Anotace:
This study observes Small and Medium Sized Enterprises (SMEs) digital internationalization and determines of online comments influences customer purchase behavior, whether different product types (utilitarian and hedonic products) would moderate the relationships among online comments and customer purchase behavior. Objective data collected by Amazon platforms and SMEs transactions of 1,366 customer orders from 2019 to 2022. We collect online ratings and open-ended comments, adopting Python package of Natural Language Toolkit (NLTK) for Natural language processing (NLP) to define a score of each comment based on a database of natural human language. Regression results reveal that customer purchasing behavior was influenced by both title and content of others’ comments, the negative content has a greater negative effect than the negative title. Product types play a moderate role. Specifically, the negative contents of utilitarian goods have a stronger negative effect on customer purchase behavior. For theoretical contributions, we found that the negative comments of utilitarian products are viewed as the most influencing cues for customers purchase behavior. Furthermore, the neutral score of both title and content has a tendency to express negative emotion and attitude. The neutral comments and ratings are valuable for exploring more in future works.