Study finds reviews from critics and audiences have similarities to movie sales


Movie reviews are often studied in marketing research as a measurable form of online word-of-mouth, said Shijie Lu, a marketing professor at Notre Dame.

Lu is co-author of a forthcoming article in the Journal of Marketing that found that similarity in review content between reviewers and general users impacts demand for movies.

Due to the wide availability of data, Lu said, movies are often the most convenient focal point of online entertainment industry research.

“Films are also relatively new products, which have a short life cycle,” Lu said. “It just makes it easier for researchers to understand how word of mouth online affects demand across the life cycle of the product.”

In the past, Lu added, research on movie marketing has looked at the two sources of movie reviews: critics and the general user, separately.

“Some research has looked at the impact of critical reviews on box office sales, and the majority of research completely ignores critical reviews,” he said. “They focus solely on user reviews, as the volume of user reviews is usually much higher than reviews.”

In the study, Lu and his co-authors found that the content of both types of reviews can be more important than the actual rating of the film.

“We’re saying that when both parties talk about the same aspects of movies, they’re more likely to drive consumer awareness of the movie,” he said.

Lu uses the award-winning potential of a movie as an example topic to help clarify the concept of topic consistency and how it might affect consumer demand.

“If critics talk about the movie’s award-winning potential and consumers talk about the same attribute, it will make that particular attribute of the movie more memorable in consumers’ minds and therefore encourage them to watch that movie in theaters,” Lu says.

To test this hypothesis, Lu said his team needed to find a way to measure the similarity of review content with text mining tools.

“If we want to process review data on a large scale, we use statistical measures,” Lu said. “The particular measure we use, this research is called topic modeling. [and] it is a statistical modeling method.

For the movie “La La Land,” Lu said, the text mining process revealed that there were 25 topics that could generally summarize all of the review content.

“Each review will have a weight on each of the topics,” Lu said. “We can summarize each review as a kind of vector along the topics. [and] we can use this vector between two opinions to measure their similarity, this is the basic idea.

Lu said his team aggregated critical reviews and user reviews, analyzing 10 user reviews for each critical review.

“We look at similarity for each pair and then aggregate similarity across all reviews. So what we’ve done is create a metric, which we call topic consistency, based on all reviews written by users and reviewers, and when that overall similarity metric or topic consistency metric is high, we find that it is generally associated with a high box office income,” Lu said.

After collecting this field data to test for positive correlation, Lu said, his team conducted lab experiments to test whether presenting subjects with different pairs of reviewers and users with high or low similarity would change consumers’ willingness to watch a movie.

Eunsoo Kim, a tagging professor at Nanyang Technological University and co-author of the study, discussed the application of the research findings.

“Common topics generated by critics and users help potential moviegoers better assess and remember the film by its (topics) attributes,” she wrote in an email.

Now that they have established a link between content similarity and box office sales, Lu said film producers could benefit.

“We suggest producers use our findings to create common themes for critics and users to… see if they can boost box office sales,” Lu said.

Kim clarified that this would likely primarily affect films with mid-range ratings.

“The analysis shows that the impact of topic consistency is positively significant, and this association is stronger when it comes to films with mediocre reviews than films with extreme ratings,” she said. writing.

Beyond movies, Lu said, theme promotion can be applied to TV commercials, online video advertising and even other industries outside of entertainment.

“This idea of ​​topic consistency can be generalized to any other industry where you have reviews from general users and experts like fashion and electronics,” Lu said.

Contact Peter at



Comments are closed.