Over the last several months, Largo has been running a series of experiments aimed at refining the accuracy of the company’s regional movie gross earning predictions.
Why is this important? Well, we will get to the results and what they mean in just a second.
You might be surprised to learn that the total yearly gross revenue for the European movie industry is roughly equal to that of the United States.
This makes it a major market for those movies that can succeed.
Understanding how audience tastes vary from country to country, and within each country, is at the very heart of this success.
Within the European market, Spain, Britain, Germany, and France are the main players.
Spain has a truly international reach thanks to Spanish being the 3rd most spoken language in the world. Today, Spanish is spoken by some 420 million people, with 350 million people speaking it as their native tongue. This makes it bigger than English which has 340 million native speakers.
There are no less than 35 former Spanish colonies that have Spanish as their official language.
An article in Cineuropa noted that “Box-office takings by Spanish films outside of Spain during 2017 grew by precisely 19.45% compared to 2016, when they grossed €80,951,788. The total sum ranked in by Spanish movies released in other countries last year rocketed to €96,709,469”.
Naturally, this international influence on the Spanish movie industry and the films that it produces makes Spain an incredibly difficult but exciting market to try to predict.
This is exactly why we chose Spain to put our sophisticated artificial intelligence software to the test.
More Important Than Ever To Get It Right
Ever since early filmmakers discovered that they could charge people to watch their movies, the need arose to predict what kinds of movies people would want to watch and which they wouldn’t.
Believe it or not but very little changed in that process on over 100 years!
A studio executive or producer sits down and reads a script. From their gut feeling, this person decides whether, in their opinion, the script has a chance of becoming a successful movie.
All this person has is their intuition of what has worked well in the past and what makes an appealing movie.
While this system managed to make the names of a few ‘super producers’, including David O Selznick, it actually ruined the careers of far more promising film producers and directors than it made.
Over the years, this system, when coupled with high budget advertising and star power, managed to keep enough bums on seats to keep the industry healthy.
With no better solution, companies essentially placed their entire financial future in the hands of just a few individuals, all with the hope that they would be able to pick the right scripts.
That was, of course, until the exciting new technology of artificial intelligence and big data analytics arrived.
The great news is that with the help of AI data-assisted moviemaking services, movie producers can now get accurate gross prediction data (by region) to empower them to make the right choices.
Not only will this exciting new technology help them pick the scripts that will see good box office returns but it will also help them budget their movies’ production costs accurately.
This is a win-win situation for both the movie industry and producers everywhere.
But before we get ahead of ourselves, this all depends on accurate regional gross predictions, which is why Largo has been putting its powerful artificial intelligence program through its paces.
LargoAI: Practice Make Perfect
Nothing speaks louder than results.
With this in mind, let’s take a closer look at LargoAI’s movie opening week predictions for Spanish releases in February in Spain.
NOTE: ‘Wide releases’ (300+ screens)
- Hasta Que La Boda Nos Separe (Spanish film)
Release Date: Feb 14th 2020
LargoAI Predicted range: 600,000 – 1,220,000 (USD)
Largo best guess: 868,000 (USD)
Actual first weekend revenue: 800,000 (USD).
Verdict – Success
- Sonic The Hedgehog (US film)
Release Date Feb 14th 2020
Predicted range: 950,000 – 1,800,000 (USD)
Actual first weekend revenue: 1,800,000 (USD).
- The Invisible Man (US film)
Release Date Feb 28th 2020
Predicted range: 797,000 – 1,500,000 (USD)
Actual first weekend revenue: 1,100,000 (USD).
Verdict – Success
NOTE: ‘Limited release’ (Less than 100 screens)
- Para Toda La Muerte (Spanish film)
Release Date Jan 31st, 2020
Predicted range: 29,000 – 152,000 (USD)
Actual first-weekend revenue: 47,000 (USD)
Verdict – Success
- Adú (Spanish film)
Release date Jan 31st, 2020
Predicted range: 456,000 – 1,000,000 (USD)
Actual first weekend revenue: 1,200,000 (USD)
Verdict: Failure
You Learn Something New Everyday
Despite scoring an accuracy rate of 75%, LargoAI’s failure to accurately predict the gross range of Adú highlights why Largo is constantly working to improve the accuracy of its predictions by conducting tests.
Our specialists analysed the post-release data and discovered that the movie had a relatively large-sized budget compared to other Spanish movies. Therefore, our model was not as well trained in analysing this data trend.
As a result of this discovery, more data is currently being added to improve the model’s accuracy in this area. This will ensure that the program will identify this data trend in the future and will factor it in when producing gross estimates.
Note: Largo provides both its SaaS as well as conventional customers with ranges for ‘70% confidence’ and ‘90% confidence’. When we apply the 70% rule in the case of Adú, we correctly predicted this revenue in this case.
Some might argue that the similar predictions can be done with the greenlighting experts, which is true. However, with two big differences: their accuracies will be as low as 50% and they will need significant time for such analysis, while Largo.ai can make this analysis in 5 min by simply uploading the script and metadata. More accurate at your fingertips.
Over the last several months, Largo has been running a series of experiments aimed at refining the accuracy of the company’s regional movie gross earning predictions.
