In several markets, firms compete not for consumer expenditure but consumer attention. We examine user priorities over the allocation of their time, and interpret that behavior in light of policy discussions over universal service, data caps, and related policy topics, such as merger analysis. Specifically, we use extensive microdata on user online choice to characterize the demand for the services offered online, which drives a household’s supply of attention. Our data cover a period of time that saw the introduction of many new and notable sites and new devices on which to access them. In our analysis, we assess “how” households supply their attention along various dimensions, such as their concentration of attention across the universe of sites and the amount of attention expenditure per domain visit. Remarkably, we find no change in “how” households allocated their attention despite drastically changing where they allocated it. Moreover, conditional on total attention expenditure, demographics entirely fail to predict our key measures of attention allocation decisions. We highlight several important implications, for policy and beyond, stemming from the persistence and demographic orthogonality of our novel attention measures.