English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

The transition of display ad exchanges from second-price auctions (SPA) tofirst-price auctions (FPA) has raised questions about its impact on revenue.Auction theory predicts the revenue equivalence between these two auctionformats. However, display ad auctions are different from standard models inauction theory. First, automated bidding agents cannot easily deriveequilibrium strategies in FPA because information regarding competitors is notreadily available. Second, due to principal-agent problems, bidding agentstypically maximize return-on-investment (ROI), not payoff. The literature onlearning agents for real-time bidding is growing because of the practicalrelevance of this area; most research has found that learning agents do notconverge to an equilibrium. Specifically, research on algorithmic collusion indisplay ad auctions has argued that FPA can induce symmetric Q-learning agentsto tacitly collude, resulting in bids below equilibrium, leading to lowerrevenue compared to the SPA. Whether bids are in equilibrium cannot easily bedetermined from field data since the underlying values of bidders are unknown.In this paper, we draw on analytical modeling and numerical experiments andexplore the convergence behavior of widespread online learning algorithms inboth complete and incomplete information models. Contrary to prior results, weshow that there are no systematic deviations from equilibrium behavior. We alsoexplore the differences in revenue of the FPA and SPA, which have not been donefor utility functions relevant to this domain, such as ROI. We show thatlearning algorithms also converge to equilibrium. Still, revenue equivalencedoes not hold, indicating that collusion may not be the explanation for lowerrevenue with FPA, and the change in auction format might have had substantialand non-obvious consequences for ad exchanges and advertisers.

Revenue in First- and Second-Price Display Advertising Auctions: Understanding Markets with Learning Agents