A tweet from MSNBC host Lawrence O'Donnell on Twitter received 18 retweets in 10 minutes, two off the prediction of 20. Photo: Twoujia: Retweet Oracle
Maximizing retweets of Twitter posts has always been a science to some Twitter users. They try to post at certain times when a lot of their followers are active. Others try to list the right hashtags or mention accounts of important people. Now, there might be a new consideration to take into account.
A Massachusetts Institute of Technology professor has developed a model that predicts the number of retweets in the entire lifetime of a tweet based on the first couple of minutes of retweet activity.
"A lot of people felt that Twitter usage was totally random and unpredictable, that it was all just noise," said Sloan School of Management assistant professor Tauhid Zaman. "But it turns out that there is a systematic, repeatable behaviour that you can model."
Zaman unveiled a website that charts the retweets of about 50 tweets from politicians and celebrities and shows how they match up with predictions. Offering a social media twist on the Ouija board, the website's called Twoujia: Retweet Oracle.
Zaman said the research could help advertisers understand the spread of ideas and give users a basis to monetise their tweets.
He worked on the research with Emily B. Fox, Amazon professor of machine learning at the University of Washington, and Eric Bradlow, co-director of the Wharton Customer Analytics Initiative at the University of Pennsylvania. They've submitted the study to statistics journal The Annals of Applied Statistics.
The latest research follows up on Zaman's previous work on determining someone's influence on Twitter. He found that during major events, one Twitter user becomes a "Superstar". Because of the high activity of their followers, these superstars garner substantially more retweets than other users.
Sign up for CIO Asia eNewsletters.