Starting Points: Movies

Apr 2, 2010

Twitter: The Killer Box Office Predictor?

by Jolie O'Dell

Twitter can predict box-office takings better than other industry-leading data sites, according to research just released by HP. Between the sentiments expressed and the rate at which buzz builds, the microblogging service outperforms other forecasting mechanisms for the film industry.

When compared to industry favorite The Hollywood Stock Exchange (HSX), Twitter trumps on predicting real-world outcomes.

The HSX is essentially a web-based game that utilizes virtual currency to predict the success (or failure) of a given film, actor, or director. But the virtual prices at which one sees filmic properties being traded on HSX strongly correlates to real-world box-office dollars – and players’ favorites can translate into professional accolades. For example, in 2007, HSX correctly predicted 82% of Academy Award nominees in major categories and around 88% of Oscar winners.

Image via CrunchBase

However, HP’s dissection of Twitter streams shows that the social site is more accurate than HSX at determining box-office revenue, even for pre-release movies, by 1-2 percent. This might not seem like a significant statistic, but when one considers the box-office take of a Hollywodd blockbuster, a percent or two can add up to millions.

Also, Twitter provides a free and open stream of data that’s fairly simple to grab and parse – something most competent social media analysts can appreciate. And in addition to simple charts showing URL mentions, retweets, and acceleration curves, Twitter also provides a rich bank of user-generated sentiment – emotionally weighted statements that further show whether users are recommending or slamming a particular movie to their friends.

These sentiments can forecast trends in sales, as well. One movie analyzed in this study, The Blind Side, had an “enormous increase in positive sentiment after release,” reads the paper. The film’s score jumped from 5.02 to 9.65 on HP’s scale. After a “lukewarm” first weekend, with sales around $34 million, the movie “boomed in the next week ($40.1 million), owing largely to positive sentiment.”

As the HP researchers note, “While in this study we focused on the problem of predicting box office revenues of movies for the sake of having a clear metric of comparison with other methods, this method can be extended to a large panoply of topics, ranging from the future rating of products to agenda setting and election outcomes. At a deeper level, this work shows how social media expresses a collective wisdom which, when properly tapped, can yield an extremely powerful and accurate indicator of future outcomes.”