Companions may enhance the enjoyment of information goods, services, and life in general, but how this enhancement varies over activities and social context remains underexplored. Furthermore, we ignore whether consumption makes company better as compared to non‐consumption. To address these questions, we implement a mixed methods approach. First, we validate a duration‐based measure of experienced utility on survey data; and, second, calibrate an econometric model of this utility measure on secondary data from the American Time Use Survey. Results show that relational proximity of companions reduces experienced utility, suggesting relational costs; whereas numerosity (cumulative number of companions by activity and day) may partly compensate for, or even reverse those costs of relating to others. Moreover, this balance of estimates of proximity and numerosity of companions is positive for non‐consumption but negative for consumption activities. Consumption emerges as a sub‐optimal locus for social interaction, with important implications for theory and practice.