Three strikes and you're out: Can data insights predict best picture three years in a row?

Bryan Melmed

Winners are chosen by the 5,800 members of the Academy of Motion Picture Arts and Sciences. Exactly who these people are is a secret, but we do know that they are fairly homogenous. In a Los Angeles Times exposé, Academy members are described as 94% Caucasian, 77% male, and having a median age of 62. With this information (industry, job title, location, ethnicity, gender, and age) we can identify 32,000 relevant user profiles, a handful of which may be voting for best picture, and many more who are the in the influencer network of voters. It's not a perfect read, but it's enough to make some smart assumptions and drive an analysis to the Best Picture winner.

Too kooky/too smart

Last year we wrote off four movies immediately as their core audiences were markedly different from the Academy profile. This year there are only two that can be eschewed so easily - The Grand Budapest Hotel and The Theory of Everything.

The cities most receptive to The Grand Budapest Hotel help illustrate how this film found its audience in a distinctive segment of American society. We see high interest in the young liberal enclaves of Austin (Texas), Madison (Wisconsin) and San Francisco and more broadly amongst the highly educated "emerging affluent". It also had the highest density of female viewers and had the youngest median age of any nominee in the category. In twenty years these viewers will look very much like an Academy voter, but the movie is out of luck this year.

Moviegoers who favour The Theory of Everything are north-eastern liberals, huge fans of NPR, work for non-profits or educational institutions, and favour electric hybrids like the Toyota Prius. They are the most educated movie audience but struggle financially - both student loans and the high cost of housing have taken their toll. Some may imagine this a perfect fit with the Hollywood stereotype, but the Academy members are Hillary Clinton voters - liberal but not left-wing, educated but not intellectual, and with a West Coast affinity for strong personalities.

Too niche

Two movies found their niche outside the gates of Hollywood, and while they have fans within the Academy, their appeal is too narrow to pull off a win. Goodbye, then, to American Sniper and Selma.

American Sniper has an unusual profile for a Best Picture nominee. It is most popular in rural states like Alaska and Montana, among military personnel and policemen, and with drivers of large SUVs. Yet this film also finds favour within the confines of Hollywood. For one thing, the audience demographics are perfectly aligned with the Academy voter - overwhelmingly Caucasian, male, and older. Looking deeper, American Sniper also has a distinct secondary audience of affluent urban professionals. It is hugely popular among owners of the BMW i3, a sporty electric car, and a favourite of lawyers, financiers, and film producers. That's enough to garner a nomination, but not enough to win.

The core audience for Selma is also quite different. These viewers are far more likely to be upper-middle class African Americans and to hail from the southern states. In other ways they resemble the American Sniper profile - an older audience, preferring traditional American cuisine and SUVs, who are likely to have a career in the military. But the distinct overlap with our Academy voter is in domestic matters. We see a shared emphasis on family life, with tight-knit multi-generational families and a great deal of money and attention lavished on the home. It may seem ironic, but more than any other film Selma reflects the sense of establishment that pervade the Hollywood elites. Again enough to capture a nomination, but not a majority vote.

Too broad

It may seem strange to rule out a film because it is widely popular, but usually we find that broad appeal also implies shallow support - and Boyhood and Whiplash found fans almost everywhere. The only telling distinction in these audiences was related to the content; in general, Boyhood was preferred by moviegoers with families and Whiplash by arts enthusiasts. Universal appeal is an ideal situation for a movie distributor but a nightmare at awards night if the consensus "feel-good" choice fails to make it to the top of anyone's list.

Our Best Picture prediction is complicated by the fact that votes are made on a ranked ballot, such that votes for losing films are redistributed towards the consensus winner. Specifically, if fans of the non-contender and niche films mentioned above share a second choice pick, Boyhood or Whiplash could steal the Best Picture nod. The correlation in our data is not strong enough to suggest this will happen. If it does, Boyhood benefit from the American Sniper and Selma vote, and Whiplash from The Grand Budapest Hotel vote. (The Theory of Everything voter will have picked Birdman second.)

The contenders

This leaves The Imitation Game and Birdman. They are the top choice of Los Angeles and New York residents, of wealthier audiences, and of those working in the film industry. The audiences of these films show the highest correlation with other Oscar best picture nominees, suggesting that they have similar tastes as Academy voters or that their preferences are shaped by the Academy nominations.

To choose between the two, we have to look past these broad strokes. Fortunately, our ad targeting is driven by precise, granular interests that can identify and describe the most highly relevant audiences. It is this approach that vaults Birdman to the top.

The audience most interested in movie cameras, video editing systems, pro audio equipment, sound stages, effects lighting, and cables favor Birdman. The audience who travelled to the Cannes film festival and the Sundance film festival would also choose Birdman. In contrast, those that prefer The Imitation Game are likely to read about film, discuss film online, and buy movie tickets or DVDs.

In short, those living as artistic creators (much like Birdman himself) are far more likely to vote for Birdman, while those passionate about film would choose The Imitation Game.

Since the Oscars are an insider's game, not an imitation one, we therefore predict Birdman will win Best Picture this year.

And if it doesn't, well, two out of three ain't bad.