Abstract

Film studios have for some time been using statistical methods to look at predicting aspects of a film such as the type and distribution of audiences, box office takings, and even the creative side. New online streaming services, including Netflix and Amazon Prime Video, also base most of their decisions on advanced data analytics and recommender systems. Such approaches, however, have a mixed track record, and most filmmakers still abide by William Goldman’s 1983 quote that when it comes to predicting the success of motion pictures, ‘nobody knows anything.’ To this day, the distribution of box office revenue and online consumption of movies remains highly skewed with asymmetric tails. In this seminar, we look at various approaches to modelling film performance and their key insights, and consider the importance of statistics and data analytics in the movie business in the future.