Preventing traffic manipulation in live-streaming selling: a signalling game perspective
构建品牌与主播的信号博弈模型，发现佣金和接受监控可作为主播真实人气的可信信号，抑制流量操纵能减少定价扭曲，建议在直播合同中嵌入监控条款。
Live-streaming selling has grown rapidly worldwide, and many brands are partnering with high-popularity influencers to boost sales. However, traffic manipulation can mislead brands into overestimating popularity, thereby distorting decisions on compensation, product pricing and live-streaming rewards. In this study, we build a brand–influencer signalling game to show how influencers use observable, costly signals to gain brand recognition. Equilibria are refined using the Cho–Kreps Intuitive Criterion to rule out unreasonable pooling outcomes. We find that commission and monitoring acceptance serve as credible signals of popularity. In equilibrium, the high-popularity influencer selects a higher commission and higher monitoring acceptance while the low-popularity influencer retains symmetric-information benchmark levels. Anticipating a higher revenue share, the brand optimally lowers prices and rewards when partnering with the high-popularity type. We then examine the role of monitoring acceptance and show that suppressing this signal imposes large losses, with identification relying excessively on commission, amplifying price and reward distortions and pushing both parties’ payoffs further from the symmetric-information benchmark. Allowing monitoring acceptance to share the identification burden attenuates these distortions and brings profits closer to symmetric-information benchmark levels. We recommend explicitly embedding monitoring-acceptance clauses and third-party verification in live-streaming contracts to support sustainable industry development.