Social media platforms serve as important tools for diffusing knowledge within organizations. The factors affecting knowledge diffusion through social media networks (SMNs) need to therefore be better understood. Accordingly, this paper focuses on two SMN‐specific characteristics – reputation mechanisms and the distribution of knowledge roles – which are argued to enhance and enable, respectively, the smooth transfer of knowledge in a SMN. To examine their effects, we distinguish between two types of reputation mechanisms – adaptive and objective – and across three distinct knowledge roles in SMNs: seekers, contributors and brokers. We argue that the extent of knowledge diffusion in the SMN depends on the type of mechanism and the relative distribution of these three roles. Using data collected through an agent‐based simulation, we find that (a) the distribution of knowledge roles affects knowledge diffusion, with distributions consisted of more brokers outperforming others and (b) objective reputation mechanisms outperform adaptive mechanisms. Furthermore, we find that reputation mechanisms and distribution of knowledge roles interact to influence knowledge diffusion. The study's implications for future research and practice are discussed in the light of its limitations. © 2016 John Wiley & Sons Ltd