Twitter Bots Match People Based On Keywords In Tweet

Sarah, meet Joe. You both used the same keywords in a post on Twitter, so the two of you must have something in common. Rather than allow you to find each other on your own, I, the socialbot, will connect the two of you through keywords used in your tweet. Well, something like that.

A group of Web researchers -- Web Ecology Project -- studying the structure and the dynamics of social media created a socialbot, a sophisticated Twitter bot that connects one real person to another based on the keywords tweeted.  

A white paper written by researchers Max Nanis, Ian Pearce and Tim Hwang reveals a study that measures the socialbots' outgoing and incoming follow and mention activity. The socialbots were programmed to vary tweet and follow activity, and variation in day-to-day activity was minimal. Because all socialbots in this study were programmed with the same settings, little variability was seen in outgoing activity across the socialbots. On average, socialbots each tweeted about 36 times and followed roughly 19 users each day.

During the 21 days of the experiment, socialbots were able to attract a total of 561 followers -- an average of about 62 followers per socialbot. Some socialbots were more successful in attracting followers than others. The most-followed socialbot, Bota, attracted 92 followers; the least-followed socialbot, Bote, attracted 45 followers.

On average, socialbots were each able to gain followers from about 16 users in their respective target groups, or roughly 5% of users in each target group.

The research paper explains how the trio will attempt to improve on the tactics designed to connect Twitter users. The group has begun to develop functions that give socialbots the ability to determine the general interests of other users.

While this research caught my attention after reading a Technology Review article published by the Massachusetts Institute of Technology (MIT), Reliable SEO President David Harry recalls tweeting one day about feeding his snakes. Out of nowhere, the person with the Indiana Jones account, from the movie, tweeted back to Harry something similar to -- Oh no, not snakes.

"It was obviously a bot seeking the keyword 'snakes' and auto-replying to my post," he said. "In that instance, I thought it was kind of cute. On the other hand, there are a lot of bots that are just spewing the same ad message randomly to users."

Most "spam" is unwanted or requested advertising, Harry said. The bots are seeking keywords and then responding based on keywords. That's a tricky strategy. Harry said it's really best to engage the user in hopes they follow the account. Spamming people rarely accomplishes the goal.