Social trading platforms are considered to be amongst the major innovations in online trading. The purpose of this article is to analyze the trading activity of traders on social trading networks by taking a behavioral approach. We investigate the factors that influence the irrational part of trading activity derived from the key characteristics of these platforms, i.e. those dealing with social interaction. Our investigation utilizes an extensive set of trading data from two major platforms in Germany to study the trading behavior. We apply a fixed effects two-stage least squares approach to quantify the relationship between trading ctivity and performance and define overconfidence as the part of trading activity that is irrationally motivated and results in negative returns. Our results provide evidence for the negative relationship between overconfidence and return on social trading platforms. The article finds that the number of followers and some platform-specific features significantly affect the trading behavior of the traders. We contribute to literature by exploring how the novel social interaction characteristics of online trading impact trading activity by giving rise to a new dimension of overconfidence. In addition, we evidence that the different frameworks of the platforms motivate heterogenous behavioral responses by the signalers. Finally, we refine existing studies by applying a distinct methodology for modeling overconfidence.