An Algorithm-based approach for Mapping customer journeys by identifying customer browsing behaviors on E-commerce Clickstream data

We propose a novel method to analyze clickstream data that can assist E-commerce retailers with marketing precision based on a better understanding of customer intentions. Design/Methodology/approach - Multiple approaches have been used with real word data from an electronic online retailer's clickstream data. This database contains 1,921,451 data from July 2019 to March 2020 from an electronic e-commerce retailer in Taiwan. In our approach, we cluster clickstream data with the K-means model and then apply the Decision tree model to identify a pattern in each cluster. On Mapping these clusters, we further analyzed that each cluster could map to the pre-purchase and purchase of customer journey model. Findings - We find a novel approach of using each session of browsing behavior as one unit to prevent a situation in which a customer may have different intentions during different periods. Furthermore, this study mainly focuses on browsing time in each category to analyze customers' preferences. Nevertheless, the results show it is a feasible approach, and one can explore customers' intentions from a new perspective. This study provides a new customer browsing intention's analyzing mode of integrating theory and practice by clickstream data, which can be applied to business. Research Limitations/Implications - Our research is limited to the browsing time in each category. One way to improve precision and accuracy would be to add other attributes to the model, such as browsing paths. Another is clickstream data cannot map retention and advocacy of customer journey. Originality/Value - We overcame the limitation of previous research, which focused on the customer journey, by taking one customer's all consumer behavior as one unit. In our study, we use each session of browsing behavior as one unit to prevent a situation in which a customer may have different intentions during different periods.