| 摘要: | 隨著資訊科技發展日新月異，3C產品對於普羅大眾而言已是生活中不可缺少的一部分，消費者對於3C產品方面的需求也有一定程度。面對消費者在消費行為方面已有有愈來愈多樣化的表現的情況下，企業在行銷策略上的布局若只仰賴單一的行銷手法已經無法滿足現今時下消費者的需求。因此台灣各家3C賣場通路無不積極發展多樣化的消費管道、獨特的顧客關係管理策略以搶佔3C零售市場大餅。
而零售業實體通路的POS交易資料就成為企業能妥善利用的資源，若能應用資訊探勘的技術於顧客交易資料上，做出顧客區隔，找出關鍵顧客群並投以精準行銷，將能提升顧客留存率同時培養顧客忠誠，創造更多的利潤收入。
因此本研究欲應用台灣某3C賣場之POS交易資料於顧客關係管理，首先將顧客的交易資料以RFM模型為基礎發展出七項顧客交易行為變數，並以這七項變數作為顧客分群的依據，藉由K-means演算法分群顧客為已流失顧客群、VIP顧客群、一般顧客群與有流失風險的一般顧客群，最後針對一般顧客群做出商品推薦。期望提升顧客未來的消費頻率與花費金額，為個案公司創造更多利潤收入。
關鍵字：3C賣場、RFM、顧客分群、推薦系統、協同過濾
;With the rapid development of information technology, 3C products have become an indispensable part of life for the general public, and consumers also have a certain degree of demand for 3C products. Since consumers’ consumption behavior has become more and more diversified, it is insufficient to satisfy consumers if enterprises only rely on a single marketing approach. Therefore, every 3C product retailers in Taiwan actively develop diversified consumption channels and unique customer relationship management strategies to seize the chance of expanding their territory in the 3C retail market.
POS transaction data in retail channels becomes a valuable resource that enterprises can make proper use of. If information exploration technology can be applied to customer transaction data to conduct customer segmentation, identify key customer groups and implement customized marketing strategy, it will help improve customer retention rate, cultivate customer loyalty and create more profits .
This study utilizes POS transaction data of a Taiwan 3C retail store to conduct customer relationship management. First, develop seven customer consuming behavior related variables base on RFM model, using theses seven variables as the basis of the customer segmentation. Then, using K-means algorithm to implement customer segmentation and divide customers into lost customers, VIP customers, regular customers and regular customers who are at risk of loss. Finally, choosing regular customers as target segment to make product recommendation. It is expected to increase the consumption frequency and the spending amount of the customers, eventually help create more profit for the 3C retail store in this case.
Keyword: 3C Retail, RFM, Customer Segmentation, Recommender System, Collaborative Filtering |
