Assessing brand equity through customer reviews: a naïve Bayes classifier approach

Customer-based brand equity has been critical to gaining a competitive advantage in the marketplace. Despite the usefulness of conventional survey-based measurement, many scholars and practitioners face difficulties in comparing brand equity across brands, sectors, and products owing to the limitations of survey-based measurement. This study proposes a machine learning method to complement traditional survey-based measures. Our approach comprises four major stages: 1) extracting seed phrases from brand equity survey questionnaires; 2) identifying response sentences from customer reviews; 3) assigning scores to brand equity; 4) replicating regression models from previous empirical studies to validate our approach. Our analysis examined 65,057 vacuum cleaner customer reviews from an online e-commerce platform, (e.g., Amazon.com) representing six major brands. This paper complements traditional studies by presenting a consistent and reliable measurement methodology.