In company filings this week, social networking giant Facebook provided a customary laundry list of risk factors that could impact the company’s business going forward. A few of those risks have raised eyebrows in the technology, advertising, and investment communities, particularly Facebook’s own assessment that some 8.7 percent of profiles on the service are “fake.”
For years, Facebook has hammered home the message that it only wants profiles from real people and businesses on the service. Folks found to be maintaining a Facebook profile under a pseudonym, handle, or a name other than the one that appears on their birth certificate have often their profiles summarily deleted from the platform.
Is the fact that nearly one in ten profiles on Facebook are “fake” a tribute to Facebook’s efforts to keep it real and only permit profiles reflecting real-life people and businesses? Or does it represent a significant failure to keep scammers, ne’er-do-wells, and “fake people” off the service?
Facebook breaks down its estimate of “fake” users into a few categories. By far the largest is people who maintain two or more Facebook accounts, in spite of Facebook’s efforts to enforce its terms of service and limit individuals to a single profile. The company estimates that about 4.8 percent of its 955 million active accounts — or about 45.8 million users — represent such duplicate accounts.
More recently, the company has rolled out an automated profiling system that issues warnings to users the company believes might be operating multiple accounts. That system appears to be generating many false positives, particularly in cases where multiple people use the same computer.
Many folks who use multiple accounts do so for entirely innocuous reasons. They may have a personal account for themselves, but also operate a separate account for a business as part of their job or profession. Some people use multiple accounts as a way to separate professional and personal uses of Facebook. A so-called social media consultant might maintain a sizable network of connections on their personal profile, but have a second (quieter) profile where they engage with family and maybe a few guilty pleasures like Farmville and Words with Friends that they don’t want diluting the “branding” of their “real” personal page. Others might set up secondary profiles for more pressing privacy reasons: for instance, to avoid the scrutiny of bullies or an abusive former spouse.
Many of these uses of multiple accounts are essentially challenges to Facebook’s “real names” policy, and Facebook’s stock response is that users must stick to a single account and avail themselves of the services’ expanded (and famously fiddly) privacy controls to regulate what information is available to different sets of Facebook users. That’s a tricky line to walk, since Facebook is notorious for opting its users into sharing information by default, meaning the best way to try to keep something private on Facebook is probably not to do it via an account everybody knows about.
Facebook would probably argue other common-but-understandable uses for multiple profiles are simply misunderstandings of the platform. That social media consultant ought to set up a business account, rather than relying on a personal account. For some folks — retailers would be a good example — Facebook’s business accounts are a good option, but for professionals whose business is social media, they fall flat. Business accounts don’t show up in search, can’t send or receive friend requests, or make apps. So, many people who are very serious about social media undoubtedly operate multiple Facebook profiles.
Of course, an unknown (and essentially unknowable) number of secondary profiles are set up for less innocent purposes, such as harassing other Facebook users, or setting up an army of fake accounts in an effort to game services offered via Facebook, or generate fake “likes” and visits to bolster visibility and advertising revenue — a variant on “click fraud” schemes that still beleaguer the online advertising industry.
Facebook further estimates that about 2.4 percent (about 23 million) personal accounts are “misclassified” — users who have set up a personal profile for a business, video game avatar, or something like a family pet. Facebook permits “entities” like these to exist on the service, but only as Pages associated with a personal profile. In other words, there’s no official way for someone’s World of Warcraft character to have a Facebook page that’s independent of the player’s “real” Facebook profile. (Mark Zuckerberg’s dog, Beast, is an example of a Page associated with a personal account.)
…and the ugly
Facebook estimates another 1.5 percent (about 14 million accounts) are operated by bad actors: “undesirables” who have set up accounts for purposes like “like” fraud and spamming. Facebook primarily identifies these accounts by analyzing their accounts activity. The behaviors of an account created purely to generate spam via comments, wall posts, and similar actions are usually quite a bit different from a “real” new user, who often upload contact lists, search for friends, surf around a bunch of pages, fiddle with preferences, and (yup) play games.
Why “fake” matters to Facebook’s bottom line
Now that Facebook is a public company, it is required to disclose more details of its business operations and status to investors. Although Facebook’s most recent 10-Q filing contains an enormous lists of potential risks to the company, that’s utterly normal for this type of disclosure. If one were to evaluate companies solely by the risks disclosed in SEC filings, one would think the climate for all U.S. businesses is unequivocal sky-is-falling doom-and-gloom. The bulk of listed risks are (pardon our French) CYA tactics: Sometimes it’s a wonder companies don’t note a zombie apocalypse as a major future threat. In the event something goes horribly wrong and Facebook tanks, the company will be able to point back to these filings and tell investors they had fair warning.
Nonetheless, the bulk of Facebook’s revenues derive from selling advertising; so far in 2012, advertising sales account for 83 percent of the company’s revenue. The value of that advertising is tightly linked to the company’s ability to collate and analyze user information, then put those ads in front of the users who are most likely to respond. Facebook might be based on a culture of “real names,” but the financial reason it doesn’t want people running multiple accounts is that it interferes with the company’s ability to comprehensively profile a user. A user might be on Facebook eight hours a day, but if she uses eight different profiles, Facebook isn’t collating information about that user anywhere near as efficiently as it would like.
By disclosing that nearly nine percent of its accounts worldwide are, in its own eyes, “fake,” Facebook effectively diminishes the value of its advertising platform. Is it nine percent less effective? Probably not, particularly since a lot of those so-called “fake” profiles aren’t going to generate much useful information that would influence Facebook’s targeted advertising. But advertisers are going to read the disclosure as “Facebook says 83 million accounts are useless to us.” That may mean they’re going to bid less money — or less frequently — to put advertisements or sponsored stories on Facebook — and that can directly impact Facebook’s bottom line.
But do these “fake” accounts have an effect on diluting Facebook’s advertising reach? Perhaps. The BBC’s Rory Cellan-Jones recently tested Facebook’s advertising platform by setting up a page for a fake business, VirtualBagel Ltd., then buying advertising intended to get users to “like” the page. VirtualBagels offered no products, no coupons, no special offers, and had no significant content. What happened? Within 24 hours VirtualBagels had 1,600 “likes.” Moreover, a significant portion of those likes came from Facebook users purported located in Egypt, Indonesia, and the Philippines…and almost none came from the United States or the United Kingdom. Facebook told the BBC that there was “no significant issue” with fake profiles or their advertising platform, and that Cellan-Jones’ test was worthless.
But here’s a line from Facebook’s 10-Q filing that’s not a CYA statement: “The loss of advertisers, or reduction in spending by advertisers with Facebook, could seriously harm our business.”