How many Screens do you need for a Blockbuster movie?
A Data Science project
I have been putting together a dataset from multiple sources like BoxOfficeMojo, IMDb and Rotten Tomatoes. Box Office Mojo has a great repository of inflation-adjusted weekend and total box office collections, and the OMDb API allows for a clean and efficient way to pull combined data from IMDb and Rotten Tomatoes if you have the film's IMDb ID. I have grander ambitions for this dataset, more data and newer connections, but while I put that together, I thought it'd be interesting to examine the current dataset one aspect at a time.
First up, let's look at the number of screens it takes to make a successful movie.
Let's get a lay of the land.
The dataset covers all movies from 2005 and 2014, and movies released up to Oct 2015, with releases in over 500 screens.
Why 500? Of the approximately 6,350 movies released from Jan 2005 to Oct 2015, here's a histogram of movie releases and numbers of screens. A little over 71% of the total films in this dataset (4,552 to be precise) had an opening weekend release in under 100 movie screens. Just under 75% of the films released in under 250 screens, and a little over 76% of the films released in 500 screens or less.
Think about that.
More than 2/3rd of the movies released in the United States in the last 10 years were released in under 100 screens.
So, after we've carved out films that screened in under 500 screens, we are left with a little under 24% of the original dataset. Let's see what this dataset looks like:
The sweet spot for a wide-release film seems to be about 3000 screens. 2000 and 2500 screens are popular as well. Studios most likely decide the strength of their screen-count well before a movie is released, based on their internal projections of how the movie will perform. That would explain the fairly clean stratifications by Opening Weekend range and Screen Count range. That is, except for the unfortunate films in the lowest opening weekend bucket ('Under $7M') that seem to be creeping past the 2000 screen count, well into the 3000s.
Releasing in over 4000 screens is almost a necessity for a $150M+ opener. Some $100M+ openers seem to have snuck in at over 3500 screens.
How does this look, split by Studio Groups?
The color gradations in each individual Studio's panel indicates how much money the film made. The studios in the first row, and a couple in the second row, specialize in very wide-release films (presumably the kind that make over $200M in an opening weekend), but also release movies all along the screen-count spectrum. Some partnerships like Paramount/Dreamworks, Sony/Screen Gems, WB/New Line come together for specific wide-release projects, while the more niche studios like New Line, Focus, Relativity and Summit stick to medium-to-wide release movies, suggesting a focus in demographics.
The Summer Blockbuster Season
It is starting to look like there's a correlation between the number of screens and the total amount of money a movie makes in its opening weekend. This intuitively makes sense, since theaters, seats, showtimes are finite, ticket prices are pretty much fixed. While the max might be capped, the opposite is certainly true for the minimum. Movies could release in over 2000 screens and still make under $5M (Mortdecai, anyone?).
Let's see what the summer blockbuster season looks like by screen-count.
The summer months show a clear uptick for the average screen count of a movie release, followed closely by the Nov-Dec holiday season.
How does adding in the Opening Weekend ($M) groups change that? Let's take a look.
More movies are released in the summer in every strata, except for limited release movies expected to make less than $7M.
Well, now. That's interesting.
The 'Under $7M' buckets take a dip in screen-count during the summer months - these are most likely movies meant for a limited release, and with the summer months over-crowded with releases, the studios likely release less of these niche market films.
On the other hand, the higher Opening Weekend buckets ($100M and higher) occur almost exclusively in the summer, with the exception of a blip in November - caused collectively by a handful of Hunger Games, Twilights and Harry Potters.
testing The Correlation
By now, it is pretty clear, there is a definite correlation. How strong is the correlation, though? Let's check.
While this suggests a high correlation, let's see what a linear regression tells us about the two numbers.
A quick correlation test shows an 81% correlation between the two variables. A linear regression however, yields a fairly unexciting 66.3% Adjusted R^2.
Why is that?
Possibly due to films that are medium-to-wide releases that make less money than anticipated, and contribute to the underwhelming performance of the regression model. Perhaps there's a way to improve the model - could user ratings help? Critic ratings? That's for another post.
Meanwhile, let's take a closer look at screens by studio.
Studios & Screens
Warner/NewLine seems to work in a pretty tight band of screen-counts, while Paramount, and ironically, Focus, seem to work in a very wide spectrum of screen counts for movie releases.
Opening Weekends and Averages by Studio
Combining the opening weekend numbers with the number of screens, we come up with an average opening weekend $ per screen ratio. What does that per-screen-average look like across the various movie studios?
In this visual, the studios are ordered by the number of total film releases in over 500 screens, in the past 10 years. Fox leads with the number of releases, but doesn't seem to have a very high per-screen-average. Disney does a decent job among the top five studios, while Paramount/Dreamworks leads the pack.
The outliers tell a more interesting story. The labeled outliers all have per-screen-averages greater than $30,000. Most of the names are recognizable franchises, and the color gradation shows that these films released in a very high number of screens in their opening weekend. The notable exceptions are Borat and Hannah Montana, with low screen-counts, but a very high per-screen-average - a testament to the buzz marketing in the former case, and the cult Disney following of Miley Cyrus in the latter.
The $200M+ openers have per-screen-averages of $48K-$50K.
Finally, the count of movies released by studios, that fall into each of the Opening Weekend buckets stacks up like this:
Disney and Universal have the only two films that cleared $200M in their respective opening weekends - The Avengers (2012) and Jurassic World (2015). The other studios have a handful of $100M - $200M openers, while studios like Relativity, Focus play exclusively in the limited-release, sub-$25M opener space.
Whether by design or not, Warners seems to have exceeded all other studios in the number of films released in the $7M-$25M category, while Universal tops the $25M-$50M category.
For an avid movie-watcher like me, movies are already an endless source of fascination and entertainment, but the data behind the numbers is beautiful and fantastic too.
The studios have tried, and will continue to try to boil the art of movie making down to predictable formulas - not the least because, hey, you need to make money to stay in business. It's interesting to see the kinds of screen-counts and per-screen-averages a movie needs to have a successful opening. The outliers tell their own story - almost every film with a $30K or higher average-opening is a franchise - is it a surprise then that studios continue to make more franchise films?
At the same time, it is also heartening to see that over 75% of films released are niche-market films (although, to be fair, they probably hoped to make more than $7M) - which is a glimmer of hope for the kind of diversity in film-making everyone looks for... after they're done watching the franchises.
See you next time.