Online advertisements are increasingly becoming an attractive channel for advertising. Pricing for online ads has evolved over time, and different auction-based policies are employed by website publishers to sell online ads. In this paper, we analyze, under different scenarios, the following three auction-based policies commonly used by the publishers: (i) invite bids for an impression of an ad, display the highest bidder’s ad, and charge the winning advertiser the submitted bid (ii) invite bids for a click of an ad, display the ad of the bidder submitting the highest bid for a click, and charge the winner the bid submitted, and (iii) invite bids for a click of an ad, use the bid and the click-through-rate of the ad to compute the expected publisher profit from the ad, display the ad generating the highest profit, and charge the winner the submitted bid. A distinguishing feature of our analysis is that it considers the effect of product-market positioning of sellers on their bids for online ads and the consequent implications for pricing policies. Our analysis provides a number of insights useful to the website publisher and the advertisers. First, interestingly, we find that the bid per impression policy and the bid per click policy that uses the expected profit to determine the winner generate identical outcomes. Second, the advertiser profits are found to be higher when the publisher is capable of targeting the consumers. From the first two results, it appears that advertisers’ angst at pricing based on impressions may be misplaced, and that a more critical determinant of the advertiser’s profitability is the publisher’s ability to target. Third, we find the surprising result that among the two bid per click policies, the one that uses the expected publisher profit can generate lower publisher profits. Fourth, we show that, similar to that in the literature on maximizing channel profits in the traditional markets, non-linear pricing of ads leads the publisher to gain the maximum profits even in this vastly different advertising distribution channel. Fifth, we demonstrate that offline ads and online ads are synergistic.
