运用商业智能构建分析型CRM
摘要
正如我刚踏入大学校园时对“电子商务（E-business）”充满好奇一样，在研究生项目管理课程中第一次听到“企业资源计划（Enterprise Resource Planning, ERP）”一词时也感到耳目一新，而现在却发现几乎所有的人都在大谈特谈网络经济时代和电子商务，周围的不少朋友也可能正在为本公司的ERP系统奔走忙碌，相关的书籍和文章更是如雨后春笋般层出不穷，更多的概念像潮水一样向我们涌来——譬如客户关系管理（CRM）。
其实对企业而言，客户关系管理并不是什么新名词，本文开头举出的街边小店的例子就很好的说明了这一点。然而要强调的是，客户关系管理并不等同于CRM。尽管从字面上没有差别，但前者更倾向于一种管理理念，而后者则更像一整套如何对客户关系进行管理的解决方案。这正是由于企业对客户关系管理一向以来的重视，并对这个概念不断丰富和扩充的结果。客户关系管理涵盖了包括发掘高价值客户、提高客户利润贡献率、降低客户流失率等一切与客户有关的活动，其目的都是为了保证“以客户为中心”战略的实施。
“管理为本、技术为用”，CRM正是在这种需求下应运而生，信息技术的发展为在企业内构建CRM应用提供了支持和保障，许多企业也纷纷建立起自己的CRM系统，然而对客户关系管理的质量却没有得到实质性的提高。一方面，CRM的功能仍然集中在对营销、销售和客户服务三部分业务流程的信息化、一体化，以及与客户进行沟通所需要的手段（如电话、传真、网络、Email等）的集成化、自动化之上；另一方面，CRM系统自身和企业的其它应用系统、外部数据源却产生了大量的未经处理的数据，这些数据的价值有90%没有开发出来。
针对目前主流CRM系统偏重流程的特点，业界提出了分析型CRM的概念，即具有对数据进行分析和提炼，产生客户智能，并支持企业战略决策功能的CRM系统。实际上，数据仓库、OLAP和数据挖掘的概念在较早的时候就已经提出来了，并且运用于商业活动的某些领域，然而，将这三者作为一个整体并冠以商业智能（Business Intelligence, BI）的名称却仅仅是两三年前的事（详见文中有关IBM商业智能体系结构的介绍），而如何用商业智能来构建分析型CRM系统却正是目前业界正在探讨的问题之一。
本文首先从客户生涯价值的角度，以Fast Lube公司进行客户保持策略前后客户生涯价值对比分析的案例论述了客户关系管理对企业的重要性。紧接着，文章对CRM的体系结构和当前CRM系统的应用情况进行了探讨，指出了当前主流的流程型CRM的主要不足之处，即无法真正明白客户需求、无法进行有针对性的商业措施，又从CRM系统投资回报率（ROI）的角度详细论述了采用分析型CRM在提高客户利润贡献率等各方面的优势，从而得出了分析型CRM将是CRM发展趋势的结论。随后，本文通过大量翔实的案例重点研究了如何运用商业智能来构建分析型CRM的问题。
* 从所用的商业智能技术来分析：
首先，建立客户信息数据仓库，对异构数据源进行清洗和转换，为以后的数据分析工作打下良好的基础。其次，通过报表查询和多维分析（OLAP）来完成有关客户关系管理的验证型问题，即观察CRM系统的实际情况，以便确定某种假设是否成立；最后，通过数据挖掘来完成有关客户关系管理的发现型问题，即从大量数据中发现未知的信息，进而对将来的情形进行预测。
* 从CRM的应用领域来分析
从广义的应用来看，商业智能可以运用于CRM系统中营销、销售和客户服务这三大领域；从狭义的应用来看，商业智能（主要是数据挖掘技术）可以运用于客户细分、客户保留、客户响应、交叉营销和市场分析等各项具体业务。
本文的写作过程也是笔者学习、探讨和实践的过程。在此期间，笔者阅读了大量国内外著作、论文，对大量资料进行了归纳和整理，也试着对该论题进行细致而深入的研究。真诚感谢我的导师陈恭和教授在论文写作期间给予我无私的指导，并提出了许多非常宝贵的意见和建议，同时还要感谢所有老师对我的帮助，没有你们的支持，我不可能顺利完成论文，也不可能完成学业。通过论文的写作，笔者自身感觉受益颇多，也希望能对读者有所帮助。
It seemed new and fresh to me when I heard the word "Electronic business" for the first time during my earlier days of university campus life, also did the word ERP, enterprise resource planning on the Project Management course during my postgraduate studies. Nowadays almost everybody is talking about them, even many friends of mine are busy working with the ERP system of their company. We can find numerous related books or articles in the market or on the net. More and more new words are flooding us, such as CRM, customer relationship management.
Actually, customer relationship management is not as new as it seems to be. The case of old corner grocer at the very beginning of the article proved it very well. What I have to emphasize is customer relationship management is not equal to CRM though it seemed no difference existed between them literally. Customer relationship management trends to be a kind of management strategy. CRM is more likely to be a set of solution to the problem of customer relationship management. People paid a lot of attention to customer relation and gave new meaning to it continuously. Customer relationship management includes acquiring new customer, retain existing customer, improving customer profitability and all the other things related to customer. The goal of customer relationship management is to carry out the "customer-oriented" management strategy.
To realize the goal of customer-oriented strategy, we need information technology to support us. Many companies had built their own CRM system, but unfortunately they had no substantial improvement in customer relationship management. On one hand, the main function of CRM system is only to keep marketing, sales and customer service in conformity with each other and automate customer communication (such as telephone, fax, network, e-mail, etc), on the other hand, CRM system itself, plus other application systems inside the company and outside data source had produced numerous data with 90% of them had not been used.
Most of all kinds of CRM software on the market nowadays are operational CRM which lies on business process. But what we need most is analytical CRM which has the ability of analyzing and abstracting data from different kinds of data source to produce customer intelligence and support decision making. Actually, the concepts of date-warehousing, OLAP and data-mining are had been put forward for a long time, but putting them together and naming it BI, business intelligence was the thing happened only two or there years ago. Now what we are discussing is how to use business intelligence to build analytical CRM.
In this article, I first discussed the importance of customer relationship management from the view of customer lifetime value (LTV). In the case of Fast Lube, I compared the difference of LTV before or after the customer retention
strategy had been carried out. After that, the technology structure and current usage of CRM system was described. The article pointed out the main shortcomings of operational CRM which is it can not understand the real needs of customer and can not help us carry out pertinent business measures. Form the view of ROI, return of investment, the article discussed the advantages of analytical CRM in the aspect of improving customer profitability and drew the conclusion of analytical CRM was the direction of CRM system. At last, the article discussed the problem of how to use business intelligence to build analytical CRM through a lot of cases and recourses.
* From the technologies of business intelligence
Firstly, we should build customer information data warehouse to abstract and transform data from different recourses to make to following steps easier; secondly, we can use query & reports or OLAP to solve verification problems related to customer relationship management; finally, we can use data-mining to solve the discovery problems, that is, discover unknown information from mass data and predict the future.
* From the applied fields of CRM
Generally speaking, business int
其实对企业而言，客户关系管理并不是什么新名词，本文开头举出的街边小店的例子就很好的说明了这一点。然而要强调的是，客户关系管理并不等同于CRM。尽管从字面上没有差别，但前者更倾向于一种管理理念，而后者则更像一整套如何对客户关系进行管理的解决方案。这正是由于企业对客户关系管理一向以来的重视，并对这个概念不断丰富和扩充的结果。客户关系管理涵盖了包括发掘高价值客户、提高客户利润贡献率、降低客户流失率等一切与客户有关的活动，其目的都是为了保证“以客户为中心”战略的实施。
“管理为本、技术为用”，CRM正是在这种需求下应运而生，信息技术的发展为在企业内构建CRM应用提供了支持和保障，许多企业也纷纷建立起自己的CRM系统，然而对客户关系管理的质量却没有得到实质性的提高。一方面，CRM的功能仍然集中在对营销、销售和客户服务三部分业务流程的信息化、一体化，以及与客户进行沟通所需要的手段（如电话、传真、网络、Email等）的集成化、自动化之上；另一方面，CRM系统自身和企业的其它应用系统、外部数据源却产生了大量的未经处理的数据，这些数据的价值有90%没有开发出来。
针对目前主流CRM系统偏重流程的特点，业界提出了分析型CRM的概念，即具有对数据进行分析和提炼，产生客户智能，并支持企业战略决策功能的CRM系统。实际上，数据仓库、OLAP和数据挖掘的概念在较早的时候就已经提出来了，并且运用于商业活动的某些领域，然而，将这三者作为一个整体并冠以商业智能（Business Intelligence, BI）的名称却仅仅是两三年前的事（详见文中有关IBM商业智能体系结构的介绍），而如何用商业智能来构建分析型CRM系统却正是目前业界正在探讨的问题之一。
本文首先从客户生涯价值的角度，以Fast Lube公司进行客户保持策略前后客户生涯价值对比分析的案例论述了客户关系管理对企业的重要性。紧接着，文章对CRM的体系结构和当前CRM系统的应用情况进行了探讨，指出了当前主流的流程型CRM的主要不足之处，即无法真正明白客户需求、无法进行有针对性的商业措施，又从CRM系统投资回报率（ROI）的角度详细论述了采用分析型CRM在提高客户利润贡献率等各方面的优势，从而得出了分析型CRM将是CRM发展趋势的结论。随后，本文通过大量翔实的案例重点研究了如何运用商业智能来构建分析型CRM的问题。
* 从所用的商业智能技术来分析：
首先，建立客户信息数据仓库，对异构数据源进行清洗和转换，为以后的数据分析工作打下良好的基础。其次，通过报表查询和多维分析（OLAP）来完成有关客户关系管理的验证型问题，即观察CRM系统的实际情况，以便确定某种假设是否成立；最后，通过数据挖掘来完成有关客户关系管理的发现型问题，即从大量数据中发现未知的信息，进而对将来的情形进行预测。
* 从CRM的应用领域来分析
从广义的应用来看，商业智能可以运用于CRM系统中营销、销售和客户服务这三大领域；从狭义的应用来看，商业智能（主要是数据挖掘技术）可以运用于客户细分、客户保留、客户响应、交叉营销和市场分析等各项具体业务。
本文的写作过程也是笔者学习、探讨和实践的过程。在此期间，笔者阅读了大量国内外著作、论文，对大量资料进行了归纳和整理，也试着对该论题进行细致而深入的研究。真诚感谢我的导师陈恭和教授在论文写作期间给予我无私的指导，并提出了许多非常宝贵的意见和建议，同时还要感谢所有老师对我的帮助，没有你们的支持，我不可能顺利完成论文，也不可能完成学业。通过论文的写作，笔者自身感觉受益颇多，也希望能对读者有所帮助。
It seemed new and fresh to me when I heard the word "Electronic business" for the first time during my earlier days of university campus life, also did the word ERP, enterprise resource planning on the Project Management course during my postgraduate studies. Nowadays almost everybody is talking about them, even many friends of mine are busy working with the ERP system of their company. We can find numerous related books or articles in the market or on the net. More and more new words are flooding us, such as CRM, customer relationship management.
Actually, customer relationship management is not as new as it seems to be. The case of old corner grocer at the very beginning of the article proved it very well. What I have to emphasize is customer relationship management is not equal to CRM though it seemed no difference existed between them literally. Customer relationship management trends to be a kind of management strategy. CRM is more likely to be a set of solution to the problem of customer relationship management. People paid a lot of attention to customer relation and gave new meaning to it continuously. Customer relationship management includes acquiring new customer, retain existing customer, improving customer profitability and all the other things related to customer. The goal of customer relationship management is to carry out the "customer-oriented" management strategy.
To realize the goal of customer-oriented strategy, we need information technology to support us. Many companies had built their own CRM system, but unfortunately they had no substantial improvement in customer relationship management. On one hand, the main function of CRM system is only to keep marketing, sales and customer service in conformity with each other and automate customer communication (such as telephone, fax, network, e-mail, etc), on the other hand, CRM system itself, plus other application systems inside the company and outside data source had produced numerous data with 90% of them had not been used.
Most of all kinds of CRM software on the market nowadays are operational CRM which lies on business process. But what we need most is analytical CRM which has the ability of analyzing and abstracting data from different kinds of data source to produce customer intelligence and support decision making. Actually, the concepts of date-warehousing, OLAP and data-mining are had been put forward for a long time, but putting them together and naming it BI, business intelligence was the thing happened only two or there years ago. Now what we are discussing is how to use business intelligence to build analytical CRM.
In this article, I first discussed the importance of customer relationship management from the view of customer lifetime value (LTV). In the case of Fast Lube, I compared the difference of LTV before or after the customer retention
strategy had been carried out. After that, the technology structure and current usage of CRM system was described. The article pointed out the main shortcomings of operational CRM which is it can not understand the real needs of customer and can not help us carry out pertinent business measures. Form the view of ROI, return of investment, the article discussed the advantages of analytical CRM in the aspect of improving customer profitability and drew the conclusion of analytical CRM was the direction of CRM system. At last, the article discussed the problem of how to use business intelligence to build analytical CRM through a lot of cases and recourses.
* From the technologies of business intelligence
Firstly, we should build customer information data warehouse to abstract and transform data from different recourses to make to following steps easier; secondly, we can use query & reports or OLAP to solve verification problems related to customer relationship management; finally, we can use data-mining to solve the discovery problems, that is, discover unknown information from mass data and predict the future.
* From the applied fields of CRM
Generally speaking, business int
引文
[1] 林宇等 《数据仓库原理与实践》人民邮电出版社 2003/01 第1版
[2] 陈京民等 《数据仓库与数据挖掘技术》电子工业出版社 2002/08 第1版
[3] Alex Berson等 《构建面向CRM的数据挖掘应用》人民邮电出版社 2001/08 第1版
[4] Bernard Liautaud, Mark Hammond著《商业智能：信息、知识、利润》电子工业出版社 2002/05 第1版
[5] 覃征等 《网络企业管理》 西安交通大学出版社 2001/09 第1版
[6] Gary P. Schneider, James T. Perry著 《电子商务》 机械工业出版社 2000/09 第1版
[7] Arthur Middleton Hughes, Customer Retention: Integrating Lifetime Value into Marketing Strategies, DM Review
[8] Arthur Middleton Hughes, How Lifetime Value is Used to Evaluate Customer Relationship Management，DM Review
[9] Tania Morr, CRM and BI Total Cost of Ownership DM Review, January 2002
[10] Stephen Cranford, How a Multinational Retailer Integrated BI, CRM and KM, DW Review
[11] 田同生:《CRM中的决策支持系统与商业智能》www.yesky.com (2001-09-03)
[12] 管政：《"前台"CRM与"后台"CRM》www.ctiforum.com 2002/04/09
[13] 李新明：《谈CRM与ERP的发展趋势与整合》www.amteam.org
[14] 李蓓：《分析型CRM：留住黄金客户》www.ctiforum.com 2001/12/03
[15] 朱爱群编著. 中国财政经济出版社.《客户关系管理与数据挖掘》 . 2001
[16] 《如何进行客户管理》亚太管理训练网 http://www.longjk.com
[17] Kennette Reed：Customer Retention Starts on the Inside
[18] GreaterChinaCRM：《中国CRM行动指南》（第二版）2003/01 p5-p10
[19] 蔡伟杰等 《关联规则挖掘综述》
[20] W. H. Inmon, Building the Data Warehouse Design
[21] 王珊 《数据仓库技术与联机分析处理》
[22] 杨永恒 《客户关系管理——价值导向及使能技术》2002/12
[2] 陈京民等 《数据仓库与数据挖掘技术》电子工业出版社 2002/08 第1版
[3] Alex Berson等 《构建面向CRM的数据挖掘应用》人民邮电出版社 2001/08 第1版
[4] Bernard Liautaud, Mark Hammond著《商业智能：信息、知识、利润》电子工业出版社 2002/05 第1版
[5] 覃征等 《网络企业管理》 西安交通大学出版社 2001/09 第1版
[6] Gary P. Schneider, James T. Perry著 《电子商务》 机械工业出版社 2000/09 第1版
[7] Arthur Middleton Hughes, Customer Retention: Integrating Lifetime Value into Marketing Strategies, DM Review
[8] Arthur Middleton Hughes, How Lifetime Value is Used to Evaluate Customer Relationship Management，DM Review
[9] Tania Morr, CRM and BI Total Cost of Ownership DM Review, January 2002
[10] Stephen Cranford, How a Multinational Retailer Integrated BI, CRM and KM, DW Review
[11] 田同生:《CRM中的决策支持系统与商业智能》www.yesky.com (2001-09-03)
[12] 管政：《"前台"CRM与"后台"CRM》www.ctiforum.com 2002/04/09
[13] 李新明：《谈CRM与ERP的发展趋势与整合》www.amteam.org
[14] 李蓓：《分析型CRM：留住黄金客户》www.ctiforum.com 2001/12/03
[15] 朱爱群编著. 中国财政经济出版社.《客户关系管理与数据挖掘》 . 2001
[16] 《如何进行客户管理》亚太管理训练网 http://www.longjk.com
[17] Kennette Reed：Customer Retention Starts on the Inside
[18] GreaterChinaCRM：《中国CRM行动指南》（第二版）2003/01 p5-p10
[19] 蔡伟杰等 《关联规则挖掘综述》
[20] W. H. Inmon, Building the Data Warehouse Design
[21] 王珊 《数据仓库技术与联机分析处理》
[22] 杨永恒 《客户关系管理——价值导向及使能技术》2002/12