A few weeks ago 'Dark Phoenix' was a very hot conversation topic in the movie business, not for the movie itself, but because it symbolized the difficult Fox acquisition by Disney.
While Disney movies have done fairly well in 2019 to put it lightly (5 movies generating over $1Billion so far, and it's only August with "Frozen 2" and "Star Wars 9" still due....), Fox is under very tight scrutiny and 'Dark Phoenix' has been heralded as the barometer for Fox movie success/failure. Especially after Disney CEO Bob Iger declared during quarterly earnings call that “the Fox studio performance … was well below where it had been and well below where we hoped it would be when we made the acquisition.”
While technically profitable thanks to higher-than-expected international results (latest IMDB results have $250M gross worldwide revenue with analyst forecasts around $280M for an estimated budget of $200M), 'Dark Phoenix''s performance will continue to be observed and scrutinized.
But this is just one movie we're talking about! Is it really fair to reduce a movie studio the size of Fox to a single movie? 20th Century Fox has been around for a while, and I was curious to compare their historical performance compared to Disney's.
First things first, data is required! I pulled all of Disney's and Fox's movies from IMDB, including key metrics in the process: IMDB rating, metascore, duration, genre, estimated budget, gross revenue...
Naturally some filtering was required to focus solely on non-short non-documentary movies released after 2000 with non-missing data.
In addition to Fox and Disney movies, I created a third category for movies produced by both.
The next plot compares the performance of each of the 512 movies that matched are criteria (331 Fox, 181 Disney), the x-axis is the worldwide revenue, y-axis is IMDB rating. Bubbles are color-coded by company and the size of the bubble is proportional to the movie's ROI (gross worldwide revenue / estimated budget):
A log-transform of the x-axis provides more insight into the lefthand-side blob, but loses the magnitude of the super high-performing movies:
It would appear that Disney dominates the high revenue end of the spectrum, with 22 out of 23 of movies generating over a billion dollars coming from Disney, the sole exception being 'Transformers: Age of Extinction', things are much more murky under that threshold.
Given the difficulty of separating the two classes I tried fitting an XGBoost model which was capable of classifying movies from each company solely based on IMDB rating, metascore, budget, world revenue and movie duration with over 75% accuracy on the test dataset, pretty remarkable, wouldn't you agree? Although to be fair, a simple logistic regression was capable of achieving a reasonable 65% accuracy.
It'll be interesting to see how the next wave of Fox movies perform, perhaps attempt a before/after comparison of the Disney acquisition...
