The impact of live comment characteristics on subsequent live streaming sales and user retention

Live comment, as a primary mean of live interaction, addresses the issue of information asymmetry that viewers face during commercial live streaming. While live comments generally enhance user engagement, the varying processing strategies viewers employ towards the same information can lead to different impacts of live comment content on their decision-making. However, existing research has yet to examine the specific mechanisms through which these differentiated processing strategies affect viewer decisions. This gap limits our deep understanding of how live interactions affect viewer behavior in social media commercial live streaming. This paper used the ELM model to investigate how different information processing strategies on live comments dynamically influence product sales and user retention. By scraping data from 61 TikTok (China) live streaming events, we collected 9,636,136 comments, segmented into 3,854 time slices. Our empirical study finds that the quantity of live comments exhibits an inverted U-shaped relationship with immediate sales and user retention. Real-time purchase information and free chat-related comments positively impact immediate sales and user retention, while comments required for participation in lucky draws have the opposite effect. Interestingly, positive product-related comments negatively impact real-time user retention, while negative product-related comments are positively associated with real-time sales. Our research contributes to extend consumer behavior models by capturing the distinct dynamic state of live streaming and provide insights for business to optimize their live streaming performance.