An Analysis on the Customer Logistic Satisfaction based on Word Clustering

Abstract

With the development of e-commerce, the importance of last mile delivery has increased and is being actively studied. These studies are mainly concerned with delivery planning, and a few focuses on customer satisfaction. If the relevance of customer characteristics and needs can be understood from the data, better services can be provided. Therefore, in this study, we analyzed the data on posted complaints regarding delivery. The proposed analysis method extracts segments of dissatisfaction by applying nonlinear clustering DBSCAN to the semantic space obtained by Word2vec. The relationship between dissatisfaction and lifestyle was analyzed by applying a correspondence analysis between dissatisfaction and customer attributes. For example, the data analysis confirmed that there is a difference in the tendency toward dissatisfaction and needs between married female, such as housewives and part-time workers, and men, such as technical company employees. Based on these results, it can be expected that the proposed method will function effectively as a guideline on how to respond and what services should be provided considering customer attributes when further enhancing customer satisfaction.