Cracking the Box Office Code

Have scholars found the formula that predicts movie hits?

Charts on a computer
BSIP/UIG Via Getty Images
The activity surrounding a movie's Wikipedia page may be key to predicting how much money it will make.
By |

Every weekend, a small industry of box office pundits and prognosticators – some using sophisticated market research, some just throwing darts – try to predict which movies will be hits and how much they'll earn at the multiplex. Of course, the very unpredictability of the box office (nervewracking for  nail-biting executives and filmmakers, exciting for everyone else) keeps the pundits in business.

Which Movie Has Ruled the Box Office for the Past Three Weeks?

Now, scholars think they've found the key to predicting how much money a movie will make its opening weekend in North America, up to a month in advance of its release. And no, it's not social media buzz – those metrics are so five minutes ago. Rather, it's the activity surrounding the movie's Wikipedia entry.

According to the study by mathematicians from Oxford University, the Central European University at Budapest, and Budapest University of Technology and Economics, who published their study in August in the journal PLoS ONE, there's a general correlation between a movie's first-weekend earnings and its Wikipedia activity – the number of page views for the movie's entry, the number of human editors who've built the page, the number of edits they made, and the diversity of the online users. Plugging these numbers into their formula yields an accurate prediction of the movie's opening-weekend take some 77 percent of the time. While not perfect, it beats the 57 percent accuracy rate the scholars claim for professional market research firms.

One of those firms is Exhibitor Relations Company, whose senior box office analyst, Jeff Bock, jokes that "if there ever was a conclusive scientific formula for how well a movie will perform theatrically, I would more or less be out of a job." Fortunately for him, the European mathematicians' algorithm seems to ignore other factors he says are crucial, like a movie's opening-weekend competitors, its genre, and the box office history of the movie's director and actors.

Stepping Into the Void: The Magic of Gravity

Indeed, the study's authors admit some blind spots. For one thing, they find their formula is a lot more accurate with blockbusters than modest-grossing films, since hit movies involve more Wikipedia activity and therefore generate more data. In other words, the system might recognize that a would-be blockbuster is probably going to belly-flop, but it can't tell the difference between a movie like that and an indie or foreign film that's a hit on its own terms, even if it earns a tiny fraction of what well-hyped studio movies opening in thousands of theaters earn.

The study also doesn't seem to take into account the omnipresence of a movie's marketing – everything from its ad campaigns to the number of theaters it books on opening weekend. A study last year by researchers at Japan's Tottori University found a correlation among a movie's earnings, its positive word-of-mouth (as measured via blog entries), and its advertising budget. In other words, movies that have more money to spend on marketing do better, which seems to be Box Office Analysis 101.

Even so, for all of science's progress in predicting box office winners and losers, the ultimate rule in Hollywood may still be the one laid down by The Princess Bride screenwriter William Goldman: "Nobody knows anything." So suggests Bock, who adds, " I mean, except that Avengers 2 and Avatar 2 will gross billions – other than that, we're all in the dark."

See Why Peter Travers Calls 12 Years a Slave "Best Picture Bait"

Besides, he says, there are always flukes that science can't explain. "After all, Paul Blart: Mall Cop and Wild Hogs are hit films. Just thinking about that makes my brain hurt." While he does believe there is predictive value in social media mentions, he says, "I don't think anyone has accurately locked down the factors at play with Twitter and Facebook trending and how that overall anticipation for a film directly translates to financial success."

If you're judging by Twitter, the market research firm Fizziology says this fall's biggest movies are Riddick and Machete Kills. Sure, they're both sequels to movies that weren't big hits in the first place, but they're generating more tweets than the seemingly surefire The Hobbit: The Desolation of Smaug. If the Tolkien film tanks while the action underdogs prosper, you heard it here first.