The paper, authored by Google analysts Reggie Panaligan and Andrea Chen, claims that Google data can predict box office results with 94% accuracy. The results are arrived at by observing Google searches for a movie's trailer, then combining that with the movie's "seasonality" and "franchise status." A movie's paid search ad click volume in the week prior to its release is also heavily indicative of box office performance.
Seasonality refers to what time of the year a movie opens, with summer and holiday weekends, obviously, affecting box office totals quite a bit. Franchise status refers to how many movie theaters a movie will be shown in. The status is broken up into several tiers, from Tier C being a movie shown in less than 2,500 theaters, all the way up to "midnight blockbusters" such as The Dark Knight Rises and the Twilight movies.
The paper shows that moviegoers are using Google more than ever to research their movie choices, with Google stating that movie category searches were up 56% from 2011 to 2012. This means that online marketing and trailer releases are becoming more important than ever to a movie's success. It also means, however, that it is becoming harder than ever to hide a poor movie's quality, with bad reviews only a Google search away.