In a world where a business’s social media clout is measured by the number of followers it has on Twitter and the number of “likes” on Facebook, it’s not surprising some game the system for profit. On Twitter you can buy followers — the going rate is $13 per 1,000 new followers. On Facebook brands can pay third party companies called “click farms,” usually populated by people in low-wage nations, who will “like” their public page for pennies a click.
Last night I came across a video from Veritasium, an Australian video blog with nearly 1.3 million followers on YouTube presented by Sydney-based physicist and TV presenter Derek Muller, who has pulled back the curtain to reveal that even real engagement brands pay Facebook for directly, isn’t that engaged at all.
Happenstance led Muller into investigating a phenomenon where page likes aren’t driven strictly by organic engagement and brand sentiment but rather these ‘click farms.’ In some ways it recalls the struggle Google has had with click fraud, where legions of people would click on ads to generate ad revenue for websites using AdSense. The problem has spread beyond ads and to Google offerings as disparate as Google Shopping and Google Play.
In May 2012, Muller says he received a series of $50 vouchers to promote his fan page for Veritasium. After spending these he noticed page likes skyrocketing while engagement plummeted. Today, Muller says Veritasium has 135,000 likes on Facebook. Because of his ad campaign, seventy-five percent of these are from India, Egypt, the Philippines, Bangladesh, Nepal and Sri Lanka. Only one percent of these fans have engaged with his content, where conversely engagement rates from other countries bounce between 20 and 60 percent.
Click farms like boostlikes.com, which Muller says pay people around $1 for every 1,000 things they like on Facebook, offer a brand chunks of likes on a platter; $120 will get you 2,000 international likes, $340 will get you 2,000 likes States-side. Facebook has nothing to do with this transaction and disavows click farms from its platform.
And anyway, Miller had paid money to Facebook, not a click farm, to be connected with genuine people. Instead, he found was purchasing a stack of fans that felt similarly fake.
To see if he could get around this, Muller created a fake shell for a fansite called Virtual Cat. Its welcome message read, “Here we’ll post only the worst, most annoying drivel you can imagine. Only an idiot would like this page.” He put $25 into an ad campaign that only targeted countries outside of the reputed click farm territory: Australia, the United States, Canada and the United Kingdom.
It didn’t make a difference. He imagined that no sane person would like his new page. Instead, it got 262 likes, all from his chosen territories, from mostly North American accounts that had liked thousands of disconnected things and seemed not to have engaged with his shell of a page at all.
Facebook has been waging an ongoing battle with fakery, deleting 83 million fake accounts in August 2012. Yet in its latest financial report to the Securities and Exchange Commission, it revealed that between 5.5 and 11.2 percent of its 1.3 billion accounts were fake. That’s between 67.7 million and 137.8 million people, which on the high end would make it the 8th largest country by population in the world.
So the vulnerabilities Muller exposes – with marketplace ads also being its lowest grade of advertising product – are not surprising. While the clickers may be faked, however, the implications are real, going to the heart of Facebook’s credibility as a marketing platform for business.
The idea that an ad will reach a living, breathing user, and that the clicks and likes an advertiser pays for are genuine, underpins the entire system of social advertising. Real engagement makes the wheels spin. Twitter and LinkedIn’s stock got hit bad when theirs fell. Facebook benefited when its rose. In the company’s recent earning announcement, Mark Zuckerberg crowed that engagement in December 2013 was up 59 percent on the year before.
As Muller says in his video, this technique isn’t just disingenuous in devaluing the worth of the whole system, it also inflates company revenue by creating a hamster wheel that requires you throw more and more money at to get the same results.
If you have a page and post something, the content goes to a small portion of your supporters. If a good amount of that first lot engages the content gets pushed wider again. If the people who like the page are clicking thousands of things with little regard, they won’t be paying attention to your content and you’re going to have to pay even more money to promote it to reach the same amount of people.
I’m always suspicious of taking news from YouTube at its word and having some experience with Facebook advertising in the past, I decided to replicate what Muller had done step-by-step and see for myself.
Step 1. Like Muller, I created a page I wanted to appear laughably awful and mocking toward any potential liker, something so vile that only someone so completely disengaged could like it. I chose the ‘The Good Luck Chuck Appreciation Society,’ starring Dane Cook and Jessica Alba and which may be the worst movie I’ve ever seen. Anyone who ventured to the page would see the greeting: “Good Luck Chuck is the worst movie ever created in the entire world and I would not respect you for a second if you liked this page.”
Step 2. I threw $20 behind the ad, targeting people like me in the United States, Canada, the United Kingdom, and Australia. By dinnertime my new page had 13 likes. All of the accounts appeared to be made by real, living people but had completely public profiles.
A couple of my new fans had liked a few hundred pages, but most had liked several thousand. There were strange geographic disconnections in the things they had liked that were readily apparent. My first fan from Arkansas City, Kansas had liked a Sonic restaurant from Oklahoma City, a pizza place in Eastchester, New York, a chain restaurant from Oahu, Hawaii and an independent eatery in New Mexico. Another fan from Beverly Hills expressed her appreciation for culinary delights in small towns in Washington, North Carolina, New Hampshire as well as San Francisco. And so it continued. Person after person was eating in a series of strange small towns spread from coast to coast.
The things these people previously liked didn’t make much sense either. One sports-minded fan liked the Detroit Pistons and its arch rival, the Chicago Bulls, as well as the Los Angeles Lakers and the Orlando Magic. A woman from the midwest liked both the Strong Men of God page and Urban Ink Magazine, two unlikely overlapping circles of interest.
Step 3. I was impatient and tweaked the formula, going past what Muller had done. What if I paid for a better ad? Would this get me a faster stream of dubious likes, or better quality engagement? By switching from a simple marketplace ad to a sponsored story, my ad would be run in the same old spots as it was before as well as the Facebook pages of the friends of people who had already liked it. My like count went from 13 to 39 in less than half an hour. None of these likes appeared more genuine than the first batch, but a couple of new fans wrote on my wall.
Fan 1: “Lol, what?”
Fan 2: “Good Luck Chuck is an awesome film. Dane Cook and Jessica Alba are the best. Hilarious and sh** and full of nudity, whats not to love?”
I felt slightly mortified when people wrote back to me.
Step 4. Muller deliberately targeted countries outside click farm territories, but I wanted to see what happened if I opened the gate. I threw $10 behind an ad that targeted people in the US, UK, Canada and Australia, alongside India, Bangladesh, the Philippines and Egypt. The ad was live for a quarter of an hour and my page received 66 new likes, none of them from the first four territories. It was the same story: each page had anywhere from a few hundred to many thousands of likes covering disconnected, unrelated things. Many of my new audience also weren’t using Facebook in English.
Muller’s video is filled with fascinating tidbits about the economy of empty likes driving Facebook. Last July, the Washington Post ripped the State Department after an Inspector General’s report said it had paid $630,000 to Facebook for two million likes, only two percent of which had ever engaged with State Department on the platform. A January 6 story on click farms also in the Washington Post found that both the Facebook Security and Google fan pages on Facebook have more fans in Dhaka, Bangladesh than any other city on earth.
I reached Muller by phone late in the evening. He told me from Vancouver that when he started advertising his page on Facebook, it became functionally useless to him as a platform. He had the same amount of followers on YouTube, but there fans interacted with his content like people, whereas on Facebook he felt like he had tens of thousands of ghosts following him. When he began to map the levels of engagement of his supporters by country he knew something wasn’t right.
“Why are the people liking these pages?” Muller wondered. A popular theory is that with Facebook cutting down on fake likes and click farms and making moves to protect genuine engagement with its valuable ad business, click farms have started liking a whole series of erroneous posts for free to confuse monitors and obscure traffic.
“I think Facebook has to know about this,” Muller says. “They have a plethora of data and the brightest minds working for them.”
In a statement to media yesterday, Facebook denied Muller’s hypotheses. The company has achieved the kind of “real-world results” that would “not be possible with fake likes.”
“He spent $10 and got 150 people who liked cats to like the Page. They may also like a lot of other Pages which does not mean that they are not real people – lots of real people like lots of things.” [Muller actually spent $25 for 262 likes.]
Except that it’s not that simple: 262 people liked Muller’s empty shell of a page that was aggressive toward any person that clicked on it. I promoted a page to more than 100 fans in less than two hours with the same tactic.
All of my new fans and Muller’s had liked hundreds of random things, atypical behavior displayed at an alarmingly consistent rate. “No one likes that many things,” Muller says. I surveyed a couple dozen Facebook friends. Two had amassed more than 100 likes. Most figured in the low tens. I am fairly indiscriminate with liking things and have amassed 90 likes in eight years on Facebook. One person I talked to with 125 page likes was embarrassed to have me call attention to it.
I’d paid $40 for my impromptu community of Good Luck Chuck fans. Like Muller, the least I could do would be check in on them.
“Hi there, Good Luck Chuckers. I created this page as an experiment in engagement in Facebook advertising for a story I am writing for PandoDaily. What brought you to like this page?”
Total silence. An hour later, five people had seen the post.
The system’s not totally corrupt. By upping my ad spend, a couple of people did actually engage with me. Facebook racked up $7.8 billion in revenue last year, it’s not all tainted fruit. But Muller’s hypotheses shows how the system can be gamed.
Because no matter little you paid for a like, if the social intent isn’t there, you paid too much.