A Social Media Based Profiling Approach for Potential Churning Customers: An Example for Telecom Industry

Abstract

Customer churn prevention has been one of the most pressing concerns for telcos to tackle in the rapidly changing telecom business over the past decade, as the industry has faced increased market saturation and intense competition. Social media electronic word-of-mouth (e-WoM) offers telcos with insights into the consumer experience. Capturing preferences and views about products/services via text messaging might increase the recovery mechanism for clients who may churn. This study collects postings from social media forums concerning Taiwan’s top five telecom firms in order to perform an opinion analysis that is thought to be strongly linked with prospective churning consumers. To collect topic-sentiment information, we employ the doc-based Heuristics-N-Phrase-Rules technique. In this study, the sentiment score of negative terms is determined, and correspondence analysis is used to highlight the features and profiles of probable churn consumers. To validate its efficacy, we propose comparing the derived profile information to the telecom company’s lost consumers.