Predicting user motivation towards retention of e-Services: an NLP-based approach

Abstract

In this modern era, the dynamic business world has led to the emergence of ‘market orientation’ and ‘social CRM’ (Kohli & Jaworski, 1990; Narver & Slater, 1990; Diffley et al., 2018). The benefits of e-Services are often not fully utilized because of users’ unwillingness to use it (Venkatesh & Davis, 2000; Devaraj & Kohli, 2003). Hence, understanding the user’s motivation in an e-service through Twitter data can help companies form better strategies to retain users. Though people adopt services quickly, they tend to discontinue the service after limited use. Productivity benefits and maximum customer lifetime value (CLV) are typically obtained in the continued use phase (Venkatesh et al., 2003; Kim & Malhotra, 2005).With the emergence of social media, extracting and processing information (Kosala & Blockeel, 2000, Sakaki et al., 2010, Russell, 2011; Crooks et al., 2013), can help in understanding the motivation of users (DeVaro et al., 2018) towards using an eService. This has motivated us to analyze Twitter data to understand customer motivation levels in eService retention. In this study, 1000 tweets were downloaded from ten different e-Service providers based on the company’s official twitter handle and analyzed. The results show that using Naïve Bayes on function and content words help in predicting retention intention. Predicting the IS continuance intention of users through tweets analysis can help companies perform better sentiment analysis and provide customized benefits to users. The limitation of this study is that though Twitter analytics can serve a good medium for analyzing users’ behavioural intentions, the words and its usage in different situations can change.