Marketing insights from sentiment distribution of high-tech products using DistilBERT

Abstract

PurposeThis study proposes a new method for detecting sentiment distribution and identifying key events in high-tech product categories, providing insights from decision-making in marketing.Design/methodology/approachUsing DistilBERT, the study analyzes the sentiment of wireless earphone reviews, categorizing them and extracting labels and scores. It visualizes trends, identifies sentiment inflection points from sentiment distribution, infers events via topic analysis and compares product utility and customer sentiment.FindingsOur approach detects key events and discloses sentiment fluctuations and product interest shifts. Innovations amplify positive responses, underscoring the dynamic consumer sentiment. This study adeptly pinpoints key events after introducing novel technologies and reveals disparities in product utility and customer sentiment pre- and post-notable sentiment changes.Originality/valueThis study presents a unique method that uses an event-based approach to analyze sentiment and topic shifts, providing new insights into sentiment trends and emerging topics. This new understanding of consumer preferences and trends over time can be especially valuable for better strategic decision-making in marketing.