Inside the ISIS Social Media Campaign

A new study from Brookings creates a snapshot of the network and how supporters use it. Reuters

Many forces have been feeding the Islamic State (ISIS), but none are seemingly more elusive than its sophisticated social media campaign. That's the focus of a new study out of the Brookings Institution, which some are calling one of the most exhaustive investigations into the inner workings of the group's social media strategy to date.

Due to the fragmented nature of the organization, previous efforts focused on small segments of the ISIS social network. But using a sample of 20,000 ISIS-supporting Twitter accounts, as well as a "sophisticated and innovative methodology," the authors of "The ISIS Twitter Census" study released Friday constructed a snapshot of the size of the network, who is part of the campaign and how ISIS supporters leverage social media to strengthen the organization.

The study estimates that between September and December 2014, there were at least 46,000 ISIS-supporting Twitter accounts, though not all were active at once. In fact, the authors found that a minimum of 1,000 ISIS-supporting accounts were suspended during that time frame. One reason for the suspensions is that ISIS's social media strategy is known for its violent content.

Around 20 percent of supporters indicated that English was their primary language when using Twitter, while about 75 percent selected Arabic. The typical ISIS supporters were situated within areas the group either controls or is contesting in Syria and Iraq. Saudi Arabia and other parts of Iraq and Syria were also major hubs.

Supporter location was determined in a variety of ways. One method was assessing the locations listed on Twitter profiles. This was deemed less reliable than analyzing the geo-locations of users who enabled the location feature on their smartphones. The number of users who allowed geo-location, however, was small.

Of the supporters who used a mobile phone to tweet, 69 percent used Android, 30 percent used Apple and around 1 percent used BlackBerry. The study notes that in mid-December, ISIS announced a ban on iPhone products within its territory, but only a 1 percent drop in the use of iPhones was reported by February.

The study found that each account, most of which were created in September 2014, had an average of around 1,000 followers. The authors also discovered that the more active the user, the more likely he or she was to be suspended.

Among the tens of thousands of ISIS-supporting accounts, the study attributed the strength of ISIS's online campaign to a small group of highly active users—numbering between 500 and 2,000 accounts—who sent more than 50 tweets on average per day.

Only 4 percent of ISIS-supporting accounts had more than 5,000 followers—a low number compared to celebrities with followings in the millions, but high compared to the average Twitter user.

The authors discovered that supporters retweet content sent by others within the network as a way to counteract the high number of Twitter suspensions of key influencers.

"Part of the reason it is so effective is because it is organic, it's from the audience that it is going after," said Shahed Amanullah, a former senior adviser at the State Department. "These young people understand youth frustration, they understand the fascination with violence, they understand that imagery and graphics that you see in Hollywood will attract these people."

Meanwhile, the lack of a comprehensive understanding of ISIS's online strategy has led to poorly informed reactions from governments and tech companies, and it has allowed the group's campaign to flourish.

"Those graphics and that messaging, they propagate pretty quickly through the Internet," Amanullah continued. "The trouble we face is to counter that, [the U.S. government's] countering messages...don't percolate through the Internet in the same way. The kumbaya message does not fly through the Internet the way a beheading video does."

The study concluded that the suspension of Twitter accounts hindered the group's ability to broadcast its message—pointing to indicators such as the diminishing usage of its main hashtag, the group's name in Arabic. The study also suggested the metrics of the network could be manipulated in such a way as to disrupt the flow of information among these accounts. But the authors recognize that suspensions and regulations set political precedents.

"Regulating ISIS per se presents very few ethical dilemmas, given its extreme violence and deliberate manipulation of social media techniques," the study reads. "However, the decision to limit the reach of one organization in this manner creates a precedent, and in future cases, the lines will almost certainly be less clear and bright."