Numerous real business stories of success are accompanied by impressive advertising campaigns. Huge amount of work has been done with regard to different functions of advertising, however less has been addressing on how advertising spending is determined. This paper uses the dynamic linear model and borrows ideas from classic advertising textbooks, which claims that three main factors, which are previous years’ sales (percentage of previous years’ sales), strategic learning process and competitive advertising, determine advertising expenditures. To investigate this issue, the author empirically tests whether firms are sophisticated and can perform all three determinants when making advertising expenditure decisions. This paper employs the Bayesian paradigm and MCMC (Monte Carlo...[ Read more ]

Numerous real business stories of success are accompanied by impressive advertising campaigns. Huge amount of work has been done with regard to different functions of advertising, however less has been addressing on how advertising spending is determined. This paper uses the dynamic linear model and borrows ideas from classic advertising textbooks, which claims that three main factors, which are previous years’ sales (percentage of previous years’ sales), strategic learning process and competitive advertising, determine advertising expenditures. To investigate this issue, the author empirically tests whether firms are sophisticated and can perform all three determinants when making advertising expenditure decisions. This paper employs the Bayesian paradigm and MCMC (Monte Carlo Markov Chain) techniques to estimate the equation system. The findings indicate that managers do consider the previous advertising performance and previous sales when making current decisions, but on average, they just look back up to two periods to collect information. Moreover, managers differ in their altitudes toward the advertising elasticity, which means the degree of importance to which different firms consider last two periods’ advertising performance can differ largely. But firms averagely have similar altitudes on the determinants of “percentage of previous years’ sales”. Given the limitation of data, the competitive effect is generally not significant and this is because either the author cannot distinguish firms’ competitors among all others in a same industry, or that in a crowed industry small and mid-size firms are insensitive to changes of others’ advertising strategy.

Based on the research experiences above, starting from researching on firms’ advertising behavior to test for and develop the general theorem within this field. My future research in this topic should extend to include all four marketing strategies, namely discounting, advertising, distribution breadth and product activity, to identify some missing dynamic effects in this study.