Page 122 - 臺大管理論叢第32卷第1期

Internet Celebrity Economy: Exploring the Value of Viewers’ Comment Features and Live Streamers’
Marketing Strategies in Forecasting Revenue

7. Conclusions and Implications

7.1 Conclusions and Discussions
This study applies the HB model to develop a predictive streamers’ revenue statistics

model and examine the effects of comment metrics, discrete emotional comments, and
streamers’ marketing strategies on viewers’ gift-sending behavior and the cross-level
effects of streamer heterogeneity. We find that the effects of viewers’ comment features and
streamers’ marketing strategies on viewers’ gift-sending behavior depend on the streamer

heterogeneity. Specifically, the more a male streamer chats with the viewers, the higher the
viewers’ gift-sending behavior. Second, the streamers with higher outward beauty receive
more excited comments, and the more they chat with and responded to the viewers, the
higher the viewers’ gift-sending behavior. Third, sociable streamers receive more total,

negative, and complaining comments, and the more they respond to the viewers’ questions
and share food features, the higher the viewers’ gift-sending behavior. A persuasive
streamer has more praising comments, and the more they share food features, the higher
the viewers’ gift-sending behavior. Further, a comical streamer has more excited, praising,

complaining, disappointed, and ridiculing comments, and the more she/he chat with the
viewers, the higher the viewers’ gift-sending behavior.
Past studies on gift-sending behavior in live streaming have pointed out that
comments related to excitement (Zhou et al., 2019) and the interaction between the

streamer and the viewer (Yu et al., 2018) positively affect the gift-sending behavior of
viewers. Similar research has also indicated that the characteristics of the streamer (such as
personalization, sociability, and attractiveness) also positively affect the viewers’ intentions
to send gifts (Wan et al., 2017; Wohn and Freeman, 2020). Nevertheless, these past studies

neglect that there may be an interaction effect between the variables mentioned above,
which would lead to biased conclusions. This study proves that streamer heterogeneity
has cross-level effects on the relationships between the viewers’ comment metrics and
viewers’ gift-sending behavior; between the viewers’ discrete emotions and viewers’ gift-

sending behavior; between streamers’ behaviors and viewers’ gift-sending behavior. The
conclusions of this study can help live-streaming platforms better understand the factors
that affect the viewers’ gift-sending behavior.
Further, prior studies have discussed paid gifts from the perspectives of social

114