Revenue management in online advertising

Online advertising is a multibillion-dollar business with a promising revenue increase for the coming years. Web publishers that generate revenues from online advertising face several challenging decisions. They need to decide on how many advertising slots to have on their website, whether to hire a sales force to attract advertisers to post ads on their website or rely on advertising networks, how many impressions to promise to deliver, and how much to charge, etc. Revenue management, in particular pricing, is considered one of the most challenging tasks and currently ad-hoc approaches are frequently used. In this dissertation, we provide systematic approaches for managing revenues in online display advertising. In the first chapter, we consider a web publisher facing uncertain demand from advertisers requesting space on its website, and an uncertain supply of impressions from viewers visiting the website. Formulating the problem as a novel queuing system we show, for example, that the optimal cost-per-impression (CPM) can increase in the number of ads rotated in a slot, which goes against the intuition of supply and demand. In the second chapter, we consider a different pricing scheme, the so-called cost-per-click scheme. Formulating the problem as another novel queuing system, we show that the general heuristic applied by practitioners to convert between the CPC and CPM pricing schemes using the so-called click-through rate (CTR), can be misleading. In the third chapter, we explore the interactions of two web publishers in a competitive setting and provide various interesting insights about their strategic pricing behavior at equilibrium. Lastly, In the fourth chapter, we obtain the optimality conditions for the advertisers' demand process when the demand distribution, instead of being Poisson, follows an arbitrary continuous distribution