Dissecting Movie Genres from an Audience Perspective : MTI Movie Classification Method

This research uncovers the critical deficiencies of the current genre-based movie categorization practice in the film industry and introduces a new method to classify movies. Using a large scale audience survey data, the authors cluster movies into nine distinct types based on eight audience-perceived movie characteristics such as 'fun', 'eye-catching', 'discomfort', 'feel-good', etc. The authors also validate their method by comparing 'movie types vs movie genres' in terms of their box-office revenue explanatory power. All three types of box-office revenues (opening week revenue, total revenue, total revenue per screen) differed significantly across movie types, while only the opening week revenue showed a significant difference across movie genres, suggesting that movie types may be a better predictor of a movie’s box-office performance than movie genres which have been frequently used in prior research on box-office performance prediction