By Emerson College, Special for USDR
A study conducted as part of a School of Communication research course at Emerson Collegefound that Twitter could demonstrate the influence of word of mouth on movies. Specifically, the study found a strong correlation between sentiments expressed on Twitter and movie revenue.
Historically, word of mouth has been thought to be a primary determinant of a movie’s success. This is evident by the absence of a formula to guarantee a predictable return on investment and the lack of other reliable predictors such as genre, cast and budget. As a result, recommendations from moviegoers, whether positive or negative, can have a significant impact on a movie’s performance.
This study consisted of manually coding a sample of 12,000 tweets out of a universe of 800,739 tweets from 17 movies in the fall of 2016. The movies selected for this study were those that were scheduled for wide-release and were not holiday-themed or released around the holidays to eliminate the possibility of data being skewed by unique audiences. In particular, hashtags used by studios to promote their films were used to define a pool of tweets related to these movies.
To conduct the analysis, a random sample of 400 hashtags were analyzed each week after the release of each movie for as many weeks up to four weeks that the data would allow. For several low performing movies, researchers were only able to gather a week of data and, in other cases, they weren’t able to gather any data at all. In the end, out of the 17 movies for which researchers collected tweets, only 11 had at least a week of data that met their sample size criteria.
Once the tweets were coded, researchers then compared the average weekly sentiments with the average weekly percentage change in revenue for each movie. Weekly revenue was defined as the percentage change from weekend to weekend to allow researchers to compare this data to another study which used this methodology. The correlation analysis between tweet sentiments and revenue found an r value of .75, which indicates a strong relationship since a correlation coefficient of .5 is considered high, .3 is considered medium and .1 is considered low. And, even though the sample size of 11 movies is small, this analysis found a p value of less than .01 which means that there is a 99 percent chance of it being accurate. Exhibit 1 below illustrates the relationship between hashtags and revenue.
Moreover, this study was developed to further confirm the findings of another study conducted by this group of researchers on the relationship between word of mouth and movie revenue. This earlier study used revenue as a proxy for word of mouth to see if significant differences could be found between high-performing films versus low-performing films.
Conducted in the fall of 2015, this study analyzed the role of word of mouth on Oscar-nominated movies and Razzie-nominated movies over the previous 20 years. Then these movies’ average percentage change in revenue over their initial four weekends was averaged during this period as shown in Exhibit 2. This resulted in an average percentage change between Oscar nominees and Razzie nominees of -30 percent.
“This study offers more evidence of the influence of word of mouth on movies,” said Owen Eagan, an Executive in Residence in the Department of Communication Studies. “Specifically, it confirms our earlier research regarding the significant relationship between movie buzz and revenue.”
SOURCE Emerson College