Comparative Analysis of BERTopic Versus LDA: Topic Modelling in Marketing Research

Topic modelling is the algorithm of future to expand the horizons of domain knowledge in marketing due to two reasons: (1) its ability to derive marketing insights from burgeoning wave of text data, and (2) its thoroughness in conducting literature reviews to extract latent meaning out of extant research in marketing. This study carries out a comparative assessment of two cutting-edge unsupervised topic modelling algorithms: BERTopic based on bidirectional encoder representations from transformers (BERT), and latent Dirichlet allocation (LDA). The sample for this study includes text data generated from 200 research papers published in Web of Science (WoS) indexed journals during the 5-year period of 2019 to 2023. This labelled curated sample comprises of 50 published papers belonging to each one of the elements of marketing mix – product, price, place and promotion. The topic modelling outputs are evaluated based on a comparison of topical solutions obtained using BERTopic and LDA. The study presents categorical evidence pertaining to superiority of BERTopic over LDA as an unsupervised topic modelling algorithm.