Abstract: Big data provides important support for enterprises doing precision marketing decision. It can significantly enhance marketing effectiveness and improve customer satisfaction. Customer identification and selection is fundamental for precision marketing. It helps companies target valuable customers and design marketing mix strategies. As a social structure analysis method, Social Network Analysis（SNA） method can effectively target the key customers by analyzing customer’s individual characteristics and group features. Firstly, through the social network centrality analysis, this paper studies the relationship between the node centrality in a network and the marketing effectiveness. Furthermore, by introducing a method called RMCL（Regularized Markov Clustering）, a solution is presented to partition social network into communities. More importantly, in order to detect communities in large-scale social networks, the RMCL method is implemented in Map Reduce Pattern, a parallel programming framework. This paper also explores the approaches to identify the core node within a social network community. Two indicators are proposed in term of customer influence degree and influence factor to measure the customer’s impact on purchasing. Lastly, combining with social network centrality metrics, CART method is used  to uncover the links between the social network and customer consuming response, and the order of variable importance is determined, which provides a feasible guidance for the enterprises adopting differentiated marketing strategies.

Key words: social network analysis, precision marketing, customer positioning, big data

摘要： 大数据为企业进行精准营销提供了重要支撑，精准营销能提升营销效果，提高客户满意度，精准营销的前提是客户识别与选择。通过分析网络个体与群体特征，社交网络分析能够定位核心价值客户。首先对社交网络的中心性进行分析，探讨社交网络节点地位与营销效果的关系，运用社群识别方法，对社交网络进行分群，提出并用MapReduce实现了针对大规模社交网络的社群划分RMCL方法。在此基础上，构建了客户影响度与客户影响因子等指标，并结合中心度指标，定位社群的核心节点，并采用分类回归树方法，研究了社交网络结构与客户消费响应关系，并确定了变量重要性，为企业采取客户差异化营销组合策略提供指导。

关键词: 社交网络分析, 精准营销, 客户定位, 大数据