Unstable diffusion in social networks

How and to what extent will new activities spread through social ties? Here, we develop a more sophisticated framework than the standard mean-field approach to describe the diffusion dynamics of multiple activities on complex networks. We show that the diffusion of multiple activities follows a saddle path and can be highly unstable. In particular, when the two activities are sufficiently substitutable, either of them would dominate the other by chance even if they are equally attractive ex ante. When such symmetry-breaking occurs, any average-based approach cannot correctly calculate the Nash equilibrium - the steady state of an actual diffusion process.