As a general rule, if you learn about it, they’ve been doing to us for years, if not decades now.
There’s this term in the processed food industry for when something is quantifiably at its most delicious, but also–critically–still unsatisfying enough that you always crave more. It’s called the bliss point, and it’s what ensures that you always want another sip of that soft drink, or the crunch of that potato chip. It’s about maximizing the bliss of consumption so that you consume more.
Now, Disney Research has developed a neural network that seems to be chasing a similar idea–but for the world of movies. The system has been trained to watch an audience of theatergoers as they watch a film. It can track reactions like smiling and laughter on hundreds of faces in a dark theater, allowing Disney to quantify whether or not a film is working as intended on a granular scale. It’s easy to imagine such technology eventually reaching well beyond the movie theater, into the real world, where Disney parks could react to your mood with real-time Disney magic.
Disney’s researchers tested the system across 150 showings of several Disney films like The Jungle Book and Star Wars: The Force Awakens in movie theaters. Studios like Disney have used the reactions of test audiences to gauge early cuts of films for decades, often making changes to the edit or ending if the film wasn’t hitting as hoped. What’s different in this instance is the sheer amount of analysis Disney Research was able to produce with this methodology: 3,179 people generated 16 million points of data.As Disney Research scientist Peter Carr put it to Phys.org: “It’s more data than a human is going to look through. That’s where computers come in–to summarize the data without losing important details.” Impressively, the neural net isn’t just capable of summarizing takeaways. Within just a few minutes of watching a filmgoer, it could predict their facial expressions throughout the rest of the film.
Disney is far from the only company working on big data in media. Snap, most notably, actually measures the shot-by-shot performance of its Snapchat Shows. That means the company can see if a certain 10-second blip is losing its audience’s attention. Such insights are actually shaping the way Snap produces content.