People frame their communications to suit their personal agendas, even when they are collaborating. However, most information systems (IS) literature assumes people working together have completely aligned agendas. The assumption of agenda alignment in prior research motivates our examination of how agendas and media capabilities influence communication performance. By relaxing the assumption of complete agenda alignment, we reconceptualize how media capabilities interact with communication processes to influence communication performance. We empirically test our model using a lab study with 712 participants interacting face-to-face or via virtual worlds on tasks with differing degrees of agenda alignment. We used a hierarchical linear modeling approach to test our hypotheses. Our results support the role of media synchronicity for explaining communication performance when agendas are aligned. However, our study shows that relaxing the assumption of agenda alignment changes communication in profound ways. We find that media synchronicity is insufficient for describing how media capabilities influence performance in partially cooperative communication contexts. Our findings reveal the importance of anonymity and communication framing when communicants’ agendas differ.