Functional Insights into Google AdWords

In this work we analyze data from Google AdWords, today's most important platform for online advertisements. There, the success of an online ad campaign depends on (time-global) seasonal factors as well as on (time-local) events, such as the importance of Valentine's Day for an online flower shop. These components are, however, difficult to assess using existing key figures and methods. In order to reveal both components from the data, we build upon the recent advances in the literature on the functional linear regression model with points of impact. Our proposed model contributes a yearly perspective on the click-through rate -- one of the most important measures for evaluating Google AdWords campaigns. For estimating the model parameters, we provide an adjusted estimation algorithm that leads to a significant improvement over the original estimation procedure.