Researchers’ Twitter Experiment Found SocialBots Increased Follower Numbers

Ever write a tweet only to get a response from a Twitter bot? Chances are you’ve come across one before, they’re an automated program designed to respond to any tweets featuring specific key words. The majority of bots are very obvious and, in turn, very annoying , but what would happen if they became more sophisticated, being capable of fooling almost anyone.

TechnologyReview report that a group of freelance web researchers created these type of bots called “socialbots” that could not only fool people into thinking that they were genuine, but speed up the natural rate of communication between real people.

The experiment began through the Web Ecology Project, an independent research group which focus on studying the structure and dynamics of social media phenomena.

The group wanted to investigate the claims of social media consultants who say that they could grow their clients Twitter networks and even increase the level of interaction between a brand and Twitter. They decided to measure these claims by getting teams of researchers to compete for the most Twitter replies. As there were no rules against using automated bots, a few teams realised they could compete better through this method.

SocialBot Experiment

After their success, three researchers, Tim Hwang, Max Nanis and Ian Pearce, created the Pacific Social Architecting Corporation (PacSocial) which focused on the study and development of socialbots. They set up a second experiment which looked at bot-human interaction and their ability to create new human-to-human connections.

One experiment included tracking 2,700 Twitter users, divided into randomly assigned “target groups” of 300, over the course of 54 days. The first 33 served as a control period, during which no social bots were deployed. Then, during the 21-day long experimental period, nine bots were activated, one for each group.

Each bot was programmed to perform simple actions such as retweeting messages, and “introducing” one real user to another by replying to one and mentioning another in the same tweet.

Surprising Results

On average, each bot attracted 62 new followers and received 33 incoming tweets (mentions and retweets). But the group discovered that the human-to-human activity changed within the target groups when the socialbots were introduced. They noticed a 43 per cent increase in follows during the 21 day experimental period, in comparison to the 33 day control period. Yet one group showed a 355 per cent increase in follows during the 21 day period suggesting that further research would be needed to determine this.

While it’s early to speculate, it could mean that in the future, when bots become more sophisticated in their approach, the entire dynamic of brand consumer interaction could be changed. Instead of outsourcing your social media accounts to consultants or companies, you could enlist the services of a programmer to help create the interaction, targeting those users by interest or location for example. It’s not going to replace community managers who build and maintain these relationships between consumer or brand, but it could take a lot of the time and hassle out of finding these people and convincing them to follow your brand.

The researchers created a visual representation of the bot-to-human connections made during the 21-day period, while the research paper detailing the entire experiment and results can be found here.