** This article first appeared in movieScope, issue 19. **
As Facebook, Twitter and the like spawn millions of amateur critics, Dr Bernardo Huberman explains how social media is becoming an effective tool for predicting box-office performance.
Dr Bernardo Huberman may hardly be a household name, but as consulting professor of applied physics at Stanford University and director of the Social Computing Lab at Hewlett-Packard Laboratories, Hollywood has certainly been taking notice of his recent research into attention and influence in the new online social media.
“Our whole recent focus is at the intersection of information technology and social behaviour—the notion of social attention, and how attention is allocated to content—if only because information is now so plentiful and free, but attention is so scarce,” he tells movieScope. “As part of that, I always had this notion that where attention goes, you can predict the future in a sense, if only because so many people are focused on that. And the reason I chose movies is because you can predict something very concrete, which is box-office revenue.”
I always had this notion that where attention goes, you can predict the future in a sense.
Huberman’s co-authored article ‘Predicting the Future With Social Media’, published in March 2010, demonstrates a direct, measurable correlation between the number of ‘tweets’ referencing a given film title in any week, and the film’s box-office takings the following weekend. Indeed, the predictions of his analytic model, which were further fine-tuned when the sentiments (positive, negative or neutral) of the tweets were taken into account, proved considerably more accurate than pre-existing predictive models, including the industry’s gold standard, the Hollywood Stock Exchange (www.hsx.com). And while the study has taken Twitter as its focus, Huberman is quick to point out that “this can be done in any social medium.
Now, Huberman has the industry’s attention. “Even though I’m not allowed to discuss particular specific names, we are in discussions with two big entertainment companies in Hollywood about the licensing of the technology.” Nor is it just Tinseltown that has come knocking. “Anyone interested in present-day marketing is interested in this, to know the future of a product, how people will feel about something, the rate at which people will pay attention to something.”
The most surprising fact to emerge from the study was just how little impact, relatively speaking, the pre-release dissemination via Twitter of promotional materials and URLs has on a film’s subsequent box-office performance. Evidently Twitter has not yet proved a good platform for manipulating opinion. “So far,” Huberman comments, “they haven’t been able to sway people one way or the other, at least through this kind of channel.” Perhaps part of the solution from a film marketer’s perspective would be to determine which tweeters have the most influence, and then to co-opt their support. As it happens, Huberman’s next co-authored research paper, ‘Influence and Passivity in Social Media’ (Aug 2010), is concerned precisely with a model for “how to identify influential people in a social network.” Now, it seems, Huberman’s own work is having an influence on an industry always desperate for attention.