CONTENTS

Optimal Temporal Policies in Fluid Milk Advertising

Editor's Notes

Distribution of Gains from Commodity Checkoff Programs:
Research vs. Promotion

Director’s Corner

Next Meeting

NEC-63 1999 Fall Meeting September 27-28, 1999 Ottawa, Canada What Can We Learn from How the U.S. and Canada Milk Promotion Boards Evaluate the Many Components of Their Programs?

Printable pdf version

Optimal Temporal Policies in Fluid Milk Advertising

by Philip Vande Kamp and Harry M. Kaiser

Previous studies that have addressed the optimal allocation of generic advertising over time have assumed a symmetric demand response to increases and decreases in advertising (Bockstael, Strand, and Lipton; Liu and Forker; Kinnucan and Forker). Consumers, however, do not necessarily respond at the same pace to increases relative to decreases in advertising. Indeed, in a recent NICPRE Quarterly article, we showed an asymmetric demand response to fluid milk advertising in New York City, i.e., demand decreases slowly when advertising is reduced compared to a relatively rapid expansion in demand when advertising is increased. An important question that naturally arises is: what are the implications of an asymmetric response to advertising for optimal advertising policies? This question has not been previously addressed in the generic commodity promotion literature and has gained only limited attention in the general marketing literature (Simon, Mesak).

The purpose of the research reported here is to determine optimal temporal patterns of advertising given an asymmetric demand response to advertising. To this end, empirical results of an asymmetric advertising-demand relationship for fluid milk are used to develop a dynamic optimization model where advertising is allocated over time so as to maximize the present value of current and future fluid milk sales. The analysis is applied to generic fluid milk advertising in New York City.

Optimal advertising strategies may take several forms. Alternatives include a uniform advertising policy where advertising is the same in all periods, or a pulsing advertising policy where periods of intense advertising are alternated with periods of low or zero advertising. Many variations of pulsing advertising policies exist depending on the shape of the advertising patterns including the intensity and length of the pulses.

We assume generic advertising program managers attempt to maximize the current and future revenue for a commodity by choosing the level of advertising in each period. This can be represented by maximizing the present value of all current and future revenue subject to the amount of funds available for advertising expenditures. A fixed level of funds is assumed to be made available for advertising each month.

These funds, which are earmarked for advertising, can be spent on purchasing advertising in the same month they are received, or can be saved for use in future months. The expenditures on advertising each month are constrained by the funds available for advertising including savings from previous months and an upper bound on advertising. The upper bound on advertising represents the maximum amount of advertising that a promotion manager would purchase in a specific month. For details on the optimization model, see Vande Kamp and Kaiser.

A number of optimizations were performed to determine the optimal advertising policy for fluid milk in New York City and test the sensitivity of the results to various assumptions. The main optimizations were performed with and without an upper bound on monthly advertising. Similar to Bockstael, Strand, and Lipton and as discussed earlier, the upper bound on advertising represents the maximum amount of advertising that can be spent in one month. Based on consultations with fluid milk advertising managers in New York, increasing advertising levels more than 300 percent may result in consumer fatigue from repetitive advertising, especially among target audiences in the television media market. Also large advertising purchases during limited time slots could increase the price of advertising.

The optimal advertising policy was determined for three possible upper bounds on advertising including two times, two and one-half times, and three times the historical average advertising level. To better understand the specific relationship between advertising and demand, results were also obtained when the upper bound on monthly advertising was removed. The optimal advertising policies for these four optimizations were compared to the results from two additional advertising policies. These two include a uniform advertising policy, where advertising was held constant every month, and the actual advertising policy for the period January 1986 to June 1995. It is important to note that the sum of real advertising expenditures over time is the same for all advertising policies.

In the first optimization, when maximum advertising was two times the historical average advertising level, a clear “steady cycle” quickly emerged with the optimal policy of three months of zero advertising followed by three months of advertising at the maximum level. Convergence to this pulsing strategy is relatively short—by the twenty-seventh month, the optimal advertising pattern has already emerged. The second and third optimal advertising policies, exhibited similar pulsing patterns. In the second optimal advertising policy, where the maximum advertising level was $2,128, the stable pattern includes four months of zero advertising followed by two months of $2,128 advertising, followed by one month of adverting at $1,731. The third optimal advertising policy, where the maximum advertising level was $2,553, consisted of two alternating seven-month cycles. The stable pulsing pattern showed four months of zero advertising followed by two months of maximum advertising, followed by one month of advertising at zero or $1,769, depending on the cycle.

In the fourth optimal advertising policy, where the upper bound on monthly advertising was removed, the optimal pulsing pattern consists of five months of zero advertising followed by one month of advertising at $5,138. While this result may not be unrealistic, it should be treated with caution since the pulse of advertising is significantly outside the range of the data, and the error advertising policies. The occasional pulsing that occurs in the actual policy likely causes it to dominate the uniform advertising policy. By using pulsing strategies, demand is shown to be 4.3 to 6.2 percent greater relative to the results of the uniform advertising policy. The optimal pulsing advertising strategies are shown to give a 3.0 to 4.9 percent higher demand compared to the actual advertising policy used in the period January 1986 through June 1995. Again, caution must be applied to the results of the fourth optimal advertising policy since the policy includes advertising significantly outside the sample advertising data. Overall, however, these results suggest that managers of this program should consider a more systematic and pronounced pulsing pattern.

As an application to fluid milk demand in New York City, this study suggests that, while holding total advertising expenditures unchanged, a pulsing advertising policy is significantly more effective at increasing demand than a uniform advertising policy. The optimal advertising policy can be characterized as a six to seven month repeated cycle consisting of several months of zero advertising followed by several months of intense advertising.

References to Optimal Temporal Policies...

Bockstael, Nancy E., Ivar E. Strand, and Douglas W. Lipton. “Pulsed Generic Advertising: The Case of Common Property.” Marine Resource Economics 7(Winter 1992):189-208.

Kinnucan, Henry and Olan D. Forker. “Seasonality in the Consumer Response to Milk Advertising with Implications for Milk Promotion Policy.” Amer. J. Agr. Econ. 68(August 1986):562-571.

Liu, Donald J. and Olan D. Forker “Optimal Control of Generic Fluid Milk Advertising Expenditures.” Amer. J. Agr. Econ. 72(November 1990):1047-1055.

Mesak, Hani I. “An Aggregate Advertising Pulsing Model with Wearout Effects.” Marketing Science. 11(Summer 1992):310-326.

Simon, Herman. “ADPULS: An Advertising Model with Wearout and Pulsation.” Journal of Marketing Research. 19(August 1982):352-63.

Vande Kamp, Philip and Harry M. Kaiser. “Optimal Temporal Policies in Fluid Milk Advertising.” In review at American Journal of Agricultural Economics.