近年來,企業積極發展目標顧客,使用資料探勘(Data Mining)的分析技術研究顧客行為(Customer Behavior),精確地鎖定潛在顧客(Potential Customer),為企業帶來更大效益,故顧客關係管理(Customer Relationship Management, CRM)更加不能忽視,而顧客價值分析是顧客關係管理的基礎,過去文獻中分析顧客價值,RFM模型是最常被提及與運用的方法,藉由RFM模型的三個衡量指標,可以簡單而清楚的看出顧客消費行為的輪廓。學者許雅涵等人(2011) 以RFM模型為基礎,考量石油在產品、定價策略上的特性改良為RFQ模型。本研究以加油站為例,以RFQ模型為基礎,提出以基因演算法演化RFQ三項指標之最佳權重,修正為GA-WRFQ模型,並配合K-means方法進行群集分析,輔助加油站將顧客依消費行為進行群集分析,做為加油站業者後續策略擬定的參考,如此一來,加油站業者可以針對目標顧客進行有效的行銷策略,並提升加油站競爭優勢。
In recent years, organizations seek to find target customers, and use analytical techniques in Data Ming domain to understand customer behavior while focus on potential customer and bring more revenue to organization. So Customer Relationship Management (CRM) becomes more important. Customer value analysis is the cornerstone of CRM. In the past studies, RFM is the method mentioned and used in customer value analysis most popularly. By using the three indicators of RFM, one can clearly understand how the customer behaves in a simple way. Considering the product and pricing strategy character in oil industry, Shiu, et al. (2011) enhanced RFM model and proposed RFQ model. In this research, we propose a GA-WRFQ model, which is based on RFQ model, and using Genetic Algorithm (GA) to evolve weights of the three indicators in RFQ model. Based on K-means clustering algorithm, this mechanism can cluster customers by their consuming behavior. We also apply this mechanism to a real oil station, help them cluster customers and develop strategies. In this way, oil station manager may develop strategies that focus on target customers and make every penny count.