Word-of-mouth (WOM) is a critical source of information for experience goods such as movies, with emotional content serving as a key driver of WOM sharing. However, existing WOM research has primarily focused on aggregate metrics such as volume and valence, overlooking the complex and dynamic role of emotions in WOM communication. To address this limitation, we investigate the temporal dynamics of emotional content in online movie reviews from pre-release to post-release period, aiming to uncover when and how emotional dynamics influence consumer responses.
We analyzed 26,055 movie reviews from the Internet Movie Database (IMDb) for 77 movies released in 2024 over six weeks. Using natural language processing, we integrated lexicon-based sentiment analysis (SentiWordNet) with chi-square feature selection to extract and quantify specific emotional content over time.
Emotional content in WOM peaks during a movie's release week, indicating heightened emotional expressions. Early reviews demonstrate a positive bias, possibly influenced by the self-selection of loyal fans. As WOM accumulates, emotional asymmetry decreases, characterized by a decline in overall and positive emotional content and an increase in negative emotional content, consistent with informational sufficiency over time.
This study contributes to the WOM literature by introducing a dynamic two-stage framework of WOM communication. The findings show that emotional content changes over time from an early emotion-driven stage to a later stage characterized by more balanced, information-driven evaluations. The study provides new theoretical insights into how the informational role of emotional content evolves over the product lifecycle, moving beyond static aggregate measures of WOM. The results provide movie studios and platform managers with a temporal framework for leveraging real-time emotional analytics during the critical release week and throughout the movie's lifecycle to better manage consumer perceptions and marketing strategies.
