Digital human calls you dear: How do customers respond to virtual streamers’ social-oriented language in e-commerce livestreaming? A stereotyping perspective

• Published in: Journal of retailing and consumer services Vol. 79; p. 103872
• Main Authors: Yao, Ruiqi, Qi, Guijie, Wu, Zhiqiang, Sun, Hua, Sheng, Dongfang
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.07.2024
• Subjects: Linguistic style; Livestreaming commerce; Purchase intention; Stereotype content model; Virtual streamers; Stereotype content model; Livestreaming commerce; Virtual streamers; Purchase intention; Linguistic style
• ISSN: 0969-6989
• DOI: 10.1016/j.jretconser.2024.103872

Social-oriented language as a linguistic style can improve customer experiences in various service contexts, involving both human employees and AI-enabled service agents. The emergence of digital human characters in e-commerce livestreaming, also known as virtual streamers, raises the question: Can social-oriented language used by virtual streamers increase the customers' purchase intention? However, the theoretical understanding of whether, how, and when the interaction effect between virtual streamers' social-oriented language and product type influences purchase intention remains far from conclusive. Based on the stereotype content model (SCM), we developed a theoretical model to explain the mechanism between virtual streamers' linguistic style and customers’ purchase intention. A multi-method study combining two online experiments, one lab experiment, and one focus group was conducted to explore the effectiveness of social-oriented language (Study 1: N = 321; Study 2: N = 292; Study 3: N = 120; Study 4: N = 9). Participants in both Study 1 and Study 2 were recruited online and were randomly assigned to a condition in 2 (linguistic style: social-oriented vs. task-oriented) × 2 (product type: experience vs. search) between-subjects experimental design. Participants were recruited in the university by three blind-to-hypothesis experimenters in Study 3 and were randomly assigned to a condition in 2 (linguistic style: social-oriented vs. task-oriented) × 2 (product type: experience vs. search) × 2 (virtual streamer type: human-like vs. animated) between-subjects experimental design in the behavioral lab. Participants in Study 4 were recruited from a large WeChat group specifically interested in online shopping discounts. The results of Study 1 indicate the effect of social-oriented language on purchase intention solely in the experience product condition. Study 2 demonstrates the mediating role of perceived warmth and perceived competence, and confirms a compensatory effect and a positive relationship between perceived warmth and competence. Study 3 verifies the moderating role of the virtual streamer type while a focus group Study 4 supports the findings from Studies 1–3. Our findings extend the livestreaming commerce literature by exploring the implications of social-oriented language and task-oriented language when the live streamer is a digital human rather than a human. Also, we provide fresh insights into the literature on human-AI interactions by unraveling both the cognitive mechanism and the boundary condition, deepening our understanding of a novel but under-investigated phenomenon. Furthermore, our findings provide practical implications for livestreaming commerce practitioners.