CONTENTS

Does Advertising Effectiveness Vary over Time?

Editor’s Notes

Manager’s Viewpoint

Asymmetric Response to Advertising

Director’s Corner

Next Meeting

Printable pdf version

Does Advertising Effectiveness Vary over Time?

by Chanjin Chung and Harry M.Kaiser

Farmers invest almost $1 billion annually for generic advertising and other programs to promote U.S. agricultural products. Since most of this money has been generated from mandatory checkoff programs, evaluating the effectiveness of generic advertising has become an important practice for farmers overseeing the checkoff fund. Although a sizable body of literature exists on the economics of generic advertising, most previous studies have assumed a constant advertising effectiveness for the time period being analyzed. This implies a constant relationship between a dollar invested and advertising effects over the entire period. However, consumers’ past responses to advertising may not be the same as their current responses because peoples’ tastes, demographics, advertising theme and managerial effectiveness, and other economic and social factors have changed. Therefore, economic models that allow varying advertising effectiveness may offer more accurate estimates, as well as provide richer information for program managers.

In this study, we develop an advertising model, which recognizes the presence of varying advertising effectiveness. Moreover, the model can identify the sources of change in advertising effectiveness. Using this model, we address the following questions: (1) does advertising effectiveness vary over time? (2) if so, what causes the change? and (3) what are the most important factors affecting the effectiveness of advertising?

Generic fluid milk advertising in New York City (NYC) is chosen as a case study for applying the general model. This market is of interest because New York state farmers spend almost $11 million a year for fluid milk advertising and more than 60 percent of this money has been invested in the NYC market. Data for milk sales, demographics advertising expenditure, and other milk campaign information (e.g., changes in copy themes) are also available for this market. Monthly data from January 1986 to June 1995 are used, and a thorough description of the data sources is presented in Lenz, Kaiser, and Chung (NICPRE Research Bulletin 97-06).

For the empirical estimation of the NYC fluid milk advertising model, per capita fluid milk consumption is assumed to depend on real milk price, income, fluid milk, advertising, competing product advertising, consumer fat concerns, food expenditures away from home, population under age 20, African American population, Hispanic population, and seasonality variables. Since the model is designed to permit the varying effectiveness of advertising over time, we hypothesize that the advertising effectiveness depends on changes in advertising strategies (e.g., copy theme change) and market environments (e.g., price, competing product advertising, consumers’ health concerns, food expenditures away from home, and age and racial mix). The estimated coefficients of advertising strategy and market environment variables are expected to determine the direction (increasing or decreasing) and magnitude of the changing effectiveness of advertising.

Major objectives of this study were to examine whether the advertising effectiveness varied over time, to identify factors changing generic advertising effectiveness, and to compare the relative importance of these factors. Analyses of the estimated coefficients of advertising strategy and market environment variables led the discussion on these issues. First, the hypothesis that the advertising coefficient was variant over time was tested from the estimated model. Test results indicated that advertising effectiveness was clearly time-variant for the study period. The advertising elasticities (i.e., percentage change in per capita milk sales given a one percent change in advertising expenditure) varied accordingly: 0.072 for the 1985-88 period, 0.038 for the 1989-92 period, and 0.076 for the 1993-95 period. A 10 percent increase in advertising expenditure, for example, translated into a nearly 0.8 percent increase in per capita fluid milk sales for the 1985-88 period, but for the period of 1989 through 1992, a 10 percent increase in advertising expenditure increased only 0.4 percent of per capita fluid milk sales. This is a clear departure from the usual view of constant advertising effectiveness.

Second, in the advertising model, we postulated advertising effectiveness could be determoned by several relevant variables such as economic and market environmental variables. We found the economic and market environmental variables provided significant influence in determining varying advertising effectiveness. Signs of the coefficients indicate the direction of the influence of each variable on advertising effectiveness. Advertising effectiveness revealed positive responses to change in variables such as price, percentage of African American population, percentage of food expenditure away from home, and change in advertising strategies, while showing a negative relationship with competing product advertising, consumers’ fat concerns, and percentages of people under age 20 and Hispanic populations. The results, for example, indicated that as the percentage of African American population grew, advertising effectiveness increased, but as the percentage of Hispanic population grew, the advertising effectiveness decreased.

Although no previous theory or empirical evidence exists in determining signs of these coefficients, we attempt to draw two possible implications from these results. First, when the market situation is unfavorable, advertising becomes more important and effective. For instance, increase in price, percentage of African Americans, and percentage of food expenditures for eating away from home are all unfavorable factors for milk sales. When these factors increase in the NYC market, advertising might play a more significant role in increasing (or even maintaining) milk demand than when the market situation is favorable. When the market is in a favorable situation due to an increase in factors such as the percentage of Hispanics or the population under age 20, however, milk consumption might go up without increasing advertising efforts. Second, heavy drinkers may have been close to the saturation point or at least may not have had room to drink as much milk as light drinkers. Since people under age 20 and Hispanics are already heavy milk drinkers, they might not be able to increase their milk consumption as much as African Americans can. Two exceptional cases from two foregoing implications were signs on coefficients of competing product advertising and consumers’ fat concern variables. Negative signs on coefficients of these variables indicated that as competing product advertising expenditures and consumers’ concerns on dietary fat grew, milk advertising effectiveness started to erode.

Finally, the relative importance of determinants of advertising effectiveness was examined. Results suggested that demographic factors were more important than economic factors and, particularly, the percentage of Hispanics in the population was the most important factor in changing advertising effectiveness.

The results of our study imply that advertising effectiveness does vary over time, and when the market is in an unfavorable or unsaturated condition, advertising generally becomes more important and effective. These particular results should be viewed with caution since the estimation was a case study for the NYC market. Further exploration of these results with different market data will permit a generalization of findings in this study. Despite this caveat, our findings have reasonable interpretations. More importantly, the availability of techniques to understand the dynamic nature of advertising effectiveness opens a potentially important avenue for developing better advertising strategies.

Chanjin Chung and Harry M. Kaiser are research associate and associate professor, respectively, in the Department of Agricultural, Resource, and Managerial Economics, Cornell University.

[ top ]