近年來資訊科技快速發展,電子商務成為許多產業亟欲發展的經營領域。透過網際網路,可大幅降低開店所需成本,而消費者也可不受時空限制進行消費。本研究的主要目的在於找出顧客的消費特性,並針對不同顧客的需求,進行行銷策略之探討。 本研究以國內音樂購物網站為研究對象,藉由資料庫中的歷史交易資料進行資料探勘,將交易紀錄利用RFM(Recency, Frequency, Monetary)分析找出顧客的價值,接著利用K-means分群法,將顧客依照其特性做適當分群;並利用顧客資料中的居住地區進行市場區隔,探討不同市場區隔下顧客之購物行為;最後透過決策樹分析找出顧客對熱門商品的消費特徵。根據分析的結果提出適當的行銷策略,做為個案公司顧客關係管理之參考。
Information technology has developed rapidly in recent years. Consequently, E-commerce has become more popular in many industries. Through the internet, stores can not only decrease cost dramatically, but also attract consumers to purchase products without time-and-space restrictions. The main purpose of this research is to identify customer characteristics and to develop the corresponding marketing policies in different customer demands. This research focuses on two main music web-stores in Taiwan and applies data mining techniques to analyze historical transaction data in three step. Firstly, we applied the RFM model to find the value of customers, and then used K-means analysis to cluster customers into groups. Secondly, we segmented the market based on customer’s geographic information and discussed customer behavior for different markets. Thirdly, we applied decision tree to identify customer characteristics in terms of popular merchandises. Finally, based on the empirical results, we proposed several marketing strategies for the companies.