A case study of application of data mining to achieve customer relationship management

• Main Author: 蘇俐文
• Other Authors: 吳信宏 博士
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/02850664609251158615

碩士 === 國立彰化師範大學 === 企業管理學系 === 100 === Companies can dig out useful customer information from the past consumption data through clustering methods to segment customers and understand customers' characteristics and purchasing behaviors by offering the right products or services to them and establishing a good customer relationship management to improve profitability, attract potential customers, and develop long-term marketing strategies or services. This study uses LRFM (Length, Recency, Frequency, and Monetary) model to transform and analyze customers’ transaction records, and customers can be classified into twelve groups by self-organizing maps and K-means method. Descriptive statistics of the twelve groups are discussed. In addition, twelve groups are reduced to five groups based on the performance of LRFM variables, i.e., LRFM values above the averages, LRF values above the averages, RFM values above the averages, and LRFM values below the averages. Further, three categories are found based on the five customer bases proposed by Chang and Tsai (2004), namely core customers base, lost customers base, and new customers. Finally, this study uses the customer relationship matrix proposed by Marcus (1998) to divide twelve groups into four categories: long-term frequency customers, short-term best customers, long-term best customers, and short-term uncertain customers. To identify the top five commodities purchased by the each groups. In accordance with the results, this study suggests the suitable marketing strategies and products for different scenarios. By evaluating different customers’ characteristics, different marketing strategies can be made to identify high-value and potential customers for the company such that the company can allocate its limited resources more appropriately to obtain the best configuration methods to improve customer satisfaction, loyalty, and corporate profitability.