Choose Language Hide Translation Bar
LavadaB
Level I

Netflix or AMC: Predicting Release Strategies in the Age of Options (2021-US-EPO-813)

Level: Intermediate

 

Lavada Blanton, Business Analytics Student, Oklahoma State University

 

As the challenges of COVID-19 made more businesses switch to virtual resources, the film industry was no exception. While billions are staying indoors and taking socially distanced precautions, production companies must decide the risks involved in the traditional movie release in theaters. Alternatives such as Netflix, Amazon Prime Video, and HBO Max were once frowned upon by critics but now have become respected players in the film industry. We now know that some movies released during COVID-19 were less than critically favorable but recent hits such as Godzilla vs Kong and Soul have given a glimmer of hope to decision makers. This project analyzed 1,605 movies, released through either the traditional movie theater format or through a streaming service since 2010. We look at key indicators in box office success such as IMDB score, critic reviews and audience reviews to evaluate the best “Movie Mix” to be distributed either in theaters or on a streaming service. In order to compare distribution type, box office success in streaming movies is predicted based on a theatrical release model. Then box office revenue is compared between streaming and theatrical to profile movies based on categories such as Genre and Release Month. Finally, a decision tree is used to streamline recommendations. These recommendations can be used by production companies in formulating the optimum release strategy and resource allocation.

 

 

Auto-generated transcript...

 


Speaker

Transcript

Mike Anderson We are now officially recording.
  All right, you understand this is being recorded for use in the jump discovery summit conference and will be available publicly in the jump user community do you give permission for this recording and use.
  Yes.
  Hello, my name is Lavada Blanton. And, I am a graduate student at the Oklahoma State University's Masters in Business Analytics and Data Science program. My project is called Netflix or AMC
  Predicting Release Strategies in the Age of Options.
  First to go a little bit introduction of what my project is about.
  Covid-19 cause production companies to decide the risks involved in the traditional movie releases in theaters.
  Streaming services such as HBO, Netflix and Amazon were once frowned upon by critics, but now have become a respected key player in the film industry.
  This project looks at key indicators and box office success, such as IMDB score, gross revenue, and critic reviews to evaluate the best movie mix to be distributed, either in theaters or on a streaming service.
  And the objective for this project is to predict whether movies, should be distributed through streaming services or box office. Success was measured by an IMDB score of 5.5 or higher, which is above average.
  I sampled 1913 movie titles released between the year 2010 and 2021.
  I had various attributes, such as genre, premiere date, duration budget, theatre box office revenue, and content rating, which were used to determine the best movie mix on the respective release types.
  As you can see right here, there were far less streaming movies than theatrical movies. To counteract that I did reduce it to 5.5 or higher for IMDB scores, and that evened it out a little bit.
  Next, I will go through my approach. First, I did data preprocessing and data collection. This involves gathering data from IMDB, Rotten Tomatoes, and the numbers that come.
  Next I filtered movies released before 2010 and took those out of the data set. And, I created a Covid flag for movies released between March 2019 and March 2021.
  And as you can see underneath here, there were some transformations specifically in duration, the pre- and post-transformation is shown here.
  Next, I did a box office prediction for streaming movies. This was produced with a neural network with a sample of theatrical movies in JMP Pro.
  Finally, I did a decision tree. This was filtered out. The data set was filtered out with movies from 5.5 and higher only. This reduced the data set to 1400. And, with the sample I was able to use a target of either streaming or theatrical for a categorical decision tree.
  The use cases for this data is an example in 2019
  the global film industry is worth around $136 billion. This means that this is a lot of risk and reward involved in every aspect of how a movie is made.
  And these movie mixes could be used by executives in top production companies, streaming services, or even theaters to make informed decisions about future movies.
  Below is a list of the top five highest grossing movies of all time. And, as you could see
  they are all 7.8 or higher. So, this cut does kind of show that these high IMDB scores do translate to box office.
  Okay next i'm going to dig down deeper into the models that I used in JMP Pro.
  First, I did a neural network. This neural network was created to predict box office sales in streaming movies. As you can see, we had an R square .87.
  And this is fairly good for the neural network. Next I did a stepwise variable selection. This was used to choose what variables would be more most suitable for the recommendations.
  Finally, I did a bootstrap forest to create the final recommendations with the target being distribution type.
  Finally, I have my recommendations. I have two movies, based on streaming, and two categories, based on theatrical.
  First, I have Cheap Feel-Good Comedies. These have an average duration of about 91 minutes. They're equal comedy and drama, as well as rated R.
  And then have an average budget of $2.1 million. Next, I have Big Budget Thrills. We have an average duration of 147 minutes. A genres
  50/50 Adventure/Horror, and, 50/50 R or PG-13. We have a budget of $146.5 million. Next, for theatrical I have Biographical Reenactments on a Budget.
  This is an average duration of 97 minutes. The genre is 41% biography and 35% adventure. We have majority PG.
  And also, we have a budget of $8.6 million. Then, finally, we have the Adventures for All. This is an average duration of a whopping 127 six minutes. 73% are Adventure.
  And then we have 71% as PG-13 and a budget of $10.3 million. Thank you so much for watching my presentation. Do you have any questions?
Comments
Raaed

data ?