How do emotional and language tones influence the abandonment rate of a movie or TV show?

Abstract

This research evaluates the effect of emotional and language tones perceived from movies and TV shows on user abandonment or users abandoning the video without completion. The motivation behind this research emerges from the increasing popularity of the Over-the-top (OTT) platform. These platforms have revolutionized the accessibility and usability with which people consume content over the internet. The increase certainly entails several implications for all stakeholders, including the OTT provider, the consumers, and the content creators. For the OTT providers and consumers, this represents the two sides to the same coin or finding the right content. The content creators, however, are trying to cope up with the increased demand. In addition, if the content being put up is often left abandoned by the users without completion, these implications are only intensified. Therefore, this research tries to shed light on a tangential solution to the aforementioned implications. Moreover, this research analyses five emotional tones and three language tones. Emotions are abstract. The literature describes how emotions often entail an action tendency of either approach or avoidance. Watching a movie or a TV show is a complex audiovisual phenomenon. Therefore, once the viewer is immersed in the fictional narrative being presented, the tones perceived are theorized to impact engagement. For emotions, this impact is dependent on dimensions such as valence, pleasantness, and activation, whereas, for language tones, this impact generally seems positive. Therefore, this research speculates that positive emotions should affect abandonment negatively and negative emotions positively. Language tones, however, are speculated to decrease abandonment generally. This research exploits the Natural Language Processing technologies to extract these tones from English subtitles. Additionally, an extensive dataset is obtained, containing information regarding the consumption of content on the platform. On this data, this research employs logit regression to study the relationship between tones and abandonment. The findings indicate that positive emotion joy decreases abandonment, whereas the impact of negative emotions is diverse. Similarly, mixed results are obtained for language tones. Future research should explore the possibilities of using subtitles in combination with the narrative and the visuals to extract tones. Furthermore, the possibility of these tones exhibiting a non-linear relationship with engagement shall also be explored.