Effect of Different Chatbot Features on Customer Satisfaction in E-commerce

Abstract

By incorporating additional questions related to chatbots, we construct a questionnaire inspired by the work of previous researchers. This questionnaire aims to assess the relationship between various chatbot attributes, including usability, language style, and response time, and customer satisfaction within the e-commerce sector. To mitigate issues of multicollinearity and optimize the number of variables within a multiple linear regression model, we employ LASSO regression. Upon conducting hypothesis testing, our findings manifest a positive correlation between the overall chatbot features and customer satisfaction index. Specifically, an increase in the levels of language style and usability is consistent with increased levels of customer satisfaction, with language style exerting a larger effect.