Predicting customer behaviour in the Prisa Group
Today, digital media are moving to a subscription and pay-for-content model that allows them to maintain recurring revenues. For this, it is essential to have a profiling of the users of your websites to predict customer behaviour.
The Data area of Prisa Noticias decided to launch in the middle of the year, a project created by Google to fund strategic media projects. This project is part of the programme "Partner Program. The objective is user profiling and the application of predictive models (subscription, abandonment, etc.).
BOSONIT was the specialised company in charge of the user profiling part of the project.
First of all, the main purpose of the project has been the definition of objectives and improvements of the system to be implemented taking into account the customer's needs (what data they wanted to collect from navigation).
Secondly, the objective has been to design and subsequent deployment of the architecture components within the Google Cloud Platform ecosystem. The components had to enable the collection, processing and storage of navigation events in real time.
These are produced by its users, both on the El País website and on its mobile app.
The aim of this process has been to creation of a customer file in real time to serve as a starting point for the development of predictive modelling, using tools from machine learning included in the ecosystem itself. In other words, it will make it possible to identify which content the user has visited, with the aim of associating each one with the advertising campaign that best suits their preferences. This will increase the likelihood that the advertising content will actually be consumed by the user.
In addition, a series of KPIs have been created to create dashboards using Business Intelligence.