Abstract: From the social network theory, this article depicts the characteristics of phone call network and message network through the analysis of the communication and message data, and consumption record provided by a mobile telecom company in the domestic market. Furthermore, the paper conducts principle component analysis and clustering analysis based on network statistics, classifies the consumers into three categories: the inactive consumer, the moderately active consumer and the opinion leadership, and divides customer value into direct and indirect value to investigate the impact of opinion leadership on customer indirect value. Regression analysis reveals that the role of opinion leadership is mainly reflected in the promotion of indirect value of customers and his/her consumption (i.e., direct value) has no significant difference with others.However, consumers who are connected with opinion leaderships have significantly more consumption because opinion leaderships are more active in a network. This finding suggests that measuring customer value should be based on individual spending, moreover, a customer value should include social relationship value; in order to remain customers in a long term, telecom companies should pay attention to the identification and maintainance of opinion leadership, because they can stabilize network relationship.

Key words: social network, consumer indirect value, opinion leadership;principle component analysis;clustering analysis;regression analysis