近年來,顧客消費行為的異質性受到普遍的重視,讓行銷人員人必須更深入地瞭解顧客消費的偏好,以達到有效的顧客關係管理外;競爭激烈的信用卡市場讓發卡機構必須透過精準的顧客關係管理,計畫出更多的行銷方案來吸引持卡人刷卡消費。 本研究是結合了層級貝氏模式與Stone提出的RFM模型,將顧客有效地分群。先透過層級貝氏模式,以generalized gamma分配來模擬顧客消費間隔時間,再利用混合層級貝氏模式將顧客分成四種狀態。而RFM模型則可以衡量出顧客過去貢獻的價值,將顧客依其價值分群。本研究將兩種分群結合不僅可以捕捉到顧客異質性的消費行為,還可以將不同狀態與價值的顧客進行市場區隔,讓行銷人員可以針對不同群的顧客採取不同的行銷策略,亦可決定行銷的先後順序。 本研究透過分層抽樣的資料,驗證兩種分群結合的有效性,可以作為信用卡發卡機構全面分析的參考。
Consumers’ consumption behavior is the most important studies in the customer relationship management (CRM). In recent years, the heterogeneous among customers let marketers have to understand deeply. In the other way, the vehement competition in the market let the credit card organizations to design more attractive promotion plans for more the credit card expense. In this study, we combine the Hierarchical Bayesian model (HBM), which is mostly used to estimate the inter-purchase time of customer, and the Stone’s RFM model, which can be use to weight the customer value, to categorize customers efficiently. In the Hierarchical Bayesian model, we uses a generalized gamma distribution for modeling inter-purchase times and then uses the mixture generalized gamma distribution to segment customer into four states. By using the RFM model, we can cluster customers into groups of various customers’ value. The contribution of the study is combing the BHM and RFM model. Not only catches the heterogeneous behavior of customer purchases but also segments the customers for several groups effectively. Good segment let marketers to adopt many diverse marketing plants and different order to promotion. This study used the stratified sampling data and check the efficiency of two classification, may provides some references to the credit card organization.