This is the first paper to use drivers’ micro-level driving data and gasoline purchasing behavior to understand how drivers choose where to stop to purchase gasoline. We model drivers’ decision as a combination of a choice of which stations to consider and which station to purchase from conditional on the consideration set. We find that consumers consider only a subset of stations nearby their route and that their consideration set is at least in part formed based on where they have recently traveled, where they have purchased from in the past, and the wholesale gasoline price level. We use the results to calculate drivers’ implied value of time. Our results show that the density of alternative fueling stations needed to make drivers willing to consider alternative fuel vehicles rather than gasoline stations is substantially lower than earlier estimates would suggest. Finally, we show that information acquisition costs are high in gasoline. Full information on which stations are nearby a driver’s route would increase consumer surplus by 10-35% of total gasoline expenditures.