Exclusive: Marginal MediaWorks acquires Fresno to "de-risk" Hollywood movie making

A new tie-up between a startup content studio and a data science startup aims to flip the way Hollywood green lights movies and shows.

Why it matters: The deal comes as media giants (think Disney, Warner Bros. Discovery and Amazon-owned MGM) are struggling to decide which movies and shows to fund, which platforms or theaters to release them, when — and for how long.

Details: Marginal MediaWorks, based in Los Angeles, has agreed to buy Fresno Unlimited, an A.I. startup backed by Chamath Palihapitiya, Zander Lurie and Jason Calacanis.

• The companies have already been working together to use Fresno's technology to help pitch projects — armed with data-backed proof of audience fit — to major studios, networks and platforms.
• Typically, movies get a focus group treatment after they're made. “Our whole thing is — why don't you start there?” Sanjay Sharma, Marginal’s founder and CEO, tells Axios. 

Between the lines: The dominance of streaming requires higher quantities of content and therefore more careful budgeting. Past projects aren't even safe in this new environment where old shows like Sesame Street, most recently, are being booted off platforms to cut costs.

How it works: Fresno’s platform soaks up publicly available, anonymized data about consumer interests, profiles and activities from platforms including Facebook and Amazon.

• Sharma and his team at Marginal then use that data to estimate how many people might like a concept (for film, TV, animation and audio) in progress.
• Once a piece of content is made, they can also use Fresno’s platform to target and to distribute shows to certain viewers as they've done with stars like Jimmy Kimmel.

What they’re saying: "The intent and the practice is about de-risking [the process]," Rob Goldberg, Fresno's founder and CEO says.

• "We can now start marching in to buyers, to financiers, studios [and] streamers to say — not only is this one of the best scripts we've read in the last three months [but also] here are ... the audience segments it hits, the themes it hits [that] are resonating with these audiences," adds Sharma.

Yes, but: Both executives stress that the data does not create original ideas.

• Instead, creators that Marginal works with may want validation, specific distribution or sometimes inspiration, Sharma suggests.

Context: Closed platforms such as Disney+, Netflix and Amazon Prime Video have a lot of proprietary data that they can already use to make similar decisions about new content projects.

• Marginal sees its role with Fresno as being able to work with any sized studio or platform — regardless of what data they have.

The deal's bigger mission is to help stories from underrepresented creators and communities see the light of day.

• New movie ideas can be tossed away easily if they’re being measured against historical data. For example, if the industry continues to measure the potential success of an action concept based only on the fact that Tom Cruise action movies do best, that can neglect current reality.
• “This high failure rate [of other projects] by using the same sort of pattern recognition system is what we think is ripe for disruption and a massive commercial opportunity,” says Sharma, who is in the process of working on a horror project with Tyler Perry.

What to watch: The deal is expected to close before early next year. Terms of the deal were not disclosed.

Our thought bubble, via Axios chief financial correspondent Felix Salmon: I've seen this movie before. But maybe the sequel will be more successful.