|摘要: ||台灣已進入資訊社會時代，但由於地狹人稠的特性，資訊商品通路也逐漸在改變，以往地區型電腦資訊通路及集中商圈為主的通路型態，近年來已經慢慢被3C連鎖通路所取代，但資訊商品所含概的客戶族群十分廣泛，一般以中型企業及大型集團性的公司所需的產品規格與服務，卻非3C連鎖通路所能提供，因此商用的資訊商品的銷售仍需依賴特定通路來完成，目前此類型通路散佈於全省各都會地區，在同業與相關產業的激烈競爭下，資訊產品品牌商除自身應該發揮強而有力的品牌價值外，通路價值的經營也是企業不能輕忽的課題。本論文的目的有三：（1）整合品牌商旗下之經銷商資料庫，並利用資料庫探勘技術及統計分析工具，將經銷商作分群找出高低價值的經銷商族群；（2）根據市場區隔理論及決策樹分析結果，發掘區域經銷商之採購特性，開發潛力型之經銷商並確認其價值之所在；（3）研究品牌商如何作到資源集中化，效率極大化，根據不同價值之經銷商及區域性特徵，進行差異化行銷策略及強化通路管理競爭力之探討。本研究將運用資料探勘(Data Mining ) 技術對個案公司進行實證研究，首先運用RFM（Recency、Frequency、Montary）模型指標的取得，再透過 K-means 集群分析將顧客進行分群，將其分群結果對應其區域特徵及 RFM 等分類變數，再依 CART（Classification and Regression Tree）方法之分類結果，觀察出目標經銷商的屬性及特徵。經由產業及實證資料分析後，本研究將以「顧客價值分析」與「顧客分類特質分析」兩個主題分別說明不同顧客類型之差異，接著針對個案參考不同類型的顧客群特性，分別擬定適當的行銷策略。|
As Taiwan enters the era of being an information society, marketing channels for information products on this densely populated island are gradually changing. In recent years, conventional marketing channels for information products — which mainly include localized distribution channels and centralized shopping districts — have been replaced by 3C (computers, communication products and consumer electronics) chain stores. However, customer segments in the market of information products are greatly diversified. Products and services generally required by medium-size businesses and group corporations, for example, are not available in these 3C chain stores. In other words, information products for businesses are sold through particular channels, which in Taiwan are distributed in metropolitan areas. Therefore, brand companies of information products not only have to exercise the power of their brand value, but should also pay attention to the management of channel value to survive heated competition from peers and relevant industries.
This thesis has the following three objectives: (1) to integrate the distributor databases of brand companies and, using the data mining technique and statistical analysis tools, group the distributors according to their values; (2) to mine the procurement characteristics of local distributors, develop potential distributors and confirm their values according to the theory of market segmentation and the results of decision tree analysis; and (3) to study how brand companies can achieve resource centralization, maximization of efficiency, and discuss differentiated marketing strategies and the enhancement of competitiveness in channel management for distributors of different values according to the localized characteristics.
In this study, the data mining technique is applied to case companies for validation. Customers are divided into groups by first determining the RFM (Recency, Frequency, Monetary value) model attributes and then performing a K-means cluster analysis, so that the grouping results match the localized characteristics and the RFM classification variables. The attributes and characteristics of target distributors are then identified according to the classification results obtained by using the CART (Classification and Regression Tree) technique.
After analyzing data obtained from the industry and the validation results, the differences among customer types are analyzed under two themes: “analysis of customer value” and “analysis of customer classification characteristics.” Finally, appropriate marketing strategies are proposed for the respective case companies with reference to the characteristics of different customer types.