Hollywood Scoffs at Google Box Office Prediction Tool

Big tracking firms and studio marketers dismiss the search engine giant's boast of 92 percent accuracy.

Movie marketers increasingly have grown nervous in the past few weeks. Tracking -- the prerelease surveys of audience interest in movies -- has failed to predict the disappointing openings of After Earth and The Internship or the surprisingly strong bows of Now You See Me and The Purge.
Amid all the anxiety, Google jumped into the breach June 6, claiming that by measuring search volume and other factors such as franchise status and the time of year, it can predict a movie's opening weekend. "In the seven-day window prior to a film's release date, if a film receives 250,000 search queries more than a similar film, the film with more queries is likely to perform up to $4.3 million better during opening weekend," wrote Andrea Chen, Google's principal industry analyst. "When looking at search ad click volume, if a film has 20,000 more paid clicks than a similar film, it is expected to bring in up to $7.5 million more during opening weekend."

So far, Hollywood isn't floored by that revelation. "It's kind of a big duh," says one studio marketing exec. "Still, given the scale of all the data points available to Google, we have to pay attention to whatever they can show us."
While Google refuses to offer dollar predictions for specific films, it claims it can make predictions with 92 percent accuracy. It further reported that four weeks before a movie opens -- before the main ad buys kick in -- measuring searches for a film's trailer can result in a box-office prediction that is 94 percent accurate.
Google doesn't plan, however, to start charging studios for such information the way other tracking services such as NGR and MarketCast do. "We're going to offer this as a courtesy to our clients," says Reggie Panaligan, Google's senior analytic lead. That doesn't mean that if a film is underperforming, Google will use its findings to encourage clients to buy more search ads. "We are not suggesting causation," he adds. "You can't really buy search query volume. That's organic; it comes from things like trailers, billboards. If a studio is unhappy about what we are predicting, we won't say search is the missing element."
Still, other competitors are asking, where's the beef? "Tell us how a movie is going to perform," says Ben Carlson, president of Fizziology, which measures chatter on Facebook, Twitter and other social media platforms to offer its own predictions. "I may be a bit of a purist, but I'd prefer that my advertising and my research not have a direct relationship."