We study a dynamic model of technology adoption featuring a network externality: the benefits for users increase with the number of adopters. Such a complementarity gives rise to multiple steady states, and to suboptimal allocations. Users differ in the benefits they get from the technology. If benefits are deterministic, then path of the adoption rate in a network displays a Ratchet effect, with one or more large upward jumps. If the benefits are stochastic, then the model generates slow adoption, as individuals optimally wait for others to adopt before doing so. We cast the problem as a Mean Field Game and characterize the aggregate dynamics using an analytic perturbation method around the steady state(s). The results reveal the local stability properties of different steady states. We analytically solve for the dynamics following a small shock, and numerically solve for the global solution, and the social planner’s optimal subsidy. We apply the theory to the gradual adoption of SINPE Mobile, an electronic means of payment developed by the Central Bank of Costa Rica. We use transaction-level data and a rich set of covariates to document the presence of strategic complementarities in the adoption choice.