WE ARE VERY STORY, Part 1: Hollywood’s Narrative Algorithms

by Yves Bergquist

NOTE: over the past 12 months my team and I at the Entertainment Technology Center have done a lot of research on how to measure narrative structures in film, advertising, and marketing. I’ve decided to lift the veil on some of that research in a series of posts as part of a series called “We Are Very Story”.

PART ONE:

I believe two things:

(1) Hollywood and Silicon Valley are in the same business: producing algorithms.

(2) To survive and thrive, the media and entertainment industry needs to start thinking algorithmically about stories.

I’ll tell you why.

Everything is story. Your car, your haircut, your job, your cereal, who you love: the entire universe is perceived through the lens of our narrative representation.

Why? Because the reality we live in is too complex for us to process in its entirety, so we need to represent it in a simplified manner. This is what stories are: a cognitive function. Just like visual perception and motor skills, they are essential to our survival.

Stories are the music we extract from the noise of reality. Stories bring structure to the immensely complex nature of our universe, which we perceive as chaos.

Just like software algorithms compress large and complex inputs (data) into a set of simple arguments aimed at solving a problem, narrative representation compresses the infinite complexity of our reality into simple cognitive scripts. And these scripts usually represent the universe’s complexity with some attribute of justice, which – as we painfully know- reality doesn’t have: Good vs Evil, David vs Goliath, the small-time traders outsmarting Wall Street in “The Big Short”. All of these stories are cognitive tools, they are procedural roadmaps which we use to make decisions about what to wear, what to eat, who to hire and who to marry. Anytime we make a decision, we activate a procedural script that has been shaped by a story, which we trust – we hope- accurately summarize the universe’s complexity.

When our stories are high-fidelity compressions of the universe’s complexity, it’s called genius. When they aren’t, it’s called hubris.

More often than not they don’t. It’s why we fail so much.

Symbols, archetypes, religions, political ideas: they are all stories. They are all cognitive scripts we use to simplify complexity into a procedure we can follow to survive and thrive. And the more complex, ambiguous or multivariate the context (ethical dilemmas, relationships, or elections) the more critical stories are. This is why political campaigns throw hundreds of millions of dollars at building narratives: “Make America Great Again” is a lot simpler than a detailed argument about how much government we need.

Because it most closely mimics the multi-sensorial nature of reality, film narrative has a very special cognitive power. This is why film stories have considerable impact on our representation of ourselves and the world, and why they carry such social responsibility.

Perhaps more than any other form of storytelling, film narrative has a structured and powerful algorithmic structure. It’s not by chance that film stories are called scripts (notice the terminology is identical to software development): in an odd similarity to a software algorithm, almost every movie script is a set of often elegantly mathematical arguments (a procedure) through which the main character will attempt to solve a problem.

So without realizing it, Hollywood has beaten Silicon Valley at its own game: it has produced and monetized tens of thousands of complex (narrative) algorithms for over 100 years.

But the industry doesn’t think that way. In fact, read the sentence above to a creative executive at any studio and you’ll get thrown out of her office.

There lies the irony: even though Hollywood has made more money designing algorithms than the tech industry, it has never thought of it that way. That’s because all creative industries are run by mathematicians who hate math, artists who can produce beautiful and powerfully algorithmic narratives with little to no conscience of their mathematical structures.

As we’ll examine in a subsequent post, John Lasseter, Pete Docter, Brad Bird, Andrew Stanton and all of their storytelling colleagues at Pixar, for example, have built some of the most cognitively powerful and mathematically cogent narratives. This, I’ll argue, largely explains their sustained success, but their process was that of brilliant instinct rather than deliberate algorithmic thinking.

So while the media and entertainment industry know that audiences have a powerful relationship with film stories, nobody ever succeeded at quantitatively measuring that relationship: namely what specific story and character mechanics (what arguments in those cognitive algorithms) drive financial (theatrical box office) performance. Even the most sophisticated box office prediction models aren’t able to account for narrative structure in their quantitative models, which would be like trying to predict the outcome of a baseball game without accounting for batting average or runs.

Until now.

My team and I at the Entertainment Technology Center have just finished our preliminary study that aimed at mapping specific, almost scene-level story and character mechanics to box office returns, unlocking the cipher to what narrative structures in film are the most popular. We’ve successfully looked at what kind of arguments within specific narrative algorithms are correlated to box office success.

Stay tuned for that one: you’ll definitely want to hear it.