Thai Commercial Banks on Twitter: Mining Intents, Communication Strategies, and Customer Engagement

Abstract

Twitter is a social media (SM) platform that rapidly generates electronic word of mouth (e-WOM). Marketer-generated content (MGC) is controllable and could enhance the positive e-WOM. Hence, in this study, the author examined the characteristics of MGC and reactions from followers based on Thai banks’ Twitter accounts. The author collected a total of 10,000 tweets from nine banks in Thailand—both high- and low-performing banks. The author conducted research with natural language processing (NLP) to uncover intents using an open application programming interface (API). The author used three data-mining techniques—association, clustering, and classification. The Twitter strategies of banks with high and low performances are quite similar. The sentiment is the intent type that dominates Thai banks’ intent strategies. Several intents could be combined to draw e-WOM in terms of favorites (FAV) and retweets (RT). Six intent patterns (clusters) were extracted. Some of these clusters are classifiers for FAV and non-FAV tweets. This study guides the application of data mining in business research and suggests MGC strategies for marketers.