MARK Thesis Presentation - Anthropomorphized Recommendation System and Attribute Type
2:00pm - 3:00pm
Online via Zoom (https://hkust.zoom.us/j/96999454229)

As online shopping gets more prevalent, the product-advising function originally performed by salespersons is being increasingly taken over by online recommendation systems. Although recommendation systems contribute a lot to e-commerce companies’ sales [1], most of consumers still indicate that they prefer to interact with humans [2]. One possible reason is that humans are perceived to be able to account for consumers’ unique characteristics [3]. To improve consumers’ online search experience, marketers sometimes anthropomorphize their online recommendation systems. For example, Baidu anthropomorphized their recommendation system by endowing it with a human-like face and named it Xiaodu to recommend food and places to visit to consumers. However, limited research has examined under what circumstances anthropomorphizing a recommendation system leads to positive consequences and under what circumstances it may backfire. Our research aims to address this question. Specifically, we examine how consumers react to anthropomorphized versus non-anthropomorphized online recommendation systems as a function of search attribute and experience attribute. We argue that consumers who focus on experience attributes would perceive the anthropomorphized recommendation system to be more accurate. In other words, consumers perceive that the anthropomorphized recommendation system is more likely to recommend options that match their preference, which in turn leads to more positive attitude towards the recommended option. On the contrary, consumers who focus on search attributes would perceive the non-anthropomorphized recommendation system to be more accurate, which results in more positive attitude towards the recommended option.


The conceptualization of the present research informs past work on several fronts. First, the present research contributes to the literature on search/experience attribute. Past research on search/experience attribute focuses on how search/experience attribute influences consumers’ trust in the product information [4], [5], [6] and marketers [7]. The present research examines how search/experience attribute influences consumers’ perception about the recommendation system and product attitude. More specifically, the present research shows that the effect of anthropomorphism on perceived accuracy of the recommendation system and the attitude towards the recommended product is depending on attribute type consumers are focusing on. Secondly, the present research contributes to the literature on anthropomorphism. Past research mostly focuses on identifying the how anthropomorphizing products or brands influence consumers perceptions on affective dimensions (e.g., emotional intensity [8], enjoyment [9]) and cognitive dimensions (e.g., perceived risk [10], perceived mindfulness [11]), as well as the downstream consequences. The present research combines the literature on online recommendation system and identifies another aspect on which anthropomorphized targets can be perceived, namely perceived accuracy, and examines the downstream consequence.


Apart from its theoretical contribution, the current research also offers managerial implications for marketers. As online shopping is becoming increasingly popular, more and more companies begin to adopt online recommendation systems. Our research sheds light on how these companies can optimize the design of online recommendation systems to maximize the effectiveness of these recommendation systems.

講者/ 表演者:
Lulu SHI
Department of Marketing, HKUST
語言
英文
適合對象
教職員
公眾
研究生
本科生
主辦單位
市場學系
聯絡方法
新增活動
請各校內團體將活動發布至大學活動日曆。