CASE STUDY - PREDICTIVE ANALYTICS

Raman passed out of campus and landed a plum role as a Predictive Analytics Consultant in a leading Data Analytics Company based at Bengaluru. His friends were envious as not many landed in the most sought-after roles. Raman had a great induction and was keen to get started.

His immediate boss Ashok assigned him to a movie optimization project. The project brief is given below.

Scheduling / Programming of movies is an important source of driving traffic and occupancy in the Cinema Business. The scheduling changes and decisions are taken every Tuesday basis the expectations from new release / performance of running movies (performance measured by occupancy), compulsions in managing Distributor expectations and relationship and intuition developed through their experience. There’s scope for Optimization and Analytics in this, which could lead to better scheduling calendar, higher occupancy and higher revenues.

The business is largely managed by their Programming Head and there’s a need for Knowledge transfer and automation in this. There’s scope for Optimization and Analytics in this, which could lead to better scheduling calendar thereby resulting in higher occupancy and higher revenues. In the Cinema industry, movies are changed every Friday but the decisions are taken every Tuesday. Multiplexes have many screens and have varying capacity. Usually there are 4 shows on weekdays and 5 shows on weekends. The task is to arrive at a predictive modelling tool that forecasts the footfall for each movie by show for a week. Ashok mentioned that they have past movie performance of more than 600 numbers, that were released over 2 years. Data like day wise, show wise, screen wise, movie wise, admits are available.

The client has 40 multiplexes and operates 180 screens in all. Their Tuesday’s are a nightmare as they need to analyze the data of last week and take a punt using intuition, experience and results.

Ashok also explained some of the terminologies which are given below. Some of the terminologies are:

a) Shows: 4 on weekday and 5 on weekends (morning show only on weekends)
b) Revenue: Revenue earned from sale of Ticket
c) F&B Sales: Food and Beverage Sales earned through money spent by customers during interval.
d) Occupancy: % of Tickets sold to the Capacity of the Auditorium.
e) Running Movie: Movie released and being screened
f) New Movie: Movie Expected to release.

Ashok clarified that the client wanted a tool, where they would input past week(s) admits data on a Tuesday for a running movie, show wise and it should forecast the admits for the week starting Friday.

Raman was curious and he asked Ashok, “How can I predict the forecasted attendance for a new movie?” It is anybody’s guess? Ashok laughed and said, welcome to the real world of corporate? What you studied won’t be useful beyond a point. You need to apply and keep trying and discover yourself. There are no teachers and there are only mentors. I can at best ideate with you, when you bring an approach document. But you need to bring an approach document, that optimizes the footfall and increases the revenue for the client.

Raman was doing some mental maths. Every Tuesday, the decisions taken are any or all of the following. a) Which running movie should I maintain status quo in the number of shows and screen size b) Which running movie should I reduce the screen size c) Which running movie should I reduce the number of shows d) Which running movie should I reduce the number of shows and screen size Raman was confident that he can take a crack on predicting the footfall for a running movie, but was clueless on how to predict the footfall for a new movie. He was hoping to find some models in the net. If you are Raman, how would you approach this project of building a predictive analytics tool to forecast to achieve occupancy optimization balancing the running movies and new movies. The client will only be happy if the predictive forecast is better than their own decisions. The clients operate at 60% occupancy and they wish to improve this.

Send an email to subbu@drawing-board.in to submit your solution to this challenge