Could Machine Learning Guarantee Box Office Success?

Warner Bros. deal with Cinelytic could change the future of cinema

In Q3 2019 Warner Bros. revenues were $3.3 billion, down 10.4% YOY due to declines in theatrical and television revenues. The AT&T Q3 2019 Investor Briefing suggested this decrease was primarily due to a mix of releases, as the same quarter the year before included The MegThe Nun and Crazy Rich Asians. But what if the company could utilize machine learning so it never had to report a downturn to investors ever again?

This certainly seems to be the thinking behind the deal that Warner Bros. has signed with LA startup Cinelytic, with the division of AT&T’s WarnerMedia planning to leverage the company’s AI-driven project management system to make marketing and distribution decisions.

The Cinelytic platform, which launched in 2019, licenses historical data about movie performances over time and cross-references this with information about a film’s theme and stars, using machine learning to identify patterns in the data.

The company’s CTO and co-founder, Dev Sen, used to build risk assessment models for NASA, and his co-founder Tobias Queisser has a financial background. These men seem confident that their business model will be useful when entertainment behemoths like Warner Bros. are competing against others in the entertainment marketplace for revenue, particularly in the festival setting, where studios get caught in bidding wars and feel pressured to put up vast sums of cash quickly without necessary ascertaining the true worth of content.

Cinelytic users can create their own production ‘Fantasy Football’ style, combining actors with a genre and budget, and then making tweaks to see how their film would fare in different markets.

You can pit Channing Tatum against Zac Efron, using the tools to see where a film’s audience would be most hot under the collar when watching the hypothetical flick.

Sounds fun, right? But how useful is the application of machine learning to the film industry? Can machine learning measure things like the chemistry between the two lead actors? Would Netflix’s Marriage Story have worked so well if Adam Driver wasn’t cast alongside Scarlett Johansson? It’s got a limited scope when it comes to impacting the fundamentally creative processes in filmmaking such as casting, which Queisser acknowledges:

“Artificial intelligence sounds scary. But right now, an AI cannot make any creative decisions. What it is good at is crunching numbers and breaking down huge data sets and showing patterns that would not be visible to humans. But for creative decision-making, you still need experience and gut instinct.”

Many people are understandably skeptical about how groundbreaking this technology is, as machine learning applications are trained on historical data and are therefore more likely to focus on patterns that ensured success in the past, rather than accurately predicting what will excite the next generation of cinema-goers.

Photo by Max Nelson on Unsplash

It would be cool if the platform could pluck a relative nobody’s name out of a hat and propel them to stardom, but unfortunately, in its current state, machine learning can’t really help studios identify bright new talent. If relative newcomer Daisy Ridley had never been cast as Rey the most recent Star Wars films, it’d be unlikely Cinelytic would link her to a film’s Blockbuster potential. The platform can tell you when and where a big, established name like Tom Cruise would carry weight with audiences though, which could be useful for companies looking to bolster their success in international territories. As the media landscape becomes more and more competitive, swapping one actor for another could be the difference between a positive investor’s statement and a negative one.

But perhaps most worryingly of all, machine learning could be hampered by AI bias, which in the film industry might look like fewer female actors or ethnic minorities being suggested by services like Cinelytic as key components in a winning film. If entertainment companies listened to the recommendations of machine learning alone then it wouldn’t take long before we found ourselves regressing from the small improvements that have been made in intersectional representation in cinema over the last few years.

Source: Hollywood Diversity Report 2019

The 2019 Hollywood Diversity Report found that ethnic minorities made notable gains since the previous report (particularly in television), and relative to their male counterparts, women posted gains in seven of the 12 key Hollywood employment arenas, including among film leads and film directors (which is particularly notable given the history of women’s underrepresentation on this front).

People of colour accounted for 19.8% of the leads in top films for 2017, up from 13.9% in 2016, whilst women accounted for 32,9% of the leads in top films in 2017, a minor increase from their 31.2% share in 2016.

Source: Hollywood Diversity Report 2019

A recent story from The Hollywood Reporter claimed that Warner Bros. would use Cinelytic’s algorithms “to guide decision-making at the greenlight stage”, but The Verge subsequently confirmed with a source at the studio that the software would only be used to help with marketing and distribution decisions made by Warner Bros. Pictures International.

The company shows no sign of giving this technology the power to greenlight productions any time soon — and that’s almost definitely a good thing for unknown actors, women, ethnic minorities and cinema lovers everywhere.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s