A Brand-New Look at You: Predicting Brand Personality in Social Media Networks with Machine Learning

Abstract

Tools for analyzing social media text data to gain marketing insight have recently emerged. While a wealth of research has focused on automated human personality assessment, little research has focused on advancing methods for obtaining brand personality from social media content. Brand personality is a nuanced aspect of brands that has a consistent set of traits aside from its functional benefits. In this study, we introduce a novel, automated, and generalizable data analytics approach to extract near real-time estimates of brand personalities in social media networks. This method can be used to track attempts to change brand personality over time, measure brand personality of competitors, and assess congruence in brand personality. Applied to consumer data, firms can assess how consumers perceive brand personality and study the effects of brand–consumer congruence in personality. Our approach develops a novel hybrid machine learning algorithmic design (LDA2Vec), which bypasses often extensive manual coding tasks, thus providing an adaptable and scalable tool that can be used for a range of management studies. Our approach enhances the theoretical understanding of channeled and perceived brand personality as it is represented in social media networks and provides practitioners with the ability to foster branding strategies by using big data resources.