This paper studies differences in consumers' grocery shopping behavior when they shop online and in a brick-and-mortar store. To do so, I assemble a new scanner dataset that tracks customers' grocery purchases in-store and on the Internet. This allows comparison in behavior of the same households, shopping in the same chain, for identical items and for identical prices, eliminating many possible confounding factors. I document that brand exploration - the purchase of a brand not tried in the past - is systematically more prevalent in-store than online. I propose three possible explanations for this finding: (i) shocks to the instantaneous utility of time correlated with the decision to shop online (ii) possibility to reduce shopping cost; and (iii) difficulty in assessing quality of unknown items while shopping online. I develop a model of consumer behavior that allows me to quantify each effect. I find that all of them contribute to hamper trial of new brands online. The counterfactual shows that altering the design of the website to remove potential obstacles to new trials increases brand exploration by 23%. More generally, in contrast to the conventional wisdom of the Internet reducing entry barriers, my work points to features of the online environment that in certain contexts actually could make entry of new brands more difficult.
Stored only when both papers are in the corpus — 90 citations exist upstream.