Abstract: | In an advertising strategy, it is preferable to optimize a sequence of periodic decisions, rather than optimizing each period's decision separately. Thus, a dynamic technique such as optimal control, which is used in this study, should be employed. Necessary conditions, as inferred from a parsimonious advertising model, are tested using data from the tobacco industry. Parameter estimation involves the use of non-linear regression. The estimates permit trajectories of optimal advertising expenditures and optimal market shares to be constructed for filter cigarettes. When compared with the actual trajectories, the optimizing trajectories exhibit a striking correspondence to reality. Therefore, interesting results can be obtained using a simple model, provided the optimizing method takes into account the dynamic nature of managerial decision-making and provided powerful methods for estimating model parameters are used. |