Online newspaper subscriptions: using machine learning to reduce and understand customer churn

Abstract

ABSTRACT Modelling customer loyalty has been a central issue in customer relationship management, particularly in digital subscription business models. To guarantee news media sustainability, publishers implemented subscription models that need to define successful retention strategies. Thus, churn management has become pivotal in the media subscription business. The present study aims to understand what drives subscribers to churn by performing a Machine Learning approach to model the propensity to churn of online subscribers of a Portuguese newspaper. Two models were developed, tested, and evaluated in two timeframes. The first one considered all Business to Consumer (B2C) subscriptions, and the second only the B2C non-recurring subscriptions. The experimental results revealed important patterns of churners, which allowed the marketing and editorial teams to implement churn prevention and retention measures.