Typically, we think of art as being inspired, not data driven. But with Amazon’s Manchester by the Sea snagging a Best Picture nomination at the upcoming Oscars, the first film from a streaming media service to do so, it’s an exciting time to discuss the role of data science in film.
In the movie business, the ability to analyze patterns in big data such as viewing behavior and user feedback cycles is changing the way companies approach content creation. Social media provides an unprecedented, real-time window into audience preferences. This makes the possibilities associated with using data science to predict the efficacy of characters, plot lines, and actors on viewing behavior nearly limitless.
In the past, studios had a difficult time predicting how well a movie would perform at the box office. William Goldman, a two time Oscar-winning screenwriter, famously said, “Nobody knows anything… Not one person in the entire motion picture field knows for a certainty what's going to work [at the box office].” That’s because studios traditionally used very coarse information based on high-level demographics such as gender and age to create broad audience target groups. Now, the proliferation of data in likes, shares, comments, and check-ins from Facebook, Twitter, and other social media platforms have made it possible for Hollywood to understand its audiences in a much more granular way.
That means modern film studios can specifically target their content to the right audience based primarily on the likelihood a group of viewers will be interested in the content, but also on the value they’re likely to bring back to the studio by engaging with it. Other tools are helping Hollywood predict the likelihood of a film’s success when an idea is conceptualized: Josh Lynn of Piedmont Media Research has an algorithm that predicts how much revenue a movie will make opening weekend at the box office based on the basic plot structure and cast.
Hollywood is no doubt taking a page out of Netflix’s book in using data science to generate successful content. Netflix disrupted the movie rental business by creating a simple mail order subscription DVD service in the late 1990s. Long before the company had the streaming data points it does now, it operated with the understanding user data would be pivotal to its success. In 2006, Netflix launched the Netflix Prize contest, which offered a million dollars to the group that could determine the best algorithm for using previous ratings to predict future movie ratings with only four data points: customer ID, movie ID, rating, and date watched. This was the beginning of Netflix’s wildly successful recommendation engine.
Today, many of Netflix’s hit shows — such as the 6 time Emmy winner House of Cards and the more recent smash Stranger Things — are the product of complex predictions based on hundreds of millions of user data points. Executives at Netflix knew House of Cards would be a hit before production even began based on audience preferences toward David Fincher’s The Social Network, Kevin Spacey, and the popularity of the original British version of House of Cards.
While Netflix is pioneering the way movies and television series are created, Hollywood would do well to emulate its data-driven approach — especially now that streaming media has made its way to the Academy Awards.
After all, show business is big business. And Hollywood can’t afford to fall behind.