The Storytelling Cipher: Mapping Precise Story and Character Mechanics to Box Office Returns

Our Data & Analytics Project held “The Storytelling Cipher: Mapping Stories & Characters to Box Office Revenue” Tuesday, December 6, 2016 at the USC School of Cinematic Arts.

This study leverages the Dramatic taxonomy of film narrative to infer which scene-level character and story attributes generate more box office returns, by genre. We are extending this study to ads and movie trailers.

The project researchers used machine learning to map 70+ story attributes for 300 films to their box office returns to extract which story mechanics or character features in film generated the most revenue. This was the first time granular story and character mechanics have been used to predict box office returns, which opens up many avenues to make more data-driven creative and development decisions throughout the industry.

What’s a good story? The question has been hanging without a scientific answer since the dawn of man. It seems that a story’s lack of clear mathematical structure and universal taxonomy would relegate such classification of stories to the qualitative – and highly subjective- empire of critics and … people.

Until now.

The event presented results from the research, discussed applications for the development and creative process, and outlined next steps.