Location

260-055, Owen G. Glenn Building

Start Date

12-15-2014

Description

We study information revelation policies for ad exchanges that use a second-price auction and the Two-Call design, which corresponds to the Ad Served on the Win Notice OpenRTB specification in the ad industry. Here, the exchange makes a call to all bidders at the beginning of an auction, and makes a second call to the winning bidder at the end of the auction; this second call enables the winning bidder to match the right advertiser for the impression. While valuations are private to bidders, there are two possibilities as far as the information available to the ad exchange is concerned: When the exchange has no reliable knowledge about bidder valuations, we develop simple information revelation policies that do not use any knowledge of the valuations and establish their performance guarantees. When the exchange has distributional knowledge about bidder valuations, we develop an effective informed heuristic that exploits this information.

Share

COinS
 
Dec 15th, 12:00 AM

Making a Good Impression: Information Revelation Policies in Ad Exchanges

260-055, Owen G. Glenn Building

We study information revelation policies for ad exchanges that use a second-price auction and the Two-Call design, which corresponds to the Ad Served on the Win Notice OpenRTB specification in the ad industry. Here, the exchange makes a call to all bidders at the beginning of an auction, and makes a second call to the winning bidder at the end of the auction; this second call enables the winning bidder to match the right advertiser for the impression. While valuations are private to bidders, there are two possibilities as far as the information available to the ad exchange is concerned: When the exchange has no reliable knowledge about bidder valuations, we develop simple information revelation policies that do not use any knowledge of the valuations and establish their performance guarantees. When the exchange has distributional knowledge about bidder valuations, we develop an effective informed heuristic that exploits this information.