Uğurcan Dündar is a Research and Teaching Associate at the Institute for Interactive Marketing & Social Media (IMSM). He presented his project, which focusses on budget allocations of online advertisers throughout the day aka pacing.

Campaign management systems offer readily available pacing heuristics or provides the infrastructure for advertisers to implement their own heuristics. His project compares the performance of the default pacing heuristic (even pacing) and the widely implemented heuristic (water level pacing) to profit-maximizing pacing heuristic, which theoretically maximizes profit. In order to investigate the differences in the deviation from profit-maximization between these heuristics, he conducted two types of analysis. First, an empirical counterfactual analysis investigated the effects between three different industries (online dating, direct banking and industrial goods). Second, in a simulation study he varied prediction errors to find under which circumstances heuristics perform better or worse. While the current industry default - even pacing - shows the largest deviations from profit maximization throughout all industries, waterlevel pacing (allocation of the budget according to traffic) wins under all circumstances and is only beaten by PMP uncertainty (profit maximizing pacing) for industrial goods or low levels of prediction error. This can be explained by the larger variation in CTR throughout the day, as it is the case for industrial goods. His findings suggest that online platforms should consider implementing waterlevel pacing and perhaps even set it as a default.

His project is the first one that includes all of these heuristics within an empirical and simulation study and therefore delivers valuable insights for online platforms and advertisers. Thank you Uğurcan for your work and your talk!