Live streaming e-commerce is becoming prevalent and its new business model attracts much attention from the marketing and information systems fields. In this paper, we examine the role of mega-influencers on the live streaming e-commerce platform. Based on a large-scale dataset and a natural experiment, we explore how influencers’ performance is affected by mega influencers. We find that there exist significant impacts from the mega influencers. Our paper illustrates that the treatment effects of the mega influencers continuously affected other influencers. Based on a novel machine learning approach, we show that there exist heterogeneously treatment effects across influencers. Our paper has theoretical contributions to influencer marketing.