Generalizing What Is Known About Temporal Aggregation and Advertising Carryover

Abstract

This paper provides a theoretical explanation for the inconsistent findings from previous econometric analyses of aggregated data concerning the duration of advertising carryover. Using this theoretical explanation and the knowledge of a data interval bias, the paper reviews and summarizes the empirical work in the area of sales response modeling, and then after adjusting for aggregation bias, presents support for an empirical generalization that the average advertising duration interval is of brief duration—typically between six and nine months.