以混和式資料採礦技術探討信用卡產業之顧客流失管理與服務組合設計

Abstract

近年來，顧客關係管理儼然成為一個重要的議題。對於服務業而言，如何建立企業與消費者的良好關係，將是企業增加利潤的主要來源。而顧客關係管理中，最為廣泛討論的議題即是對於流失顧客的辨別與保留。本研究針對台灣信用卡產業進行顧客流失分類，除了找出可能流失顧客外，並判斷影響顧客流失之關鍵因子；最後再根據顧客偏好建立信用卡服務組合，期望刺激顧客辦卡及消費欲望。不同於以往，過去研究多半使用顧客交易資料進行流失分類；本研究則採用顧客滿意度資料進行流失分類，並證明其分類效果較佳。 本研究除了建立決策樹分類規則供企業判斷流失顧客外，更歸納出影響流失之關鍵因素在於其對銀行所提供之“道路救援服務”感到不滿意；為快速且有效地保留顧客，企業應優先改善該項服務。最後則根據顧客偏好，利用關聯規則建立卡片組合。本研究建議企業在成本允許之下，規劃一張卡片擁有“量販超市優惠”或“加油停車優惠”或是“休閒娛樂優惠”為主要功能，並可搭配“居家生活購物優惠”或是“住宿餐飲優惠”為選配功能。此卡片組合乃是根據顧客偏好所建立，故其相當符合顧客期望。

Customer Relationship Management (CRM) has become an interesting research topic over many years because of its significant impact on increasing revenues in the service industry. One of the most popular topics is discussing how to detect the potential customer churn and then employ customer retention policies before final withdrawal. In this thesis, Decision tree, Rough sets and Association rule are combined to construct classification rules (based on customer satisfaction with service policies) to identify potential customer churn and to configure the optimal service portfolios (based on customer preference for product features) for hundreds of Taiwanese credit-card users. Instead of using the transaction data or customer profiles, customer satisfaction as well as customer preference has been demonstrated much more effective to accomplish the aforementioned goal(s). In summary, the main factor causes customer churn is“roadside assistance service” provided by credit-card company. The factor should be improved immediately because the more disappointed the customer experienced, the more chance they will stop the transaction with the company in the future. For credit-card service portfolio, one of the following features is suggested to act as a “primary” functionality, such as (1) Co-branding with supermarkets, (2) Discounting on refuel and parking, and (3) Discounting on entertainment. And one of the following features can be selected as a “secondary” functionality, like (1) Discounting on purchasing household goods and (2) Discounting on hotels or restaurants.