Reciprocity — which psychologists define as a social norm of rewarding one positive action with another — underpins the social media ecosystem. Most people implicitly know that one follow deserves another, a like should be acknowledged with a like, and so on.
Taking this practice to another level, however, groups of Instagram users are inflating engagement rates to artificially boost the popularity of certain content.
That’s according to researchers at NYU’s Tandon School of Engineering and Drexel University, who call these groups “pods.”
Pods — some of which number in the thousands — have been adept at manipulating curation algorithms.
Motivations for such activity range from increasing the reach of promoted content to amplifying political rhetoric through a tactic known to the researchers as “reciprocity abuse.” That’s where each member agrees to reciprocally interact with content posted by other members of the group.
Rachel Greenstadt, associate professor of computer science and engineering at NYU Tandon and lead author the report, was taken aback by the effectiveness of such abuses.
“One of the most surprising findings was how effective reciprocity abuse is at not only raising the visibility of a post, but in increasing real, organic engagement,” Greenstadt wrote.
How are pod people finding each other? Apparently, pods are commonly advertised on the message boards of other pods, according to Greenstadt.
For the most part, any user can join a pod. In fact, only 4% of pods require users to have a minimum number of followers before they can join, the researchers found.
All told, they analyzed 1.8 million Instagram posts belonging to 111,455 unique Instagram accounts, advertised across more than 400 Instagram “pods” hosted on the instant messaging service Telegram.
Of course, manipulating social algorithms is nothing new.
But whereas most other attempts to game the system have involved techniques such as automated bots and scripts, pods “involve humans taking action manually, so they are harder to detect,” wrote Ph.D. candidate Janith Weerasinghe.
For their analysis, the researchers collected metadata from pod groups, gathered Instagram data associated with both the pods and control posts to train a “classifier” — a machine learning function used to assign labels to particular data points — to detect pod engagement, and then analyzed the efficacy of the pods to discover if using them increased organic interaction.
They then used a machine learning model to predict with a high degree of precision whether or not an Instagram post was part of a pod, regardless of the levels of interaction and engagement.
By exploring how interactions with a post changed over time across users’ profiles, they found that posting in pods boosted organic post interaction.
On average, each pod had about 900 users, but some had as many as 17,000.
Correction: This original version of this column described Telegram as a unit of Twitter. It is not.