What is the value of Facebook "likes"? Not much, if a new video released by the YouTube channel Veritasium is correct.

Veritasium channel owner Derek Muller drew on the 2012 work by BBC technology correspondent Rory Cellan-Jones, who'd learned that advertisers were complaining about low return rates from campaigns on the Facebook ad platform. Cellan-Jones decided to investigate the merits of Facebook (NASDAQ:FB) “likes” by setting up a fake Facebook page called “VirtualBagel.” After spending $60 on advertising promotions, Cellan-Jones found that the majority of his likes came from Egypt, Indonesia and the Philippines. When he shifted his ads to target the U.K., the number of likes to his page plummeted.

Facebook explained one case to the BBC, claiming that a particular ad campaign was positioned as “scattergun advertising,” meaning it didn’t target a particular group.

Muller used Cellan-Jones’ experiment as a prompt for his video piece, which dissects the anatomy of likes obtained by using Facebook’s ad platform.

After using free advertising credits from Facebook, he found that his Veritasium page skyrocketed from 2,000 likes to 70,000 likes over a few months, about the same number of subscribers to his YouTube channel at the time.

While the number of likes to his Veritasium page grew dramatically, actual user engagement was significantly lower from the majority of the users who liked the page. Muller found these high-likes/low-engagement visitors came overwhelmingly from Egypt, India, the Philippines, Bangladesh, Indonesia, Nepal and Sri Lanka.

So what's with these fake likes? According to Muller, Facebook’s algorithm seeds a post to a small percentage of users who like a page, and increases or decreases exposure based on user engagement.

Even after refocusing a campaign to exclude the countries where the majority of low-engagement fake likes were coming from, Muller found his page still being liked by fake accounts.

Jaron Schneider shares his hypothesis on this matter in an article on The Next Web, explaining that these click-farm accounts "like" a diverse set of pages in order to hide from Facebook's fraud-detection algorithms.

Muller thinks Facebook does not remove fake likes from Facebook pages because that would force them to admit that a significant amount of ad revenue was generated from non-genuine user accounts.

Watch Muller’s full experiment in the video above.