What if Twitter could predict our success?


I saw this on Emilie Ogez’s blog (@eogez), who posted some surprising news from BBC Mundo.  Apparently, two HP researchers, Sitaram Asur and Bernardo Huberman, have created an algorithm that allows one to predict the success of a movie based on 3 criteria: the number of tweets generated about it, the longevity of the tweets, and their nature (positive, negatigve or neutral).  This algorithm is founded on the analysis of 3 million tweets concerning 25 films, including Avatar.  The results have 97% reliability.  The precision of certain results is quite surprising: the algorithm predicted 16.8 million dollars in profits for the opening week of The Crazies—and the profits turned out to be 16.6 million!  The same algorithm predicted 30.7 million dollars for the opening week of Dear John (see tear-jerker trailer below), which garnered 30.5 million dollars!

The algorithm is particularly good for movies whose target audience is similar to twitter users: young, tech-savvy, with a high purchasing power.  As a result, the algorithm is less accurate for other target audiences, such as children.  One wonders if this means of predicting a movie’s success based on Twitter could also predict the profitability of brands or of a political candidate….


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>