| Summary: | 碩士 === 國立臺北大學 === 統計學系 === 100 === In this study, a refined RFM model and two clustering methods, two-stage clustering and integrated self-organizing map and K-means clustering (SOM+ K-means), are proposed to do customer clustering, and to find out the optimal customers’ groups based upon customers’ consumption behaviors in the case of online service. The study results show that the SOM+K-means clustering method has the best performance, which effectively groups customers into four groups, namely, VIP group, potential group, developing group, and low-value without contact group. Besides, this study generalized thirteen customer satisfactions into four aspects by factor analysis of ordinal data, namely, professional, system performance and database factor. By comparing customer’ satisfaction of various customer groups and their characteristics, company can give priority to major customers for improving their service qualities. Finally, based on the study results of customer clustering and satisfaction analysis, we provide suggestions about marketing strategy for each group and recommend about quality improvement to increase company’s profit and also reduce the operating cost as well.
|
|---|
