In this paper, we link findings from a demographically representative discrete choice experiment (DCE) in eight European countries on the adoption of smart thermostats with an agent-based model (ABM) in a methodologically consistent way. We employ the ABM to simulate the diffusion pattern of smart thermostats until 2030 and to examine the effects of subsidies and recommendations by specific agents. Our findings highlight the importance of allowing for within- and across country heterogeneity in preferences for these policies and for technology attributes such as heating cost savings. Further, social interactions reinforce country differences in technology stock in the starting year of the simulations. We find that subsidies moderately accelerate the diffusion of smart thermostats, but they are less effective in countries with a large stock of smart thermostats in the starting year, strong preferences for heating cost savings, and when smart thermostats lead to a strong reduction in heating costs. For some countries, targeting subsidies at particular socio-economic groups (in our case low-income households) slightly mitigates free-riding effects. Our policy simulations further imply that recommendations by energy providers or by energy experts accelerate the diffusion of smart thermostats compared to recommendations by peers.