Consumer Feedback Analysis Using LDA Approach in Cross-border E-commerce

Abstract

Purpose - The objective of this research is to gain insight into consumer feedback regarding cross-border e-commerce, which is a pivotal avenue for companies venturing into global markets. It seeks to investigate the variety of themes present in consumer feedback in order to gain a more profound understanding of their preferences and concerns, ultimately influencing future market strategies. Design/Methodology – The present study utilizes an extensive collection of textual online evaluations and a text-mining LDA framework to analyze and classify potential themes in consumer feedback pertaining to cross-border e-commerce platforms. By quantifying the sentiment tendencies in consumer-related comments, sentiment analysis models that rely on emotion dictionaries as a foundation identify the polarity and intensity of emotions. Findings – The present study examined 1,500 consumer evaluations pertaining to Korean cosmetic products, sourced from a cross-border platform in Southeast Asia. The analysis unveiled five key subjects frequently referenced in consumer feedback. A subsequent sentiment analysis revealed that a significant proportion of consumer sentiments were positive. Nevertheless, consumers identified certain concerns in the minority of negative reviews, including packaging and delivery, product quality, and the absence of freebies. Originality/value - In a practical sense, this provides significant ramifications for the optimization of consumer experiences and the development of successful market strategies in the ever-changing global marketplace. This is especially advantageous for organizations striving for success in the fiercely competitive cross-border e-commerce industry. This study exceeds conventional survey methods in its methodology by offering more comprehensive and datacentric understandings of consumer attitudes and behaviors.