专家简介：

杨学成，北京邮电大学经济管理学院讲师。主要研究领域为客户行为深度分析、客户关系管理、营销模式优化及网络行为研究。自中国人民大学商学院毕业并取得管理学博士学位后，进入北京邮件大学经济管理学院任教，讲授市场营销、电子商务、供应链管理、技术创新战略等课程，已在国内外学术期刊发表研究论文18篇，其中4篇被EI、ISTP收录。

参与国家级课题1项，主持省部级课题3项，并为多家企业/机构提供过咨询和培训服务，这些企业/机构包括：摩托罗拉（中国）、电讯盈科（香港）、中国移动（集团公司、山东、山西、内蒙古、重庆、天津）、中国电信（互联网增值业务部）、原中国网通（宽带在线有限公司）、中国联通（增值业务部）、总后勤部军需装备研究所、国家发改委（宏观经济研究院）、北京市发改委、北京市贸促会、北京青鸟健身有限公司、北京恒拓远博高科技发展有限公司、北京北人羽新印刷厂、山西云中制药厂、北京国华能源投资有限公司等。

专家观点概述：

Due to deregulation, new technologies, and new competitors, the telecommunication industry becomes more competitive than ever. This competition for new customers in mature markets leads to the phenomenon known as “churn”. To survive or maintain an advantage in such a competitive market, transactional record data are used to provide opportunities to segment the customer population into different groups based on differences in their purchasing behavior. Furthermore, many telecommunications companies are turning to data mining techniques to solve such challenging issues as fraud detection, customer retention, prospect profiling  and even customer relationship management (CRM). Therefore, many models or theories were proposed to forecast the cross selling opportunities. For example, provided a model named Latent Trait to find potential customer of cross selling in financial industry. adopted Acquisition Patterns to analyze the cross selling opportunities in financial services sector. The other models or patterns, such as Next-Product-To-Buy (NPTB), Survival Analysis and RFM (Recently, Frequency, Monetary) etc., also can be found in the previous literatures.

Among the existing CRM tools, cross selling has recently gained increased attention for forging stronger relationship with customers. In telecommunications industry, successful cross selling allows the vendors to learn more about the user‟s preferences and buying behavior, thereby increasing its ability to deliver “the right product/service to the right customer at the right time.” However, a practical and effective implementation of cross selling is not easy to accomplish. What makes this particularly difficult is that the current standard model which are mostly based on clustering procedures are not sound enough to effectively predict cross selling opportunities. For example, many forecasting method mainly focus on part sides of customers as background characteristic and service selection, which makes vendors difficult to find „right customers‟ for product or service. Then more efficient and accurate forecasting model is required in this area.

In this research, two data mining tools, decision tree and association rule was combined to forecast cross selling opportunity of a new service, WAP (Wireless Application Protocol). The analysis result of the actual transactional record data shows that this new method can greatly improve the accuracy rate of forecasting and help vendors find a successful cross selling policy.

The remainder of this paper is organized as follows. In Section 2, we provide a framework describing the research method. Then, Section 3 uses our approach to deal with the actual transactional record. Specifically, we discuss the data, cross selling forecast procedures, the results and the combination of the two tools. Next, we discuss the managerial implications of our findings. Finally, we conclude the paper with directions for future research, as well as limitations.