Location, Location, Location: Repetition and Proximity Increase Advertising Effectiveness

We examine a large-scale field experiment that measures the effect of online display advertising. Yahoo! Research and an American apparel retailer collaborated on the experiment that tracked the online and offline sales of 3 million retailer customers. We measure the impact of the frequency of advertising exposure using a simple randomized experiment: users in the 'Full' treatment group see the retailer's ads, users in the 'Control' group see unrelated control ads, and users in the 'Half' treatment group see an equal probability mixture of the retailer and control ads. We find strong benefits to repeated exposures among users who see up to 50 ads, with revenues increasing approximately linearly at a rate of 4 cents per exposure. We find especially high ad effectiveness for the retailer's best customers and those who live closest to its brick-and-mortar locations; these findings are consistent with advertising in a Hotelling model of differentiated firms.

First Version: July 2011

This Version: June 2016

Download the manuscript.