Most collective action models conceptualize a social or political movement as actions taken by a rather homogeneous group of activists that come together for a common goal and put in the same amount of effort with similar costs. The availability of online traces of protest participation allows us to move beyond such simplifying assumptions and study heterogeneity in contentious politics along various dimensions. In this talk, drawing from data on various cases of contentious politics that differ in scales, geographies, and goals, I will discuss various forms of heterogeneity observed in online social and political movements. First, through the case study of #occupygezi, I will examine heterogeneity in political attitudes of participants and the implications of exposure to such heterogeneity for behavioral changes observed for the protesters. Second, through an examination of Black Lives Matter and Women’s rights movements on Twitter, I investigate the heterogeneity in participation intensity and protester type. Findings highlight the potential for a better understanding and modeling of movements by leveraging data driven methods that identify such heterogeneity.