Measuring the Cumulative Effects of Advertising on Sales: A Response to Peles

A year or two ago a colleague introduced me to a lawyer at the FTC as an expert on the duration of advertising effects. The lawyer's direct and utterly justifiable response was, Oh, how interesting. How long do they last? To my chagrin, I realized that a few years of involvement in cumulative advertising studies had given me no basis upon which to answer that most important question. My attempt to obtain a substantive, broadly based answer to that question began the research which led to the survey published in the JMR in November 1976. I think if one understands the genesis of the study, then some choices I made become more easily understood and some motives attributed to me by commentators are recognized as inappropriate . Because I began the survey with the intention of answering a reasonable real world question, I was forced to set some priorities. First, I wanted a number for an answer to the question rather than a philosophical skirting of the point. Second, to be useful in a real-world sense, that numerical answer had to be based on as broad an experience base as possible. Third, my interest in the question was primarily substantive and not methodological. As a result of these priorities I was forced to take the most broadly used model of cumulative advertising effect that I could find. Anyone who tries to accomplish goals similar to these would be forced to choose the Koyck model exactly as I was.