Consumer sentiments toward brands: the interaction effect between brand personality and sentiments on electronic word of mouth

Consumers can exchange real-time brand opinions through social media, hence requiring brands’ urgent attention. Therefore, can customer sentiments expressed in social media influence electronic word of mouth (e-WOM)? This research identifies customer sentiments’ influence on e-WOM and their differences based on brand personality. Brands have significantly different personalities; thus, we argue that sentiments (e.g., anger, anticipation, disgust, fear) that increase or decrease e-WOM also differ based on brand personality (sincerity, excitement, competence, sophistication, and ruggedness). Data from the top 100 most valuable global brands and more than 50,000 brand tweets were collected. We conducted a sentiment analysis on these tweets to identify customer sentiments. Then, computed regression analyses were performed, and a random forest model was constructed. This study consistently shows that the interaction terms between brand personality and sentiments are crucial to customer sentiments that increase or decrease e-WOM based on brand personality. For example, joy increases and decreases e-WOM of brands with an exciting personality and a competent and sophisticated personality, respectively. Furthermore, we conducted a cluster analysis to identify main brand clusters based on their personality and shed light on brand structure in the marketplace. We also discussed managerial implications and avenues for future research.