A Data-driven Methodology for Agent Based Exploration of Customer Retention

Abstract

© 2016 IEEE. Customer retention is of crucial importance for companies in the mobile services industry (MSI) because of the increasing competition in the mobile services sector. As a result, companies are increasingly exploiting data to address customer retention. Traditional data collection methods such as questionnaires and interviews have been utilised to address customer retention. In addition, data mining techniques such as classification and clustering have been applied to understand customer retention. The effectiveness of utilising these techniques to investigating customer retention is debatable as a result of the complexity of the mobile services market. This study proposes a novel data-driven methodological approach to investigating customer retention in the MSI, using agent based modelling and simulation (ABMS). The dataset for this study is extracted from Twitter using specific keywords to gather data from mobile services companies of interest. The design science paradigm is adopted as the overarching methodology along with a selected framework, the social media domain analysis framework (SoMeDoA) to generate customer satisfaction determinants and agent models. ABMS is performed with the established determinants to uncover the social effects of customer retention. Data analysis show that customers' environment is a key influence on their decision to remain with (or leave) their mobile services provider, with word of mouth as an important factor. Importantly, ABMS is able to explore further determinants in a wider market place.