6 Conclusion

As already mentioned, it was logical and expected before starting this project to find a positive correlation in budget and profit. However, it would have been interesting to be able to better understand the huge stakes involved in film financing, to understand what makes people dare to put several hundred million dollars at stake, for a return on investment that may seem somewhat random. Therefore, it is interesting to note that if one decides to invest in a very high-budget film, it seems preferable from a financial point of view to invest in a film with a budget above $100 million, where the correlation between budget and profit is strongest, than in a film with a budget in the $10-$90 million range, so it may not seem intuitive although as Bill Gates said “To win big, you sometimes have to take big risks”, it seems to hold true in film production industry, at least nowadays. Why this barrier around the $100 million budget? It seems important that this barrier is valid at a given time (2020) and is by no means fixed, that it is necessary to take a closer look at the cost structure of productions in order to understand this, and that this structure will most certainly change in the years to come, especially in view of the economic crisis that is coming, which is (and will be) having a particularly negative impact on the arts, culture and leisure sector.

However, on a more positive note, it can be seen that many films considered to be low budget (i.e., from $1 to $10 million) have profits reaching $100 million in many cases, demonstrating the possibility for even the smallest ones to succeed against the industry’s Goliaths. This by producing, as shown by the data, films where the genre makes the budget less strongly correlated with revenues, such as horror or war films. This conclusion is however nuanced, many other parameters coming into play such as the trend of the moment, the period of release of the film during the year and many more. It is nevertheless an interesting path to explore in a future project in order to develop, for instance, machine learning skills to be able to determine how a film will be welcomed by the movie lovers even before its release. In addition, one of the surprises of this project was certainly not to find statistically significant results regarding the impact of the production studio on the correlation between budget and profit. Faced with a time constraint to make this project, we had to make choices and decided to focus on other aspects of this project, including to learn whether a trend based on the released year of the film could have an influence on the budget-income correlation. In the end, it was rewarding to be able to observe this upward trend since the 2000s.

Now, to conclude, many interesting questions seem to arise from this project, for which a more technical knowledge of the film industry will be necessary in order to determine, among other things, whether a high-budget film generally generates higher profits simply because it sells its rights more expensively to distributors? But also, for example, to be able to determine in advance the rating a film will get from the media, and much more. All in all, these are exciting opportunities to deepen our knowledge about both the film industry and Data Science at the same time. To be continued.