As part of the Association of Information Systems (AIS) annual analytics competition, I competed in a group of three on the NBC Universal Analytics Challenge. As part of the competition, we were provided decades of rough film industry data and tasked with finding the key metrics that indicate the success of a movie at the box office. After significant data cleaning and multiple analyses via SQL Server, Tableau, and multivariate regressions using R, we derived the primary insights for predicting the success of a film. We placed first in the nation for both our presentation and report, competing amongst students from hundreds of universities across the country.
You can read more about the competition and my team’s success from Belmont University.
Sequels, remakes, and movies in a series consistently rank as the highest revenue generators at the box office. By analyzing the top grossing movies in each quarter between 2010 and 2017, we found that 58 percent were considered a sequel, remake, or in a series. A perfect example is the year 2014, when seven of the twelve top three grossing movies per quarter were in one of these categories. Only five were originals.
Sequels, remakes, and series accounted for only 8 percent of all movies made between 2010 and 2017. However, they garnered 43 percent of all movie revenue in the same period. It’s obvious that they generate a disproportionately high level of market share. But it is important to note that this level of market share has diminishing returns.
In a series, the second movie earns an average of 5 percent less than the first, while the third movie earns 19 percent less than the first. If you’re releasing an original film, timing your release to where you are competing against the third movie in a series leaves the most room for potential market share. When developing a movie series, it is generally the case that the first two films perform significantly better than the third. Regardless, it seems that successful films hold a history with their audience. That’s why sequels, remakes, and series generate such high revenue.
A team’s ability to work together is crucial to the success of any project. So it’s easy to see why combining strong actors and a strong directors yields significantly higher return. We analyzed strong actor-director combinations by identifying pairs that worked together on three or more films and looking at their average ROI per movie. Between 2000 and 2017, movies with strong actor-director pairs held an average return of $23 million. By contrast, the average return for all movies was $6 million. Strong actor-director pairs yielded nearly four times the return. Of the movies with these power pairs, exactly 50 percent fell under comedy or adventure. Based on this, we can accurately assume that a movie with a strong actor and director combination in either comedy or adventure will be certain to yield a successfully high return.
Using multiple regression analyses, we constructed three predictive models. One for small budget movies, one for large budget movies, and one for our entire sample. To reflect the most recent box office trends, these models included only movies made after the year 2000. Our initial models used 44 variables and we hypothesized that all or most of these 44 variables would be necessary for our model to accurately predict box office success. But we couldn’t have been more wrong. Only 10 of the 44 variables were statistically significant. Of those 10, only one had a strong positive correlation with box office revenue across all models—budget. Our research shows that if you want to fill a theater you should focus first on increasing your budget. More importantly, it also shows that other factors typically associated with a film’s success, are not consistently correlated with it. Our final model has an r-squared of 0.61 and a residual standard error less than one standard deviation from the mean. The accuracy of this model shows that it has the potential to be a powerful tool for predicting box office success, and in turn increasing revenue.
Our three investigations revealed the most influential factors in a film’s success. History tells us that sequels, remakes, and series consistently outperform originals. Team chemistry tells us that strong actor and director pairs, especially in comedy, are bound to do well. And predictability tells us that time of year, weather, and economic trends have little impact on success, despite what one might think. Using these simple insights, we can make better and more impactful films.