Los Angeles startup Cinelytic uses artificial intelligence to project box office returns. It licenses data about past movie performances and cross-references that with themes and actors in order to identify patterns. The software allows the user to input script and cast, and then try to see how specific actors would likely impact box office receipts. Other companies are experimenting with applying AI to film production, including the Belgium-based ScriptBook, which its founders say can predict success via an analyzed script.

The Verge reports that Israel-based startup Vault “promises clients that it can predict which demographics will watch their films by tracking (among other things) how its trailers are received online.” The company Pilot said it can “forecast box office revenues up to 18 months before a film’s launch with ‘unrivaled accuracy’.”

Even 20th Century Fox revealed in November 2018 that it uses “AI to detect objects and scenes within a trailer and then predict which ‘micro-segment’ of an audience would find the film most appealing.” (Although the results weren’t always very helpful.)

According to Cinelytic co-founder/chief executive Tobias Queisser, “on a film set now, it’s robots, it’s drones, it’s super high-tech, but the business side hasn’t evolved in 20 years.” “The data is very siloed, and there’s hardly any analytics,” he added. Queisser and his co-founder/chief technology officer Dev Sen come not from Hollywood but industries that have long relied on data and analytics, as do many who head these startups.

ScriptBook data scientist Michiel Ruelens reported that, with a list of 2017/2018 movies, his company “correctly guessed whether a film would make money 86 percent of the time … twice the accuracy rate of what the industry achieved.”

But AI has its flaws, one being that “the predictions made by machines are frequently just blindingly obvious.” Algorithms are also unlikely to “account for cultural shifts or changes in taste that will happen in the future.” U.K.-based Ingenious Group director Andrea Scarso noted that, at his company, humans are required to catch all the nuances; at his company, Cinelytic software is a supplementary tool.

“Sometimes it validates our thinking, and sometimes it does the opposite: suggesting something we didn’t consider for a certain type of project,” Scarso said. “I don’t think it’s ever changed our mind.” ScriptBook’s Ruelens believes the analytics tools haven’t been widely used because “people are embarrassed … [to turn] to the cold-blooded calculation of a machine” in an industry that prizes creativity. Queisser agreed with his assessment.

Netflix is what convinced many of these conservative Hollywood execs to overcome their distaste of big data, since “the streaming behemoth has always bragged about its data-driven approach to programming,” claiming its algorithm is worth $1 billion a year. In 2016, Netflix head of product innovation Todd Yellin stated, “we have one big global algorithm, which is super-helpful because it leverages all the tastes of all consumers around the world.”

Ruelens stated that, four years ago, big companies in Hollywood were “super skeptical.” After those companies did their homework to find out if algorithmic predictions got it right, they learned to trust it. “They’re starting to accept our technology,” he said. “It just took time for them to see.”