從資訊覓食理論探索消費者對於電子商務網站推薦系統的點擊行為

Abstract

電子商務網站中的推薦系統，可以幫助消費者過濾出潛在所需的訊息。推薦系統的點擊紀錄除了可以回饋推薦系統的推薦結果，同時也可作為後端演算法優劣的比較依據。然而過去對於推薦系統的點擊研究較為稀少，因此本研究試圖用資訊覓食理論來探索消費者於推薦系統的點擊行為。首先採用實地實驗法，從推薦系統的「提示資訊」與推薦演算法產生的「商品資訊」來探索消費者點擊的差異性；再利用質化訪談的方式，深入理解消費者點擊背後的影響因素。研究結果發現，消費者會受到提示資訊與商品資訊影響其對推薦系統的點擊。其中提示資訊又比商品資訊在瀏覽過程中率先吸引大多數消費者的目光。此一消費者點擊行為可用資訊覓食理論來加以解釋與闡述。

Recommender systems in e-commerce assist consumers to filter out overloaded messages. The clicking log can not only be the feedback for improving the recommendations but also serve as a comparison basis for the recommendation algorithms. However, past studies on recommender systems rarely focus on the clicking behavior from the interface perspective. Therefore, this study attempts to apply the Information Foraging Theory to explore consumers` clicking behavior in the recommender systems. This study firstly conducted a field experiment to explore whether consumers` clicking behavior would be affected by the ＂title＂ of the recommendations or the ＂product information＂ generated by the recommender algorithm. We also conducted a qualitative interview to deeply understand the influencing factors behind consumers` clicking behavior. The result shows that consumers` clicking behavior is affected by both prompt and product information on the recommender system but the prompt information can catch users` attention earlier in the browsing process. This consumers` clicking behavior of the recommender system can be explained by the information foraging theory.