In the increasingly competitive environment between e-commerce companies, for more accurate implementation of marketing strategies, e-commerce websites often choose to subdivide the consumer market of the enterprise to identify site users’ characteristics to find their needs. In this paper, we subdivide consumer market from the four dimensions of behavior, geography, demography and psychology and propose a model to describe the characteristics of potential user market segments. Based on the web log data and user transaction data, a classification algorithm is used to analyze user behavior data in Web log to find the potential user segments and the user's descriptive characteristics in user transaction data are clustered to obtain the distribution of consumer characteristics under various product categories, then we use the product categories in e-commerce website as an intermediary to give every single potential user in potential user market segments the descriptive characteristics, which can provide data support for the realization of precision marketing. The proposed model is applied to the actual data of a certain insurance e-commerce platform, and based on the results, we gain some implications for marketing of the e-commerce website.