Quantifying Facebook's problem with fake brand likes

By James Robinson , written on February 17, 2014

From The News Desk

Last week Derek Muller, an Australian TV presenter and popular YouTube video blogger, caused a stir with a long video post detailing his argument that the genuine "likes" brands pay Facebook for closely resemble the shoddy, fake "likes" that can be purchased from "click farms." I copied the steps Muller laid out and found exactly the same thing. I could promote my own fake page to people who seemed suspiciously inauthentic.

The issue raised was whether Facebook was profiting from paid page like campaigns that were turning up fake results, driving down engagement for brands from these pages and forcing them to pay more money to reach the same audience.

With Facebook having announced $7.8 billion in revenue for 2013 at the end of January people were ripe to get worked up. The chance to predict Facebook’s demise is never passed up on and the tenor of many comments and social chatter was that the company was guilty of fraud.

Facebook commented after Muller’s video that its ads achieved the sorts of “real-world results” that would “not be possible with fake likes.” Further digging proved and disproved this claim. There was evidence of potentially fake traffic for the largest brand pages on Facebook, but its product has evolved so far past the basic ads Muller and I were discussing that maybe it doesn’t matter.

Looking at the 20 largest private company pages - ranging from Levis with over 21 million fans up to the Facebook fan page itself with over 117 million fans - a lot of the traffic is odd.

Only six brands out of the top 20 were engaging with more than one percent of their audience, with only one brand reaching more than two percent. Target, in last place, was engaging with just .05 percent. One post going out to its 22,722,008 fans had only 12 likes.

Even accounting for the drop in engagement brought about the change in Facebook’s news feed algorithm, it doesn’t seem like a good investment.

Many of these brand pages were most popular in the sorts of cities where ‘click farms’ could thrive: high density urban areas with a base of available low-paid workers. Coca-Cola and Facebook, the two largest corporate pages on Facebook with 79 and 117 million fans, were more popular in Istanbul, Turkey than any other city on earth. Samsung and Intel were hugest in Cairo, Red Bull and Levi’s in Taipei, Nickelodeon, Monster Energy, Victoria’s Secret, Converse and Disney in Mexico City.

But overlooked in last week’s furore was that Facebook ads have evolved well past the basic brand page.

Facebook doesn’t release a breakdown of how much of the $7 billion it made from advertising in 2013 comes from which streams. You need to look at third-parties for an educated guess. San Francisco-headquartered Marin Software manages $6 billion in marketing spend for its clients. Anil Channappa, Marin’s Director of Product Management, says that only eight to 10 percent of client spending on Facebook is currently going toward advertising a basic brand page.

“It’s still a lot of money,” Channappa says.

But it has fallen markedly. A year ago Channappa estimates that this figure was in the 50-60 percent range. This dip has coincided with Facebook becoming more sophisticated as an advertising platform. Campaigns can now be measured by clicks through to a website and actual sales. Advertisers can target Facebook users based off of pages they visited on a website, finely micro-target to a specific audience and sell to people based off actions their friends have taken.

Channappa says that online advertising has become too results driven and accountable for clients to be able to sink big money into maintaining a fake audience. Simply getting someone to like your page is an outmoded assessment of success. “Likes don’t mean anything without outcomes,” he says.

Using Marin Software as the vaguest of ballpark polls for how much money is spent on promoting brand pages and getting likes, puts it somewhere between a $560 to $700 million industry for Facebook. With examples like Muller’s and the company copping in its latest SEC filing to between 5.5 and 11.2 percent of its accounts being fake, it’s a safe assumption that a large audience on Facebook doesn’t mean what it used to and brands might not get the best value for money at that lower end. But it doesn’t mean Facebook’s business is broken, it’s just evolved passed a point where a like needed to mean something.