Researchers Call Out Twitter Celebrities With Suspicious Followings

Security researchers recently shed a bright light on the multimillion-dollar underground market for fake Twitter followers. Now, they are highlighting what they believe to be some of the market’s high-profile clientele.

In a follow-up to their earlier report, two Italian security researchers, Andrea Stroppa and Carlo De Micheli, call out Twitter accounts that added or lost a large number of followers in one day. Their list includes brands like Pepsi, Mercedes-Benz and Louis Vuitton; politicians like Newt Gingrich, Representative Jared Polis and Dmitri Medvedev, the Russian prime minister; and the rappers 50 Cent and Sean Combs, known as Diddy.

Social media experts say there are several reasons why Twitter users would want to acquire large volumes of Twitter followers. For some people, it simply feeds the ego. For people and brands, a large Twitter following or Facebook fan base helps increase their visibility. If followers are constantly clicking on links to a brand’s landing page, it also lifts the brand’s position in Google’s search results.

“It’s natural for brands to want to build their Twitter and Facebook accounts because they are constantly looking for ways to expand awareness of their products and services and expand opportunities to create consideration of their products compared to others,” said Susan Etlinger, an industry analyst at the Altimeter Group. “The more content they publish, and more links people click on, the better their Google search results. ”

Ms. Etlinger added that there was corporate pressure to justify a company’s investment in social media. Twitter followers and Facebook “likes” offer some of the few seemingly hard metrics in an otherwise squishy realm of social media.

“Many brands struggle to measure the top line value of social media,” Ms. Etlinger said. “So there is a thirst to show momentum in different ways, one of which is to show that the brand has a bigger audience today than it did yesterday.”

Some major brands have expressed skepticism about the impact of social media fans and followers. Last month, Coca-Cola, whose flagship brand has over 700,000 Twitter followers and more than 60 million Facebook fans â€" more than any other brand on Facebook â€" said a corporate study found that online buzz had no quantifiable impact on short-term sales.

Social media fans and followers tend to be volatile, brand experts say. They may start following a brand for a specific contest or sweepstakes, then drop off when the campaign ends. But Mr. Stroppa and Mr. De Micheli said the follower changes they charted were drastic enough to warrant suspicion that they were purchased.

For example, Mr. Polis, a Democrat from Colorado with 22,140 Twitter followers on his personal @jaredpolis Twitter account, gained, on average, 15 new followers a day for two years. Then, last July, he added 19,705 new Twitter followers. A few months later, he lost 13,332 Twitter followers in one day.

A representative for Mr. Polis denied he had purchased fake followers and said one explanation for the sudden gains and subsequent losses was a “follow-back” campaign. Last July, Mr. Polis said he followed a large number of new people from his personal account and asked them to follow him back. In March, he said he stopped following those accounts because he was becoming frustrated by the lack of relevant content in his Twitter feed. He now follows only 2,200 people. He surmised that a large number of followers may have stopped following him, too.

But Mr. De Micheli said that reasoning does not justify the sudden 19,705 jump in Mr. Polis’s followers in one day. Twitter’s own policy prevents users from following over 1,000 new people a day: “Every Twitter account is technically unable to follow more than 1,000 users per day,” Twitter says on its Web site. “Please note that this is just a technical limit to prevent egregious abuse from spam accounts.”

As for the sudden drop, Mr. De Micheli said it would be extremely rare for 13,332 followers to stop following the congressman on the same day. To have them stop following his account, Mr. Polis would have had to block 13,332 people manually, since the only way to stop an account from following you is to manually block it.

“Nearly the exact amount of followers ‘magically appeared,’ then disappeared,” Mr. De Micheli said. He suggested that the losses was more likely due to Twitter, which routinely deactivates accounts it deems fake.  He said Mr. Polis’s follower drop matched “what the typical ‘low quality’ fake followers acquisition-drop graph looks like” when Twitter deletes those accounts.

The researchers also call out Diddy, whose verified @iamdiddy account gained 185,399 Twitter followers one day last June â€" a 3,063 percent increase from the account’s average daily gain â€" and then inexplicably lost 393,665 followers one day last month, 6,504 percent more that his average daily follower loss.  A representative for Mr. Combs did not return a request for comment.

Likewise, 50 Cent lost more than 190,342 Twitter followers over the course of one day last January, a 5,370 percent jump from his average daily follower loss. A representative for 50 Cent did not respond to a request for comment.

Mercedes added 28,283 followers one day in October 2012 â€" a 20,992 percent jump from the brand’s average daily follower gain. A spokesman for Mercedes did not respond to a request for comment.

Likewise, Pepsi added 71,686 Twitter followers one day in November 2011 and has not added that number of Twitter followers in one day since. Before the bump, Pepsi followers trailed the number of people who followed Coca-Cola, which had more linear growth. After the bump, Pepsi’s followers surpassed Coca-Cola’s.

Jeff Dahncke, a spokesman for PepsiCo, said that the bump was because of promotional campaigns. “The spikes correlate with paid activations with Twitter â€" such as promoted Tweets â€" that were designed to boost our following around key brand activations,” Mr. Dahncke said, citing a Pepsi summer concert series and “X Factor” and N.F.L. promotions that year. “The followers are validated as real followers and are not fake accounts or bots.”

But the researchers note that a one-day gain of more than 70,000 followers because of a promotional campaign is unlikely. “The peaks are very high even through traditional advertising, and the shapes of the curves don’t really convince us,” Mr. De Micheli said, that the peaks are due to “traditional Web advertising.”

Mr. Stroppa said that a major Pepsi partnership with Twitter last year did not result in the same bump.  Pepsi’s followers looked particularly suspicious when compared to those of Starbucks, one of the longest-running major brands on Twitter. Starbucks regularly advertises on the service and pays to promote its posts. Starbucks’ biggest daily follower gain is 17,562, compared to Pepsi’s 71,700. (Sixty-six percent of Pepsi’s followers are legitimate, according to Status People, an online service that claims to decipher real followers from fake and inactive accounts.)

There are now more than two dozen online services willing to sell fake followers. Based on the number of fake accounts offered through those services â€" excluding overlapping accounts â€" Mr. Stroppa and Mr. De Micheli estimate that there are now more than 20 million fake accounts on Twitter. Those accounts can be sold to multiple buyers. At quoted rates, the two said a conservative estimate is that fake Twitter followers offer potential for a $40 million to $360 million business.

Twitter said discerning real accounts from spam accounts can be difficult. “Spam is a problem that faces the entire Web,” said Jim Prosser, a Twitter spokesman. “We have a variety of manual and automated methods that evolve over time for dealing with spam, and have even sued many of the most prominent spam organizations to keep them off our service. Users can also flag potential spam accounts for our review.”

Ms. Etlinger of Altimeter cautioned brands from dealing in fake followers. “There will always be people who try to game the system in every nook and cranny of business,” she said. “But brands should know that Twitter and Facebook are getting very good at weeding out fake fans and followers. So any gains would just be temporary.”