Using Data Mining to Increase Customer Life Time Value

Transcription

1 Using Data Mining to Increase Customer Life Time Value Chu Chai Henry Chan Hsin-Chieh Wu E-Business Research Lab Department of Industrial Engineering and Management, Chaoyang University of Technology Wufeng, Taiwan, R.O.C. Abstract: - In the recent years, customer relationship management (CRM) is an important issue for marketers. Most of marketing researchers intend to compute customer value and provide valuable activities when customers demand are occurring. Unfortunately, the linkage between customer value and campaign management is still missing. This research proposes lifetime value (LTV) model in analyzing customer lifetime value to construct new promotion plans. To condcut a case study, this investigation extracts more than ten thousand customer data from a Nissan dealer for finding the customer detecion. From evaluating the customer lifetime values, we found two interesting outcomes. The first one is that the activties of maintenance item is more valuable than those activies of car purchasing and car issurance. The second one is that a lot of new customers lost their loyalties on the second year. To slove the two problems, this research suggest the studied Nissan dealer shoutd take two corresponding promotion stratetgies. First, the dealer should provids five-years assurance for new car purchasing customers to increase their loyalty. Secondly, they need to decrease the price of maintenance and speed up the process of maintenance to attract customers. Key-Words: - Customer Relationship Management, Customer Lifetime Value, Data Mining 1. Introduction Customer Relationship Management (CRM) is one of the most important strategies for many corporations to retain customers and keep profits in the global competitive environment. The major tasks of CRM include customer acquisition, customer development and customer retention [4]. Most existing researches focus on using lifetime value to compute the value of customer. For example, Verhoef applied a regretion model to prodict the potential value of customers in the insurance industry[11]. Hwang used LTV model on the wireless telecommunication industry in year 25 [6].. Unfortunately, most existing researches still stay in academic study. Most of private sectors did not know how to use customer value to increase their value and profitability. This paper is aiming to demonstrate the application of using the customer lifetime value model in finding the problems of customer detection and create new strategies for value creation. Based on the proposed procedure, most companies can find the customer detection and provide a better strategy for increasing customer value. By using the customer value evaluation processes, the potential customer value can be added. In this paper, the proposed concept is applied in a case study of automobile industry. 2. literature Review Two significant issues must be addressed in this study. First, the existing data mining researches would be reviewed. Secondly, the previous studies of custom life time value should be depicted shortly. 2.1 Customer Value Most companies dedicate all of their resources to attract customers, but not all of customers are valuable. Since 198, many corporation and researches developed customer relationship management to sustain the profitable customer as longer as possible [5,9,12,13,14]. In previous study, CRM comprises the activities; customer acquisition, customer cultivation, and customer retention [6]. The customer value are classified into three categories; current value, potential value and loyalty [6,11]. The value creation process consists of three key elements: the value customer receives; the value organization receives; and, by successfully managing this value exchange, maximizing the lifetime value of desirable customer segments [1]..

2 2.2 Data Mining Even CRM had employed in the real world for half of a decade. Still a lot companies face the painful problem in losing their most profitable customers daily. Some studies began to use data mining as a tool to dig out the hidden problems of missing customer loyalty [2,17]. The technology of data mining is defined as a sophisticated data search capability that uses statistical or intelligent algorithms to discover patterns and correlations in data [3,15]. Several studies use data mining to segment customers in bank, insurance, telecommunication industry, retailer, hospital [1,6,7,8,11,15] 3. Mining Customer Value Recognition of the importance of relationships in recent years has inspired marketers to focus on the maintenance of exchange relationships rather than on the accumulation of transient transactions[16]. IBM Corp. recommends four steps to employ relationship marketing with valuable customers (see Fig. 1). First, product information has to be prepared by marketers. Secondly, marketing and merchandising techniques need to deliver key information to customer through multiple mediums, such as webs, , POS and call. Thirdly, the merchandising results would be analyzed into the forms of business intelligence. Finally, the business intelligences would be utilized to construct the profile of customer for understanding the behaviors of customers. The purpose of CRM is to min customer value for finding a good successful promotion plan. Most of existing researches only focus on mining customer behavior and fail to link with any promotion plan. To cope with this problem, this research proposes a customer lifetime value model to conduct data mining in finding the changes of customer value. From the analyzed result, automobile dealer can figure out the trend of value creation processes. Then a valuable promotion plan would be generated to establish a good relationship with customers and increase profitability (see Fig. 2). 4. Customer Life Time Value To analyze the lifetime value of customer, the core costs of customer in automobile industry incorporates three items; purchase, insurance and maintenance. The purchase information represents the cost of purchasing a new car. The insurance information denotes the cost of customer attending an insurance policy. The maintenance cost is the expenditure on maintaining a car annually. - Product Information - Other Web Info. Customer information -Profile - Demographics - Behavioral - Business System's Data Personalization Capabilities Content Business Manager Flexible Rules Builder Merchandising Center Business Intelligence - Marketing and Merchandising Techniques Fig. 1 Relationship Marketing Cycle Enhanchements By IBM [4] Customer Data Collection Customer Value Creation New Promotion Plan What's delivered W eb POS Call center Print - Data mining - User analysis - Usage analysis - Commerce analysis Built Customer Lifetime Value Data Mining for Customer Retention Fig. 2 The Framework of Mining Customer Lifetime Value Furthermore, the revenue for each customer and items are different. For example, the average revenue of purchasing a new car is close to seven percentages. The average revenue of inviting a customer attending an issuance is fifty percentages. Based on this notation, the average revenue for each items are different. In this paper, the LTV model is defined as following:

3 Current Value = Ci n n ( Vm + V I + V M )(1+r) i ic x cxt y cyt Ci = t= 1 t= 1 Where: t-1...(1) V i = The Average Revenue of Car Purchasing Profit M ic = Car Purchasing Profit Contribution of Customer i At Period c V x = Insurance Profit contribution of Customer i I cxt = The average revenue of Insurance Profit V y = Maintenance Profit contribution of Customer i at period t M cyt = The average revenue of Maintenance Profit r= Interest Rate t= The length of years which cars are purchased 5. Case Study To conduct an empirical study, this study collect historic data of 1456 customers from a Nissan automobile dealer in Central Taiwan (see figure 1). From the aspect of customer data in table 1, we found that 86.97% of customers purchase only one car and 13.3% of customers own two cars or more. Under the investigation, the most popular models are Sentra, Cefiro and March (see Table 2). Based on the analysis of our study, the customer lifetime value is gradually increasing annually (see Fig. 3), but not all value item are increasing. From Fig. 4, we found that only the value of maintenance item is increasing annually. Car purchase and insurance would not the most value-added processes. Fig. 5 shows that the numbers of customer are decreased gradually after the second year. It means that most customers lost their loyalties on the second year. After data mining, we suggested that the studied Nissan dealer need to launch a new promotion plan to retain customer on the second year and increase customer value on maintenance. Table 1 the Basic Customer Information based on the Length of Year Year Number Percentage 86.97% 1.45% 1.59%.35% Year 5 6~12 Over12 Number Percentage.21%.37%.6% Table 2 The Distribution of Car March Sentra Cefiro X-Trail Serina Number Percentage 16.43% 37.1% 22.2% 6.94% 2.% Teana Primera AD Cabstar Others Resort Number Percentage 1.19% 2.8%.9% 1.12% 1.14% After several creation meetings, a credited-based program is designed by the Nissan deal. In this program, the dealer would provide five-years assurance for all of new car purchasing customers to increase their loyalty. From the marketing surveys, we also found that prices and time consuming on maintenance are two key factors of affecting customer s maintenance. To improve such problems, the dealers began to decrease the price of maintenance and speed up the process of maintenance to lure customers. The details of two proposed strategies are shown in table 3. 16,, 14,, 12,, 1,, 8,, 6,, 4,, 2,, Current Value Potential Value Life Time Value Average Average Average Fig.3 the Customer Life Time Value Distribution

4 8,, 7,, 6,, 5,, 4,, 3,, 2,, 1,, Fig. 4 the Customer Value Distribution for each Purchasing Items 2,5 2, 1,5 1, 5 Fig. 5 the Customer Number Distribution Table 3 Value Creation Strategy Problems Value Strategy Creation Process 1. Only the value of maintenance item is increasing Provide a good offer on maintenance to attract Decrease the price of maintenance and speed up annually. customers and the process of retain loyal customers. maintenance to attract 2. Customers lost their loyalty on the second year. Car Purchase Maintenance Insurance Average Average Average Car Purchase Maintenance Insurance Implement a promotion plan to retain customer customers. A five-years assurance program are launched 6. Conclusion Even CRM had applied in several industries by many well-known companies such as IBM, Cisco. At the mean time, a lot of previous researches proposed data mining to find the patterns of customer behavior [3,15]. For instance, Hwang and Verhoef applied differnt models to compute the customer value[4,11]. The link between customer value and promotion plans is still msssing. To deal with the obstacle, this paper proposes a mechanism to mine the value of customer lifetime value for support a valuable promotion plan. The major contributions for this paper as follows: Propose a customer value model in automobile industry Connect the customer value and promotion by data mining This study compute the customer value of 1456 data. From evaluating the customer lifetime values, we found two interesting outcomes. The first one is that the activties of maintenance item is more valuable than those activies of car purchasing and car issurance. The second one is that a lot of new customers lost their loyalties on the second year. To slove the two problems, this research suggest the studied Nissan dealer shoutd take two corresponding promotion stratetgies. First, the dealer should provids five-years assurance for new car purchasing customers to increase their loyalty. Secondly, they need to decrease the price of maintenance and speed up the process of maintenance to attract customers. After this study presents a LTV model to find the problems of customer detection and provide value creation strategies, one of major problems for promotion is to search the right and profitable customers. In the future, more researches should be conducted in the ares of targeting customers to assist automobile retailers implementing effecitve promotion plans. Acknowledgments The authors would like to thank the EMPower Corporation and National Science Council of the Republic of China, Taiwan for financially supporting this research. References: [1] Atakan O ztu rk, Sinan Kayalıgil, Nur E. O zdemirel, Manufacturing lead time estimation using data mining, European Journal of

5 Operational Research, No.173, 26, pp [2] Ching-Yao Wang, Shian-Shyong Tseng,Tzung-Pei Hong, Flexible online association rule mining based on multi dimensional pattern relations Information Sciences, No.176, 26, pp [3] Chris Rygielski, Jyun-Cheng Wang, David C. Yen, Data mining techniques for customer relationship management Technology in Society, No.24, 22, pp [4] Emmy Liu, CRM in e-business era, CRM Conference, 21. [5] F. R. Dwyer, Customer Lifetime Profitability to Support Marketing Decision Making, Journal of Direct Marketing, Vol. 3, No.4, 1989, pp [6] Hyunseok Hwang,Taesoo Jung, Euiho Suh, An LTV model and customer segmentation based on customer value: a case study on the wireless telecommunication industry, Expert Systems with Applications, No.26, 24, pp [7] Jules J. Berman, Confidentiality issues for medical data miners, Artificial Intelligence in Medicine, No.26, 22, pp [8] Nan-Chen Hsieh, An integrated data mining and behavioral scoring model for analyzing bank customers, Expert Systems with Applications, No.27, 24, pp [9] P.D. Berger and Nasr, N.I., Customer Lifetime Value : Marketing s and Applications, Journal of Interactive Marketing, Vol. 12, No. 1, 1998, pp [1] Payne Adrian, The Value Creation Process in Customer Relationship Management, Working paper, Cranfield School of Management Cranfield University, UK, 22. [11] Peter C. Verhoef, Bas Donkers, Predicting customer potential value an application in the insurance industry Decision Support Systems, No.32, 21, pp [12] R. S. Duboff, Marketing to Maximize Profitability, The Journal of Business Strategy, Vol 13, No.6, 1992, pp [13] Robert C. Blatterg and John Deighton, Manage Marketing by the Consumer Equity Test, Harvard Business Review, Vol.74, July-August, 1996, pp [14] S Gupta., & Lehmann, D. R.. Customers as assets. Journal of Interactive Marketing, Vol,17, No. 1, 23, [15] S. C. Hui, G. Jha, Data Mining for customer service support, Information & Management, No.38, 2, pp [16] Sharon E. Beatty And Morris Mayer, James E. Coleman, Kristy Ellis Reynolds, Jungki Lee, Customer-Sales Associate Retail Relationships Journal of Retailing, Vol.72, No.3, 1996, pp [17] Show-JaneYen and Yue-Shi Lee, An efficient data mining approach for discovering interesting knowledge from customer transactions, Expert Systems with Applications, No.3, 26, pp