news
Study finds similarities in critic and audience reviews affect movie sales
The Observer | Thursday, October 13, 2022
Movie reviews are often studied in marketing research as a measurable form of online word of mouth, Notre Dame marketing professor Shijie Lu said.
Lu is a coauthor of a forthcoming paper in the Journal of Marketing that found that similarity in review content between critics and general users impacts movie demand.
Due to the vast availability of data, Lu said, movies are often the most convenient focal point of online entertainment industry research.
“Movies are also relatively new products, which have a short product life cycle,” Lu said. “It just makes it easier for researchers to understand how online word of mouth affects demand throughout the product life cycle.”
In the past, Lu added, movie marketing research has looked at the two sources of movie reviews: critics and the general user, separately.
“Some research has looked at how critical reviews affect box office sales, and the majority of research ignores the critical reviews completely,” he said. “They just focus on user reviews, because the sheer volume of user reviews is usually much higher than that from critics.”
In the study, Lu and his coauthors found that the content of both types of reviews may be more important than the actual evaluation of the film.
“We are saying that when both parties are talking about the same aspects of the movies, they are more likely to make consumers be aware of the movie,” he said.
Lu uses the award-winning potential of a movie as a topic example to help clarify the concept of topic consistency and how it might affect consumer demand.
“If critics are talking about the award-winning potential of the movie and if consumers are talking about the same attribute, this will make this particular movie attribute more memorable in consumers’ minds and therefore encourage them to check out this movie in theaters,” Lu said.
To test this hypothesis, Lu said his team needed to devise a way to measure the similarity in review content with text mining tools.
“If we want to process the review data on a large scale, we use statistical measures,” Lu said. “The particular measure we’re using, this research is called topic modeling [and] it’s a statistical modeling method.”
For the movie “La La Land,” Lu said, the text mining process found that there were 25 topics that could generally summarize all review content.
“Each review will have a weight along each of the topics,” Lu said. “We can summarize each review as a kind of a vector along the topics [and] we can use this vector between two reviews to measure their similarity, that’s the basic idea.”
Lu said his team aggregated critical reviews and user reviews, analyzing 10 user reviews for every one critical review.
“We look at the similarity for each pair, and then we aggregate the similarity across all reviews. So, what we did is to create a measure, which we call topic consistency, based on all the reviews written by users and critics, and when this overall similarity measure or topic consistent measure is high, we find that it’s usually associated with high box office revenue,” Lu said.
After gathering this field data to test the positive correlation, Lu said, his team ran lab experiments to check whether presenting subjects with different pairs of critical and user reviews with high or low similarity would change consumer willingness to watch a movie.
Eunsoo Kim, marking professor at Nanyang Technological University and coauthor of the study, discussed the application of the research results.
“The common topics generated by critics and users help potential moviegoers better evaluate and remember the movie by its attributes (topics),” she wrote in an email.
Now that they have established a link between content similarity and box office sales, Lu said movie producers may be able to take advantage.
“We are suggesting that producers can use our findings to create some common themes for critics and users to … see if they can nudge box office sales,” Lu said.
Kim clarified that this would probably mostly affect movies with middle-of-the-road ratings.
“The analysis shows that the impact of topic consistency is positively significant, and this association is more prominent when it comes to movies with mediocre reviews than movies with extreme ratings,” she wrote.
Beyond movies, Lu said, topic-driven promotion can be applied to television commercials, online video advertising and even other industries outside entertainment.
“This idea of topic consistency can be generalized to any other industry where you have reviews from both general users and the experts like fashion and electronic products,” Lu said.
Contact Peter at [email protected]
