On the Coordination of Value-Maximizing Bidders

TL;DR

The paper investigates coordination among value-maximizing auto-bidders in repeated second-price auctions, motivated by scenarios where multiple advertisers are managed by a single agent or a single advertiser controls many campaigns. It proposes a simple coordination scheme in which only the highest-value coalition member bids against the outside market, while others abstain, and formally analyzes value maximization under a $RoS$ constraint. The authors prove that coordination improves per-bidder utility and $RoS$ compliance under mild assumptions, and that the coalition achieves higher total value than independent bidding when using a state-of-the-art mirror-descent bidding algorithm, especially as the time horizon grows. Empirical validation on synthetic and real-world data corroborates the theory, showing consistent gains in both utility and value and convergence of bid multipliers toward zero, indicating robust coordination benefits in auto-bidding environments.

Abstract

While the auto-bidding literature predominantly considers independent bidding, we investigate the coordination problem among multiple auto-bidders in online advertising platforms. Two motivating scenarios are: collaborative bidding among multiple distinct bidders managed by a third-party bidding agent, and strategic bid selection for multiple ad campaigns managed by a single advertiser. We formalize this coordination problem as a theoretical model and demonstrate that a straightforward coordination mechanism, where only the highest-value bidder competes with outside bids, strictly dominates independent bidding, improving both Return-on-Spend (RoS) compliance and the total value accrued for each participating auto-bidder or ad campaign. Additionally, our simulations on synthetic and real-world datasets support the theoretical result that coordinated mechanism outperforms independent bidding. These findings highlight both the theoretical potential and the practical robustness of coordination in auto-bidding in online auctions.