Research

“How brand perception (really) changes over time: Decomposing real from nominal changes using time-aware natural language processing” Vincent Chen, Ming Hsu, and Zhihao Zhang (Job Market Paper)

Consumer perceptions of brands are rarely static. Consequently, successful branding demands timely and accurate tracking of the evolving dynamics of how consumers perceive brands. While measuring brand perception has long been a standard practice for marketers, a less recognized challenge of brand tracking is that the language describing attributes of brands—such as how Social a brand is—can itself change over time. In this paper, we propose the need to introduce a conceptual distinction between real and nominal measures of brand perception changes, as the latter can deviate substantially from the former due to changes in attribute meanings over time. Specifically, we develop a data-driven approach to dissociate real from nominal perception by accounting for changes in attribute meanings. We do so by leveraging time-aware natural language processing (NLP) models and brand tracking data covering 20 years between 2001 and 2020 to extract attribute meanings over time from contemporaneous longitudinal text corpora. We further present evidence supporting the validity of our decomposition using a combination of behavioral and NLP approaches. Taken together, this work expands the ability of researchers and practitioners to capture changes in brand perception over time and provides a theoretical foundation for understanding their managerial implications.

“Windfall, similarity, and mental accounting” with Ellen Evers - Under review at Journal of Consumer Research

Although money is fungible, consumers often spend it differently depending on its source. One well-known example is the windfall effect, where consumers appear more willing to spend windfall (vs. hard-earned) money. Three competing theories have been proposed to explain this phenomenon. Two popular theories suggest consumers have a higher marginal propensity to consume (HMPC) when spending windfall money, either in general or specifically for hedonic products. A third theory, mental accounting (MA), suggests that people mentally categorize windfall and hard-earned money differently. However, past research has not specified exactly how these mental categories are formed, resulting in unfalsifiable explanations. We address this issue by proposing a similarity-based model of mental accounting, which predicts that consumers are more likely to spend money on products similar to the source of the money, whether the money is a windfall or not. Across four controlled experiments and one field study, we find consistent evidence supporting a similarity-based process. In contrast, we find no evidence that simply receiving a windfall increases spending, challenging the widely accepted HMPC explanation. We further show that prior evidence for HMPC can be explained by selective stimulus sampling: when a broader range of products is tested, the windfall effect disappears.