Optimal internet media selection

In this study we develop a method that optimally selects online media vehicles and determines the number of advertising impressions that should be purchased and then served from each chosen website. As a starting point, we apply Danaher's (2007) multivariate negative binomial distribution (MNBD) for predicting online media exposure distributions. The MNBD is used as a component in the broader task of media selection. Rather than simply adapting previous selection methods used in traditional media, we show that the Internet poses some unique challenges. Specifically, online banner ads and other forms of online advertising are sold by methods that differ substantially from the way other media advertising is sold. We use a nonlinear optimization algorithm to solve the optimization problem and derive the optimum online media schedule. Data from an online audience measurement firm and an advertising agency are used to illustrate the speed and accuracy of our method, which is substantially quicker than using complete enumeration.