Analyzing Impact of Emotional Words in Product Descriptions: A Lexicon-Based Method

Abstract

This study examines the impact of emotionally charged words in product descriptions on sales performance in e-commerce. We utilize a quantitative research design with secondary data drawn from the Online Retail dataset, comprising 521,963 validated transactions from a UK-based online retailer. Using a lexicon-based approach grounded in the NRC Emotion Lexicon, this paper classifies product descriptions into emotional and neutral categories based on the presence of positive emotional keywords. Sales performance is measured at the product level in terms of total quantity sold. The empirical analysis applies the non-parametric Mann–Whitney U test to account for the skewed distribution of sales data. The results show that products with emotional descriptions achieved an average sales volume of 1,664.13 units, compared to 1,380.09 units for neutral products, representing an increase of approximately 20.6%. This difference is statistically significant (p < 0.001), with a small but meaningful effect size (r = 0.114), indicating that emotional wording exerts a measurable influence on consumer purchasing behavior. These findings provide empirical support for affective priming theory by demonstrating that emotional cues embedded in product descriptions create added affective value that translates into higher sales outcomes. This study contributes to the literature by focusing on seller-generated product descriptions rather than post-purchase reviews and by employing an interpretable lexicon-based method. From a practical perspective, the results suggest that e-commerce sellers can enhance sales performance through low-cost textual interventions, such as incorporating relevant emotional keywords into product descriptions.