Study Of Customer Segmentation Based On Clustering Analysis

Along with the rise of the internet industry, there is a growing competition among different companies.And the essence of this competition is the competition of customer resources, which means that those companies that can effectively develop and maintain customers tend to be more competitive.However, a company just has limited resources so that it is difficult to meet the needs from all customers. In order to get the most benefits, companies should assign resources as much as possible to those customer groups which contribute most.Therefore, companies need classify their customers into different groups, and identify the characteristics and value of each group.according to the demand and value of different groups,companies can effectively improve their competiti veness and get more profits.On the basis of related theories and methods about Customer Classification, this paper selects the customers’ consumer data from Company A, and then uses three clustering methods(K-means, Two-step and Kohonen) to conduct the classification. Moreover, several test methods are used to evaluate these results, and the evaluation shows that the the Two-step method perfoms better than the other two. Afterward, this paper summarizes the characteristics and value of each group based on the classification result of Two-step, and proposes several different suggestions which help to make the operation strategy of the company.