Case Study GECA: Audience Prediction With ML

“We are particularly proud to have been able to develop this unique and revolutionary tool for international TV. We believe that this solution is key to the TV industry and we are confident that it will enable broadcasters to move into new frontiers and achieve optimal results.” Enrique Lozano, GECA Director

We are at a time of great changes and transformations affecting economic activity and society. One of the sectors most affected by this revolution is the audiovisual sector, which has become one of the driving forces behind the transformation of other more traditional activities outside the cultural sphere.

GECA is one of the leading groups in the European audiovisual sector, integrating audiovisual content, production, and distribution. It provides the creativity and technical solutions needed to design, produce, and distribute audiovisual or multichannel projects.

CLIENT: Geca

SECTOR: Media

SERVICE: AI

TECHNOLOGY: Azure ML, Azure DevOps

Audience measurement is a key economic driver for the media, so we have developed an automated prediction service using Machine Learning and Azure services to improve conversion rates and know consumer behaviour in advance.

The Challenge

Technological advances, the evolution of content, free-to-air broadcasts shared with platforms and social networks, fragmented audiences, niche channels... Many variables must be considered to predict, anticipate and innovate in the changing audiovisual world.

A platform that, through Machine Learning, had to be created would study the current data and learn from the new data being received to make audience predictions over 30-day days.

Results

The audience forecasting tool uses state-of-the-art algorithms and advanced data analytics to provide highly accurate estimates of TV viewing behavior. This artificial intelligence and machine learning system can analyze various factors, such as demographics, viewing preferences, social trends, historical data, and content minutes, to deliver robust, agile, and actionable projections.

The solution has been built using the Azure cloud, following the MLOps Framework. Using the latest advances in Machine Learning, a study of current data and learning from new data is carried out. Datasets including two years of historical data, shared data divided by gender, program trawling, etc., are considered.

After a training period, new actions can be executed, allowing for improved conversion rates and knowing in advance the behavior of the consumers and the revenues obtained.

Highlights

At a time of great changes and transformations in the audiovisual sector, having a tool like this that helps to predict the audience will be key to generate and retain value, as well as a unique differentiating point from the competition, as advantages have been achieved such as: -Robust information: Content and programming teams will obtain accurate data, facilitating decision-making. - Optimisation of programming: Programming managers can propose different scenarios with their respective projections, helping to optimize resources and maximize audiences. - Improved advertising investment: Reliable audience projections will help design strategies to maximize advertising effectiveness.

The solution has become the perfect tool to cope with the speed at which the context and sector will change. We will face an even more internationally competitive environment in which the generation and retention of value will be key. Therefore, having a solution that helps us to predict the audience will be the key to differentiating ourselves from the competition.