當今產業界的競爭要素,由以往的提升企業內部核心競爭力,轉變成目前以滿足顧客需求為主軸,唯有提供貼心的服務,才能真正抓住客戶的心。對企業而言顧客的管理是非常重要的,根據Ravi Kalakota 和Marcia Robinson(1999)認為:若要妥善地管理顧客週期,就必須分成三個階段來執行顧客關係管理。這三個階段分別為獲得新顧客、強化現有顧客的獲利性,以及維持現有顧客的終身價值。一般來說,企業大多著重於開發新顧客上,而忽略了與舊顧客維持長期關係。根據Peppers,Don 和Rogers,Martha 認為:開發一位新顧客所花費的成本是留住一位舊顧客的五倍。此外,大部分的企業每年平均有高達25%的顧客會流失,而若顧客流失率減少5%,則企業的利潤將有100%的成長。因此探討顧客流失及顧客價值為本研究的主要動機之一。 本研究是以A公司為主的個案探討,所選擇的產業為運動休閒娛樂產業,此產業與顧客的關係有著密不可分的特性,而個案中所探討的公司係為國內此產業的標竿,本研究希望從此個案公司瞭解如何應用公司的會員資料庫進行顧客流失分析,研究中採用區別分析、羅吉斯分析、類神經網路分析及決策樹分析等肆種分類技術,進行顧客流失數學模式的推導,推導中逐步找出重要的預測變數,並比對顧客流失分析數學模式的判別率,以確立判別率較高的數學模式;另一方面是將個案中的未流失及己流失的會員中取得相關的交易資料,採用Hughes,Arthur〈1994〉RFM顧客價值分析,同時應用顧客的購買期間進行MLE、WMLE的顧客價值趨勢分析。本研究的最後成果會提供給個案中的A公司,做為A公司及同業顧客流失、顧客價值及顧客價值趨勢分析的相關參考依據。
The competition element in nowadays industry has transferred from promoting inner core competition to satisfy customers’ needs. Only providing intimate service can seize the hearts of customers. Customer management is very important to an enterprise.According to Ravi Kalakota and Marcia Robinson(1999), CRM must be implemented in three stages in order to manage customer cycle properly: gain new customers, strengthen the profitability of the exiting customers and maintain the lifelong value of the existing customers. Generally speaking, most enterprises focus more on developing new customers than maintaining long term relationship of existing customers. According to Peppers,Don and Rogers,Martha the cost of developing a new customer is five times more than retaining an existing one. Besides, 25% average customers have been lost every year. The companies would have 100% profit growth if the customer lose rate is decreased by 5%. Therefore, investigating customer lose and customer value are one of the major motives in this research. This research is focused on A company, which is an industry combining sport, leisure and entertainment. This industry has crucial relationship with customers and is the model in domestic industry. This research is to understand how to apply members database to analyze customer lose. It adopts four classification techniques: discriminant analysis, logistic analysis, neural networks analysis and decision trees analysis to guide mathematics model in customer lose. Then, the main forecasting aviations will be conducted gradually and be compared to differentiate rate in order to establish higher differentiate rate of mathematics model. On the other hand, it adopts RFM customer value analysis of Arthur Hughes (1994) to analyze related transaction information from existing and lost members. In addition, it also applies customer purchase period to implement MLE and WMLE customer value trend analysis. The result of this research will provide A company the related reference guide in analyzing customer lose, customer value and value trend. Key works: Data Analysis、Multiple Classification、DataBase Marketing、Customer Relationship Management、Customer Lose、Customer Value.