Leakage Detection in telecommunications companies: a Sky Italia case study.

Abstract

Telecommunications companies have traditionally collected vast amounts of data through call records, text messages, audience measurement services and customer habits. Nowadays, these companies not only store data resulting from human interaction, but also, and most importantly, from customer clicks in an online environment. Faced with this massive amount of information, the challenge is to identify the valuable informations that are profitable for the company. This work attempts to develop a model capable of improving the digital experience of customers based on the data provided by Sky Italia. The aim is to develop a data-driven method, based on different classification models, for predicting "leakage detection", i.e. the switch from digital to human channels. Leakage risk is predicted from the digital customer journey, which is a collection of contract data and contact behavior. Predicting calls before they occur will divert customers to messaging. The value of the classification will therefore have business relevance in terms of operational efficiency, as a higher productivity of chat vs. call, and a customer experience with faster issue resolution.