No matter how much movie studios may wish for one, there is no crystal ball that will predict how much money a flick will make at the box office. But there may be something almost as useful: Google. In the high-stakes game of betting on what may be the next Hollywood blockbuster, searches for movie trailers are a metric with surprising accuracy.
A whitepaper released today by the search giant's Think Insights group called "Quantifying Movie Magic with Google Search" reveals that trailer-related searches done four weeks before a film's premier can be used to determine opening weekend revenue. According to the paper, coupling that "key leading indicator" with the current movie season and a film's "franchise status" — a metric that evaluates whether a movie is part of a top-tier franchise like James Bond films, or a "midnight" blockbuster like a Twilight film — can predict the box office take with 94 percent accuracy.
"While we've always known that trailers are an important aspect of the movie decision-making process, we were surprised at how strongly trailer search volume is linked to intent at this particular point in time," Jennifer Prince, Google's head of industry, media and entertainment, said in an email. "While news articles or movie reviews can give you the facts from another person's perspective, movie trailers allow individuals to experience the film firsthand, and that's what can really help a moviegoer form an opinion about a movie."
The predictive ability of movie trailers, the paper noted, could help movie marketers — ostensibly Google's customers — immensely because the potency of that search peaks a month before release, when the information is exceptionally valuable to studios.
"While an accurate opening weekend forecast calculated the day before premiere is certainly a valuable data point for planning post-release marketing strategy, it doesn't leave movie marketers with very much time to react," the paper notes. "Fortunately, Google (and YouTube) search data gives a great indication of where a movie is headed as early as four weeks from release week." (FYI for math fans: Google's prediction system is a linear regression model. So, for their trailer search analysis, they found that 64 percent of the variation in box officer performance could be explained by looking at trailer search data. When that's coupled with franchise status and seasonality, the level of accuracy raises to 94 percent.)
How good is Google's model? For one, it could've predicted the male-stripper flick Magic Mike was going to be a bigger hit than anyone thought. Going into the film's opening weekend its studio Warner Bros. was hoping for an opening weekend between $15 to $20 million, with some believing it could hit $25 million. But thanks to a lot of buzz, and the magic of Channing Tatum, the film hit $39 million in its opening weekend. What would Google's search data model predict? $40 million. It can also help curb overly ambitious expectations. Battleship was expected to reach some $40 million in its opening weekend, but raked in a disappointing $25.5 million. Google's data put its opening weekend at less than $30 million. For something like The Hunger Games, which had a strong franchise following and "midnight blockbuster" status, Google predicted it would top $160 million, "not too far off from the $152 million that the film opened with," Prince notes.
In other areas, the search company's research also found that on average movie fans consult 13 sources before they decide what flick to see and film searches on Google were up 56 percent from 2011 to 2012. The study, which examined the 99 top box office titles in 2012, also found that moviegoers tend to pick what film they're going to see the same day they purchase their ticket.
Traditionally, factors such as the number of screens a movie is opening on, the bankability of particular directors or stars, and that nebulous thing known as "buzz" can all be seen as good indicators of how a film might perform at the box office, but the inclusion of Google search data here is fascinating. In another example in the whitepaper, the search giant's researchers found that combining a movie's franchise status with theater count and data on the volume of searches done on the film a week before release proved to be 92 percent accurate in predicting box office results. Moreover, the research showed that if a movie had a quarter-million more search queries than a similar flick that movie is likely to bring in up to $4.3 million more in its opening weekend than the less-searched film.
So what does Google Box Office (not a real thing) mean for moviegoers? Probably not much. But for studios it could mean that the plethora of trailers released in various forms prior to a movie's release really do have some impact. It also means that—at least according to Google—movie marketing should place a renewed emphasis on search-based advertising.
"Ultimately, it is this online engagement that gives us tangible insight into intent," the whitepaper concluded. "Arming movie marketers with actionable data in their never-ending quest to quantify 'movie magic.'"
Now if only it could predict how good The Internship will be.
