隨著網路時代的發達,線上購物蓬勃發展,根據經濟部統計處統計,電商佔整體零售業比重已經連續8年成長,其中近幾年以直播的方式銷售商品的賣家更是如雨後春筍般地出現,直播拍賣相較傳統的電商多了與顧客互動的即時性,吸引許多網拍業者投入,然而卻不是人人都能經營得成功。過去的相關研究都是針對顧客的特性進行分析,找出最有價值的客群;本研究透過電商直播訂單管理App所收集的賣家交易記錄,針對賣家的銷售行為模式進行分析,使用階層式分群(Hierarchical Clustering)演算法將賣家分成三群,並導入RFM模型,探討不同類型的賣家特徵,對於不同特性的直播主,找出在不同的時間點要賣哪類型的商品才能更有效率地獲利,提供賣家客製化的銷售策略。
With the development of the internet, online shopping has grown exponentially. According to the Department of Statistics of Ministry of Economic Affairs, the proportion of e-commerce in the overall retail industry has maintained 8 consecutive years of growth. Recently, there are more and more sellers selling their products through live streaming. In contrast to traditional e-commerce, sellers of live streaming auction can interact with customers in real-time. This feature attracts many vendors to try this new way to sell products, but not everyone can achieve success. In the past, relevant researches focus on the characteristics of customers to find out who is the most valuable customer. This research analyzes sales patterns of different sellers by RFM model and hierarchical clustering based on transaction records collected by an e-commerce live streaming order management App. For different kinds of live streaming vendors, find out which types of goods should be sold at different periods to earn most revenue, provide sellers customized sales strategies.