The rise of interactive platforms like conversational AI assistants introduces a new dimension to ad auctions: time. As a conversation evolves, a platform can learn more about a user's interests, creating a fundamental trade-off: delay ad delivery to acquire precise information about ad quality, or deliver the ad immediately to capitalize on a thick market before potential bidders are revealed to be poor matches. We study this trade-off in a model where a revenue-maximizing platform can commit to an auction format (first- or second-price) but not to its timing, endogenously choosing when to run the auction after observing its beliefs about advertisers' ad quality.

We study the design of credible auctions when a seller has private information about her costs and cannot commit to public announcements. Akbarpour and Li [2020] show that when the reserve price is common knowledge, the first-price auction is the unique mechanism that is simultaneously optimal, static, and credible. Our paper demonstrates that this conclusion is overturned when the seller is privately informed about her cost, a common feature in many real-world markets.

In this paper, we initiate the study of autobidding where the signals for each bidder can be noisy and correlated. Our first set of results showcases the failure of traditional auctions such as the second-price auction (SPA) and the first-price auction (FPA). In particular, uniform bidding is not an optimal bidding strategy for SPA and both SPA and FPA can have arbitrarily poor efficiency. To circumvent this, we propose the Contextual Second Price Auction (CSPA), a novel mechanism which mitigates the aforementioned adverse effects by leveraging multiple signals to adjust the allocation of SPA. We show that uniform bidding is an optimal bidding strategy in CSPA and we prove a tight bound on the price for anarchy for CSPA of 2, thus recovering the well-established results in the independent setting. Finally, we show that CSPA always achieves at least half the welfare of SPA; moreover this is also tight.

How did the launch of Sputnik 1 affect NASA's process of developing an artificial Earth satellite? How do nuclear programs in countries attempting to acquire nuclear weapons compare to the Manhattan Project? Whether agencies attempt an original discovery or redicovery, they undergo a process of search in an unfamiliar environment. They converge upon successful designs through a process of trial and error. However, an agency attempting rediscovery searches with the comfort of knowing that the technology it hopes to reproduce is feasible. Meanwhile, the agency attempting an original discovery has no such guarantee. We study how simply knowing that something is discoverable affects the process of search in unfamiliar environments. Intuitively, the process of rediscovery seems simpler than original discovery, and we crystallize this idea.

Learning algorithms are proliferating in a variety of business contexts, ranging from automated bidding in online auctions to pricing on shopping platforms and setting rents. This diffusion has been accompanied by fears that such automation could facilitate collusion. A number of recent papers on algorithmic pricing show in simulations that learning algorithms coordinate on less-than-competitive outcomes.

Motivated by online advertising auctions, we study auction design in repeated auctions played by simple Artificial Intelligence algorithms (Q-learning). We find that first-price auctions with no additional feedback lead to tacit-collusive outcomes (bids lower than values), while second-price auctions do not. We show that the difference is driven by the incentive in first-price auctions to outbid opponents by just one bid increment. This facilitates re-coordination on low bids after a phase of experimentation. We also show that providing information about the lowest bid to win, as introduced by Google at the time of the switch to first-price auctions, increases competitiveness of auctions.