Advertising Research - Budget Decision
Advertising Research - Budget Decision
Providing analytical, research based judgments as to the optimal advertising budget is surprisingly difficult. However, there are research inputs that can be helpful. Tracking studies that show the advertising is either surpassing or failing to reach communication objectives can suggest that the budget should either be reduced or increased. Forced exposure testing of multiple exposures can suggest the optimal number of exposures per month for an audience member. Such a number can help guide the establishment of an advertising budget. More direct approaches include regression analysis of internal sales and advertising data, field experimentation, and split-cable experimentation.
Regression Analysis
A direct approach is to use internal data of advertising and sales over time, or perhaps over geographically defined sales regions, or perhaps both. The challenge is to explain sales movements using advertising as an independent variable. The regression coefficient would then be the change in sales obtained from a change in advertising.
Although such a regression analysis is relatively inexpensive and can be useful there are many serious difficulties with it.22 First, there is often little variation in advertising (except that due to seasonal factors). With little variation, it is impossible to detect the impact of changing advertising. Second, it is difficult to control for other marketing variables, competitor actions, and other factors affecting sales, because data describing them are unavailable. Third, if a business spends a certain percentage of sales on advertising, a change in sales could cause a change in advertising, instead of the reverse. Fourth, and most important, it is most difficult to detect the delayed or carry-over impact of advertising on sales.
Field Experimentation
The lack of variation in advertising can be remedied by a field experiment
in which advertising is deliberately increased or decreased in certain areas. The subsequent sales are then monitored. The Budweiser experiments, in the mid-1960s, discussed in Chapter 10, led to dramatic reductions in the Budweiser advertising budget.
22David A. Aaker and James M. Carman, "Are You Overadvertising?" Journal of Advertising Research, 22 (August/September 1982), 57-70.
Field experiments have problems of their own. First, it is virtually impossible to control for the other factors, particularly competitor factors, that can affect sales. In fact, competitors, in order to protect their share or distort a test, may engage in extraordinary actions. Second, field experiments are costly to run, partly because of the costs of deliberately over- or underadvertising in some areas. Third, reliable results are still not guaranteed because of the variability between markets. Fourth, because of the elapsed time of the test the results could be rendered obsolete before they can influence decisions. Finally, field experiments are conducted in view of competitors who are free to observe and monitor them.
Split-Cable Testing
The scanner-based split-cable tests solve many of the problems of field experimentation, and provide an excellent option for those products using television ads and distributed through markets that are monitored by the scanner systems. The advertising budget tests can be tightly controlled in that matched households can receive different levels of advertising. It is also hidden from the view of competitors. Although it can cost from $150,000 and up, and can require six months to a year or more to read because of potential carry-over effects, it can provide a valid estimate of the impact of different levels of advertising.
Hundreds of budget tests have been run on split-cable systems during the past decade, but few tested have reduced advertising levels. Aaker and Carman suggest that the relevant decision makers might be biased against such "down-weight" tests.23 Agencies basically get paid a commission on advertising and thus have a conflict of interest. Further, it is psychologically and logically inconsistent to be excited and positive about a new campaign and propose a budget reduction. Finally, the product manager who controls the advertising and is measured in part on short-run sales, views advertising as a way to protect short-term market position and react to competitor activities.
Providing analytical, research based judgments as to the optimal advertising budget is surprisingly difficult. However, there are research inputs that can be helpful. Tracking studies that show the advertising is either surpassing or failing to reach communication objectives can suggest that the budget should either be reduced or increased. Forced exposure testing of multiple exposures can suggest the optimal number of exposures per month for an audience member. Such a number can help guide the establishment of an advertising budget. More direct approaches include regression analysis of internal sales and advertising data, field experimentation, and split-cable experimentation.
Regression Analysis
A direct approach is to use internal data of advertising and sales over time, or perhaps over geographically defined sales regions, or perhaps both. The challenge is to explain sales movements using advertising as an independent variable. The regression coefficient would then be the change in sales obtained from a change in advertising.
Although such a regression analysis is relatively inexpensive and can be useful there are many serious difficulties with it.22 First, there is often little variation in advertising (except that due to seasonal factors). With little variation, it is impossible to detect the impact of changing advertising. Second, it is difficult to control for other marketing variables, competitor actions, and other factors affecting sales, because data describing them are unavailable. Third, if a business spends a certain percentage of sales on advertising, a change in sales could cause a change in advertising, instead of the reverse. Fourth, and most important, it is most difficult to detect the delayed or carry-over impact of advertising on sales.
Field Experimentation
The lack of variation in advertising can be remedied by a field experiment
in which advertising is deliberately increased or decreased in certain areas. The subsequent sales are then monitored. The Budweiser experiments, in the mid-1960s, discussed in Chapter 10, led to dramatic reductions in the Budweiser advertising budget.
22David A. Aaker and James M. Carman, "Are You Overadvertising?" Journal of Advertising Research, 22 (August/September 1982), 57-70.
Field experiments have problems of their own. First, it is virtually impossible to control for the other factors, particularly competitor factors, that can affect sales. In fact, competitors, in order to protect their share or distort a test, may engage in extraordinary actions. Second, field experiments are costly to run, partly because of the costs of deliberately over- or underadvertising in some areas. Third, reliable results are still not guaranteed because of the variability between markets. Fourth, because of the elapsed time of the test the results could be rendered obsolete before they can influence decisions. Finally, field experiments are conducted in view of competitors who are free to observe and monitor them.
Split-Cable Testing
The scanner-based split-cable tests solve many of the problems of field experimentation, and provide an excellent option for those products using television ads and distributed through markets that are monitored by the scanner systems. The advertising budget tests can be tightly controlled in that matched households can receive different levels of advertising. It is also hidden from the view of competitors. Although it can cost from $150,000 and up, and can require six months to a year or more to read because of potential carry-over effects, it can provide a valid estimate of the impact of different levels of advertising.
Hundreds of budget tests have been run on split-cable systems during the past decade, but few tested have reduced advertising levels. Aaker and Carman suggest that the relevant decision makers might be biased against such "down-weight" tests.23 Agencies basically get paid a commission on advertising and thus have a conflict of interest. Further, it is psychologically and logically inconsistent to be excited and positive about a new campaign and propose a budget reduction. Finally, the product manager who controls the advertising and is measured in part on short-run sales, views advertising as a way to protect short-term market position and react to competitor activities.
Comments
Post a Comment