資料庫行銷之顧客購買行為特性分析：以直銷公司之會員為例

Abstract

本研究以直銷產業為研究對象。有別於傳統通路的經營方式，直銷產業大多採用會員制的經營模式，透過與顧客間直接且一對一的溝通，建立一種長期且穩定的顧客關係，也是企業利基的來源。隨著資訊科技的進步，此種特殊的經營模式結合了資料探勘的各種方法與技術，使得企業得以從現存之資料庫中挖掘出各種顧客的特徵與屬性、消費行為的趨勢、購買產品的關聯性等等。 本研究以一間知名的直銷化妝品公司為個案分析對象，利用企業內部之會員及交易資料庫，進行資料庫行銷之實證分析工作，研究主題包含兩個部分，第一部分為顧客分群，第二部分為產品關聯性分析。傳統上進行顧客分群時，通常會針對顧客的價值，也就是對企業的貢獻程度來區分顧客；本研究將從顧客購買商品的行為特性面出發，採用新的構面進行分群，將顧客區分為購買商品以自用為主的「消費型」顧客，以及會邀約親友一起購買，擁有分享產品、轉賣等行為的「分享、銷售型」顧客；進行完顧客分群之後，再利用關聯規則分析來挖掘最佳顧客的消費傾向，並與整體顧客之關聯規則進行比較，最後提出具體的策略意涵。 研究結果顯示，利用顧客購買行為分群模型所區分出來的最佳顧客--「分享、銷售型」顧客，對於企業而言也為具有價值之顧客；此外，關聯規則分析的結果也顯示「分享、銷售型」顧客之產品關聯規則與整體顧客之關聯規則非常類似，具有相當程度的一致性。由於顧客購買行為分群模型是從消費行為面切入，對於企業而言，此一分群結果將可確實反映出具有不同購買行為特性的顧客，使得行銷人員可以依據分群結果針對顧客的行為面特性制定行銷策略。

This thesis focuses on direct-sales industry, which differs greatly from others in the business model. In this industry, there is a long-term and stable relationship between the company and its customers. Such an inseparable relationship comes from direct and on-to-one communication with customers, and contributes to the corporate profit. Due to the improvement of Information Technology, many direct-sales companies integrate advanced data mining techniques into existing database in order to explore some important information, including customers’ characteristic, attribute, and purchasing behavior, etc. A Cosmetics Direct-Sales company is chosen to do a case study. There are two subjects in this thesis--Customer segmentation and Product association. In customer segmentation, the most popular and frequently-used model is RFM model, which takes customer value as criteria. This research attempts to propose a new customer segmentation method- Purchasing Behavior Customer Segmentation (PBCS). Customers are clustered based on their purchasing behavior and objective. “Consumption type customers” are those who purchase for personal use, whereas “Sales-and-Sharing type customers” are those who purchase not only for private use but also for selling or sharing. After customer segmentation, we conduct Association Rule Analysis to analyze product association of “Sales-and-Sharing type customers”. The result shows that the best customers--“Sales-and-Sharing type customers” are also valuable customers. In addition, their product association represents the similar pattern with whole customers’ product association. This segmentation model provides expertise another method to classify customers, and thus expertise can work out marketing strategy based on customers’ purchasing behavior.