Sunday, February 25, 2024

Facebook’s and Twitter’s Dirty Little Secret

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Information recently revealed from a variety of sources suggests that a huge percentage of users on social sites like Facebook and Twitter aren’t actually human.

Millions upon millions of social accounts are either created by “bots” or have been created for some other purpose than social networking.

And I believe the problem is much worse than is being reported.

The Trouble With Facebook and Twitter

In a company filing earlier this month, Facebook reported that 83 million Facebook accounts are fake users.

A social media management company called Status People created an online tool called Fakers, which estimates the relative percentages of anyone’s Twitter following that’s fake, inactive and good.

For example, Twitter’s number-one user, Lady Gaga, supposedly has about 30 million followers. But Fakers says 71% of those little monsters are either fake or inactive. In reality, Twitter’s number-one user has fewer than 9 million actual followers, if Fakers is accurate.

Personally, I doubt the real number of fake Facebook and Twitter accounts are actually that low. The reason is that in neither case are the methodologies used to reach these numbers transparent. And I think it’s nearly impossible to detect a well-managed fake account.

I’m certain that I could create a fake Facebook or Twitter account, visit it weekly for ten minutes of activity, and no bot or machine or survey could ever determine that it’s a fake account.

In former Defense Secretary Donald Rumsfeld’s lingo, detection software can identify only the “known unknowns” and is clueless and powerless in the face of the “unknown unknowns.”

I think the percentages listed above are minimums, and that actual percentages of fake accounts is much higher.

The Fake User Industrial Complex

Where do fake user accounts come from? They come from many sources and serve many purposes. Here are some of the major sources for fake user accounts.

Fake user market. At least 58 companies currently appear online to sell fake users. Pay one of these services less than $20, and your Twitter following will go up by 1,000 users in 24 hours. A million new fake followers will cost you more than $1,300 and take a few days. You can also buy followers for Facebook, YouTube, Instagram and others.

Spammers. People hoping to sell something also create fake user accounts in large numbers.

Trolls. Most online trolls have multiple fake accounts, which they use to mess with people.

Astroturfers. Anyone who wants to fake a grass-roots campaign of some kind can higher low-cost workers to create very large numbers of fake accounts, which they can then use to leave opinion comments, creating the illusion that an idea is more popular than it appears.

Many politicians in the US and in other democracies are almost certainly doing this, too. For example, Barracuda Labs reported that within a period of 24 hours during the month of June, the Twitter followers of presidential candidate Mitt Romney grew by 116,992. Here’s the kicker: One quarter of those “users” had joined Twitter within the prior four days.

Regular people with multiple accounts. Some “fake” accounts are merely created for a variety of purposes. I personally have created about a dozen Twitter accounts over the years for various projects. I never closed any of them.

Abandoned accounts. A huge number of social media accounts are created, then abandoned. They’re not “fake” users, but nobody’s home. A user is only a user if he uses.

Developers. Developers often create multiple fake social media accounts so they can test an app or some other project.

Criminals. Identity thieves, drug dealers, terrorists — you name it. It’s not very smart to coordinate or execute various law-breaking activities from a real social media account. It’s easy enough to just create a new fake one, and use that from a public terminal.

Social Network Founders. Reddit founder Steve Huffman admitted in an online computer science class that Reddit was launched with a large number of fake users. The two founders wrote a capability available only to them whereby adding a user name automatically registered that username as a real user.

For the first few months, Reddit appeared to have hundreds or thousands of users, but they were mostly just the two founders. This is probably pretty common, especially for apps that interface with social networks (as many now do).

The Fake ‘Like’ Industry

BBC reporter Rory Cellan-Jones tested the value to businesses of Facebook advertising and “Likes.” He created a fake company called VirtualBagel, which offered a bogus service: downloadable electronic bagels.

He bought $60 worth of Facebook advertising, and got 3,000 “Likes.”

Upon closer inspection, however, he found that many of the “Likes” were made by fake accounts, and by serial “Likers” — for example, one fake account in Egypt that had also “Liked” 3,000 other brands.

He also noticed a wildly disproportionate number of “Likes” coming from certain countries — Egypt, Indonesia and the Philippines.

A bioscience consulting company called Comprendia discovered something similar. Some 54% of “Likes” for their a company came from India, Mexico, Indonesia, and Portugal, which collectively represent only 15% of Facebook users. They estimate that about 40% of the Facebook “Likes” they got were from fake accounts. Facebook recently announced a new initiative to automatically remove fake “Likes,” which it says come from “malware, compromised accounts, deceived users, or purchased bulk likes.”

But this program appears to me to be more of a PR stunt than an effective elimination of fake “Likes.” For starters, Facebook estimates that “on average, less than 1% of Likes on any given Page will be removed.” However, the actual percentage of fake “Likes” is surely much higher, possibly as high as 40%, as Comprendia estimates.

The Fake User Problem

I believe the fake social user account problem is going to become a far bigger challenge than most people believe.

It’s an arms race, comparable to the contest between email spammers and spam filter companies. As companies find new ways to detect and close fake accounts, the fakers will find new ways to open undetectable ones.

Fake accounts will increasingly sway public opinion, spam social streams with sophisticated and hard-to-detect soft pitches for products and services, socially engineer people for criminal purposes and much more.

The fake social user account problem will be devastating for Facebook and Twitter, which will have to work hard to convince advertisers that users and “Likes” and tweets are living, breathing human beings, and not bots or scammers.

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles