Mining the Text of Online Reviews to Explore Brand Positioning: Emotional and Psychological Brand Associations

Abstract

Purpose – So far, text mining techniques used in literature to study the content of online reviews are quite heterogeneous, so it might lead to confusion for scholars and companies willing to reproduce the analyses. This paper aims to propose and illustrate a unified and structured research procedure to uncover emotional and psychological brand associations from online reviews, with the goal of analyzing brand positioning and identifying brand segments. Design/methodology/approach – A set of 62,496 online reviews belonging to 44 brands of the product category of “blush” was collected from a cosmetic’s online retailer. The lexicon-based tool Linguistic Inquiry and Word Count (LIWC) was used to conduct the text mining analysis. A total of 26 textual variables were selected as inputs for brand positioning and brand segmentation analyses. Findings –We illustrate how textual data can be used to uncover emotional and psychological brand associations. In the cosmetics category studied, words related to positive emotions are the most common. Moreover, words representing perceptual processes, such as see and feel, words associated to body and those reflecting time and space issues are also quite related to the product category. Based on brand associations, this study found four brand clusters in the category. Practical implications – This study provides marketing practitioners, especially from small and medium companies, with a an easy-to-implement proceure to uncover the textual content of online reviews to study brand positioning and segmentation. Originality/value – This research is pioneering in providing a unified and structured guide to uncover emotional and psychological brand associations from the textual content of online reviews to study brand positioning and to identify brand segments.