I 摘要 我国电信业“三足鼎立”格局形成后，运营商之间一直进行着激烈的市场竞 争，客户资源的争夺是各通信运营商关切的头等大事。 随着通信市场的高度饱和，新增客户数量和质量大不如前，客户流失却不断 加剧，造成企业运营收入增长放缓甚至下降。因此提升现有客户保有率，预测并 挽留潜在流失客户日益成为电信企业各级单位工作的重中之重。但是大多客户的 流失尤其是中高端客户的流失前期表现的形式就是成为零次用户，所以如果能够 准确的预测出潜在的零次用户，就能对客户展开有针对性的营销，进而完成对客 户有效的保有工作。 本文根据客户生命周期、客户价值、客户维系理论，运用数据挖掘技术对潜 在流失客户进行预测分析，以移动S分公司的客户流失数据为研究对象，利用该 公司的数据挖掘工具，并运用决策树C5.0算法，建立了潜在流失客户预测模型。 在建模过程中，充分考虑了客户的在网时长、品牌资费、归属渠道和县域分布等 维度，对零次户进行了细化分类，生成多个零次户的结果集合，并以此预测出潜 在流失客户的清单。最后，根据流失预测模型结果优化了移动S分公司潜在流失 客户挽留体系，并针对客户全生命周期各阶段提出了客户挽留策略，重点阐述了 提高服务能力以提高客户忠诚度和提高营销能力以增加客户粘性，各部分有针对 性地对流失客户采取了一系列的客户挽留措施，取得了很好的效果。本文的研究 结果，对移动S分公司解决客户流失问题具有现实的使用价值，并可以指导其他 企业的客户流失预测和挽留工作。 关键词：潜在流失客户，预测模型，客户挽留 Abstract III Abstract After the formation of the situation of tripartite confrontation in China's telecommunications industry, operators have been engaged in fierce market competition. The competition for customer resources is a top priority for all communication operators. With the high saturation of the communication market, the quantity and quality of new customers are not as good as before, and the loss of customers has intensified continuously, which causing the growth of business operating income to slow down or even decline. Therefore, improving the existing customer retention rate, predicting and retaining potential customers are increasingly becoming the top priority of all levels of telecommunications companies. However, the loss of most customers, especially the high-end customers, is shown as the form of becoming zero-usage users. Therefore, if the operators can accurately predict the potential zero- usage users, they can conduct targeted marketing and complete the effective retention work for customers. Based on customer life cycle, customer value and customer maintenance theory, this paper uses data mining technology to conduct prediction analysis on potential lost customers, and takes the customer churn data of Mobile S Branch as the research object by using the company's data mining tools and Decision Tree C5.0. The algorithm establishes a potential customer churn prediction model. In the modeling process, the customer's active network time, brand tariff, attribution channel and county distribution are fully considered, and the zero-usage users are refined and classified, and the result sets of multiple the zero-usage users are generated and predicted a list of customer churns. Finally, based on the results of the churn prediction model, a potential customer churn retention system is constructed, and the customer retention strategy is proposed according to each stage of the customer’s life cycle. The focus is on improving service capability to enhance customer loyalty and on advancing marketing ability to increase customer viscosity. We have targeted a series of customer retention measures for lost customers and achieved good results. The research results of this paper have practical use value for the Mobile S Branch to solve the customer churn problem, and enable to guide the customer churn prediction and retention work of other companies. Key words: potential customer churn, prediction model, customer retention 目录 1 目 录 摘要 ................................................................................................................ I Abstract ........................................................................................................... III 第1章 绪论 ..................................................................................................... 3 1.1 研究背景与意义 ............................................................................ 3 1.1.1 研究背景 ............................................................................... 3 1.1.2 研究意义 ............................................................................... 5 1.2 研究文献综述 ................................................................................ 6 1.2.1 关于客户流失预测的研究 .................................................... 6 1.2.2 关于客户挽留策略的研究 ................................................... 11 1.3 研究内容及研究方法 .................................................................. 13 1.3.1 研究内容 ............................................................................. 13 1.3.2 研究方法和技术路线 .......................................................... 14 1.4 创新之处 ..................................................................................... 15 第2章 理论基础 ......................................................................................... 17 2.1 客户关系管理理论 ...................................................................... 17 2.1.1 客户关系管理的概念 .......................................................... 17 2.1.2 客户关系管理的主要内容 .................................................. 18 2.1.3 客户关系管理的作用 .......................................................... 18 2.2 客户生命周期理论 ...................................................................... 19 2.2.1 客户生命周期理论体系 ...................................................... 19 2.2.2 电信业客户生命周期 .......................................................... 20 2.3 客户价值理论 .............................................................................. 22 2.3.1 客户价值的概念 .................................................................. 22 2.3.2 客户价值评价体系 .............................................................. 23 2.4 客户维系理论 .............................................................................. 24 2.4.1 客户维系理论的概念 .......................................................... 24 2.4.2 客户维系及其意义 .............................................................. 25 2.4.3 客户维系的必要性 .............................................................. 25 第3章 移动S分公司客户流失问题分析 ................................................... 27 3.1 移动S分公司基本情况及发展现状 ........................................... 27 3.2 移动S分公司SWOT分析 ......................................................... 28 3.3 移动S分公司客户流失问题现状 ............................................... 30 目录 2 3.3.1 移动S分公司客户流失情况 .............................................. 30 3.3.2 移动S分公司客户流失成本评估....................................... 33 3.3.3 移动S分公司流失客户结构特征和流失原因分析 ............ 34 3.4 移动S分公司客户流失管理存在问题 ....................................... 38 第4章 移动S分公司潜在流失客户预测模型的构建 ................................. 41 4.1 移动S分公司客户流失预测技术介绍 ....................................... 41 4.2 移动S分公司潜在流失客户预测模型数据来源 ........................ 44 4.3 移动S分公司潜在流失客户预测模型数据分析 ........................ 46 4.3.1 潜在流失客户决策因子判定表 ........................................... 46 4.3.2 数据描述性分析 .................................................................. 48 4.3.3 异常值处理 ......................................................................... 52 4.4 移动S分公司潜在流失客户预测模型的验证 ........................... 52 4.5 移动S分公司潜在流失客户预测模型的应用 ........................... 54 第5章 移动S分公司潜在流失客户挽留策略 ............................................. 57 5.1 移动S分公司潜在流失客户挽留的总体思路 ........................... 57 5.2 移动S分公司潜在流失客户挽留体系优化策略 ..