit in its most recent quarter. Much hay has been made of the fact that mobile advertising is now 23% of Facebook’s advertising revenue, when in the previous quarter it was only 14%. But mobile ads are still worth much less than desktop ads, so now that more people are accessing Facebook on phones than on PCs, this shift is exactly what we should expect. It doesn’t mean Facebook is doing any better than it was—it merely means that Facebook is doing what it takes to see its revenue not shrink as people spend more time on phones and tablets and less on computers. But, you might say: So what if Facebook’s advertising revenue isn’t growing significantly, as long as the site continues to add users? Here Facebook has an even bigger problem—as the site reaches saturation in rich countries, almost all its growth is coming from developing markets, and the company simply can’t make as much money from those users.
If not advertising, then what?
Despite being a pretty innovative company, Google makes almost all of its money by giving users things for free and selling advertising against them. It’s hard to see how it will ever be any different for Facebook. Facebook’s experiment in charging its users to do something—otherwise known as Facebook Gifts—has not made serious money. Pundits often make the argument that “surely with all that personal data, Facebook must be able to do something.” A great many startups have launched in the hope that if only they get enough data about their users, they’ll find a way to make money from it. But looking at the history of Facebook and related companies like Google, there’s little to back that up. Most of Facebook’s engineering resources appear to be devoted to simply keeping up with users, giving them ways to access Facebook on every newfangled device. It’s hard to argue that the company has done anything particularly innovative since its founding; all the updates to Facebook since then come down to tweaks to the core experience of reading through a stream of updates from your friends. (Or, since the recent advent of Facebook’s Graph Search, finding new ways to sort them.) Unlike traditional web search, which is something we often do when we’re looking to spend money, time spent on Facebook does not appear to be time when we’re particularly open to advertisers’ messages. Facebook has had some luck selling ads to those who make mobile apps, but being a parasite on Google’s and Apple’s app economies is never going to be a winning business.
Revenue growth, yes; profits, no
For Facebook to grow significantly, it’s got to figure out how to increase the amount of money it makes on every user, and not by an incremental amount, but some multiple. All without committing significantly more servers and programmers. Facebook is currently making a little more than $5 per user. It needs to get to $10, $20 or more per user while keeping its costs about where they are if it’s ever going to justify a higher stock price. Facebook is not going to get there with current advertising models, and there’s little evidence it has any better ideas.