A novel susceptible–infected model on graphs was examined for modeling information diffusion in social networks. We assumed that the dynamics of web search queries on a specific topic can be an adequate reflection and proxy approximation for the quantitative representation of the information propagation process in complex social networks. Therefore, we reject the hypothesis of homogeneous mixing and consider the network structure, focusing on information propagation in scale-free networks. We apply the proposed model to analyze the phenomenon of media content propagation in social networks, which is projected onto quantitative characteristics of web search queries.