Using Neurologic Immersion to Predict Movie Ticket Sales
Immersion predicts movie ticket sales with 61% accuracy
We live in Los Angeles, so like most Angelenos we have one degree of separation between us and Hollywood. Recently, a movie producer friend asked us to investigate why so many movies miss the mark (often badly, who thought Gigli was a good idea?). Could neuroscience be used to predict movie ticket sales? If so, would neuroscience measures have predictive ability beyond the typical survey questions of “familiarity,” “liking,” and “intent to watch”? We love a challenge, and we had run experiments using video clips in our academic research, so we thought we would give it a try.
In science it is always best to start small, so picked two recently released movies and tested if we could predict movie ticket sales using our proprietary neuroscience measure of immersion called the Immersion Quotient.
The Immersion Quotient varies from 0-10 and is benchmarked across thousands of people, with higher values denoting greater immersion. The Immersion Neuroscience platform provides real-time measures of immersion. Our software also quantifies peak immersion and shows when people are frustrated and disengage with content.
In the present study, we recruited 70 adults 18-49 (57% men; average age: 25) to view two movie trailers, “Fault in Our Stars” and “Get Hard” in the laboratories of Immersion Neuroscience. Participant neurophysiology were collected using our Immersion Neuroscience platform and sensors. Participants completed a brief survey assessing content liking and information retention after each presentation.
Fault in Our Stars grossed $48M in its opening weekend whereas Get Hard, with a bigger budget ($40M vs. $12M) and greater star power, only grossed $33.8M in its opening weekend (Figure 1).
Self- reports of being entertained by the trailer, enjoying it, and even wanting to share the trailer with friends did not predict box office success. Get Hard was higher on all liking measures than Fault in Our Stars, yet Fault in Our Stars opening week ticket sales were 35% higher than Get Hard. (Figure 2). This inability to predict actions (sales) from self-reports is something we have found in our previous studies.
In contrast to “conscious” viewer responses to the movie trailers, the Immersion Quotient, our proprietary measure of unconscious neurologic engagement, predicted box office sales with high accuracy (Figure 3). Fault in Our Stars generated a 41% higher immersion than Get Hard, nearly matching the difference in first week movie sales.
These results gave us confidence that immersion was useful in predicting movie ticket sales. But movie studios use sophisticated statistical models using economic data to predict movie performance. Would immersion continue to predict ticket sales when economic factors were included?
For this study, we scaled up our analysis to nine movies of varying quality, including one that had not yet been released. We recruited 49 adults 18-70 (43.5% men; average age: 27.5) to view movie trailers for “47 Ronin”, “American Hustle”, “Birdman”, “Her”, “Inherent Vice”, “Dawn of the Planet of the Apes”, “The Hundred Foot Journey”, “The Master”, and “Transformers IV” in the laboratories of Immersion Neuroscience. Participant neurophysiology was collected as in Study 1. Participants also completed a brief survey assessing familiarity with the movies and content liking. Inherent Vice was included so we could make an out-of-sample prediction; it had a movie trailer but had not yet been released at the time of our study. Transformers IV and Planet of the Apes opened at over $100 million, while the other movies all opened under $40 million; Birdman opened under $3 million, a movie that would eventually win four Academy Awards, including Best Picture.
As in Study 1, the self-reported likability of the movie trailer did not predict first week movie sales. In contrast, immersion was strongly predictive of ticket sales. Immersion alone predicted 25% of the variation in first week sales. Importantly for this small sample of eight movies, a movie trailer’s immersion correctly classifies movies as having high or low ticket sales with 61.1% accuracy.
So what else predicts ticket sales? We looked at two-dozen economic variables, including the movies’ stars (star power), directors, genre, the number of views of the Wikipedia page about the movie, and movie budget. The most important factor is, not surprisingly, the movie’s budget. A bigger budget means bigger stars and a more famous director and more money generally spent on marketing. A movie’s budget explains 36% of first week ticket sales (Figure 4). Importantly for model-builders, immersion and a movie’s budget are statistically unrelated so that immersion substantially improves the predictive ability of standard economic models of ticket sales.
We calculated immersion for Inherent Vice and built a predictive model that included immersion and its $20 million budget to predict first week ticket sales. Inherent Vice’s broad opening was January 9, 2015 (its limited opening was December 5, 2014 in five theaters). January is a typical “dump” month so we adjusted the model to account the month of release. The model predicted first week sales of $7 million. Inherent Vice’s actual ticket sales for the week ending January 15 were $5.3 million. Immersion provided an improved prediction of first week ticket sales over standard economic models.
Neurologic testing of movie trailers shows that brain activity predicts ticket sales in a manner that audience self-reports does not, while improving models that only use economic forecast variables (e.g., seasonality). Measuring neurologic immersion, as we did here, can be of substantial value to movie distributors when they seek to determine how much to invest in promoting a movie. It can also guide directors on edits and can replace self-report scores from focus groups or dial testing. Using these approaches, neuroscience can improve the movie-watching experience for consumers and movie ticket sales for studios.