Sponsored Search Auctions Research Articles

Private Data Manipulation in Sponsored Search Auctions

Private Data Manipulation in Sponsored Search Auctions

A practical multi-objective auction design and optimization framework for sponsored search

We propose a simple, flexible, efficient, and general framework for running large-scale highly-dynamic sponsored search auctions. Our framework aims at exploring optimal tradeoffs among various objectives of three parties: platform, advertisers, and users. We model the optimal tradeoffs as an online linear program problem, which can be addressed by a simple approach based on semi-Lagrangian duality. Experimental results conducted on large-scale real-world traffic show that the proposed framework can significantly improve all objectives considered, as well as the platform's revenue.

Bid Coordination in Sponsored Search Auctions: Detection Methodology and Empirical Analysis*

Bid delegation to specialized intermediaries is common in internet ad auctions. When the same intermediary bids for competing advertisers, its incentive to coordinate client bids might alter the functioning of the auctions. This study develops a methodology to detect bid coordination and presents a strategy to estimate a bound on the search engine revenue losses imposed by bid coordination. When the method is applied to data from auctions held on a major search engine, coordination is detected in 55% of the cases of delegated bidding and the search engine's revenue loss ranges between 5.3% and 10.4%.

Sponsored Search Auction and the Revenue- Maximizing Number of Ads Per Page

Sponsored Search Auction and the Revenue- Maximizing Number of Ads Per Page

Nonparametric Estimation of Sponsored Search Auctions and Impacts of Ad Quality on Search Revenue

Nonparametric Estimation of Sponsored Search Auctions and Impacts of Ad Quality on Search Revenue

Retracted: Design of Sponsored Search Auction Mechanism for Federated Learning Advertising Platform.

[This retracts the article DOI: 10.1155/2022/5787491.].

Implementing optimal outcomes through sequential auctions

AbstractWe study sequential auctions as optimal mechanisms for selling multiple heterogeneous items to unit‐demand buyers. We find that as long as the items can be ordered in decreasing variation of the buyers' values, any combination of static standard auctions in this sequence achieves full efficiency and the constrained optimal revenue, subject to (BIC), (IIR), and the all‐sold constraint. In addition, with increasing threshold types and optimal static reserve prices, any l‐th price auctions in the same sequence can achieve the optimal revenue subject to (BIC) and (IIR). Our results have implications for practical auction design, such as sponsored search auctions.

Design of Sponsored Search Auction Mechanism for Federated Learning Advertising Platform.

Federated learning has demonstrated strong capabilities in terms of addressing concerns related to data islands and privacy protection. However, in real application scenarios, participants in federated learning have difficulty matching. For example, two companies distributed in different regions do not know that the other party also needs federated learning in the case of information asymmetry. Therefore, it is difficult to build alliances. To enable suppliers and consumers to find one or more federated learning objects that are relatively satisfactory in a short time, this paper considers the idea of establishing a federated learning advertising platform, where data transactions need to consider privacy protection. A sponsored search auction mechanism design method is introduced to solve the problem of ranking the presentation order of participant advertisements. Due to the potential malicious bidding problem, which occurs when using the classic sponsored search auction mechanism under the federated learning scenario, this paper proposes a novel federated sponsored search auction mechanism based on the Myerson theorem, improving upon the ranking index used in the classic sponsored search auction mechanism. A large number of experimental results on a simulation data set show that our proposed method can fairly select and rank the data providers participating in the bidding. Compared with other benchmark mechanisms, the malicious bidding rate is significantly decreased. In the long run, the proposed mechanism can encourage more data providers to participate in the federated learning platform, thus continuously promoting the establishment of a federated learning ecosystem.

Dominant strategy truthful, deterministic multi-armed bandit mechanisms with logarithmic regret for sponsored search auctions

Dominant strategy truthful, deterministic multi-armed bandit mechanisms with logarithmic regret for sponsored search auctions

A Strategic Group Analysis of Competitor Behavior in Search Advertising

Firms compete intensely in sponsored search. Their bidding strategies hinge on understanding who competes with whom, how they compete, and how consumers react to competing advertisements. In this context, we investigate how firm competition impacts consumers’ click-through behaviors in search advertising from a strategic group perspective. Using search results from Google and consumers’ clickstream data, we found strong negative externality for competitors within the same strategic group relative to competitors across strategic groups: firms reap fewer click-throughs when an advertisement of another firm from the same strategic group is also displayed in search results, relative to when other displayed advertisers are not from the same group. This indicates that when competitors from the same strategic group are likely to appear in the results of a sponsored search auction, the focal firm would be better off avoiding head-to-head competition in the auction. However, we did not find empirical evidence of such firm behaviors, suggesting myopia or the inability of firms to avoid such competition. We also show that when multiple firms from the same strategic group appear in search results, the closer the focal firm is located to such competing firms, the more click-throughs the firm accrues. This suggests that firms should stay close to their within-group competitors when they compete in the same search auction. Further, our empirical results indicate that firms are indeed doing so. Using another set of data from Google AdWords reports, we show that our findings are also robust to multi-keyword bidding scenarios. These findings represent the first attempt to understand the impact of strategic groups in search advertising and provide interesting implications for advertisers and search engines.

Reinforcement Mechanism Design: With Applications to Dynamic Pricing in Sponsored Search Auctions

In many social systems in which individuals and organizations interact with each other, there can be no easy laws to govern the rules of the environment, and agents' payoffs are often influenced by other agents' actions. We examine such a social system in the setting of sponsored search auctions and tackle the search engine's dynamic pricing problem by combining the tools from both mechanism design and the AI domain. In this setting, the environment not only changes over time, but also behaves strategically. Over repeated interactions with bidders, the search engine can dynamically change the reserve prices and determine the optimal strategy that maximizes the profit. We first train a buyer behavior model, with a real bidding data set from a major search engine, that predicts bids given information disclosed by the search engine and the bidders' performance data from previous rounds. We then formulate the dynamic pricing problem as an MDP and apply a reinforcement-based algorithm that optimizes reserve prices over time. Experiments demonstrate that our model outperforms static optimization strategies including the ones that are currently in use as well as several other dynamic ones.

A 2020 perspective on “Let’s play the search game: Strategic and behavioral properties of online sponsored search auction mechanisms”

Abstract Search engines, blogs, social media, and other high-traffic web sites and mobile apps are increasingly involved in sales through the ads they display, yet it is not clear how these ads should be chosen and paid for in the most optimal way. We describe three different paths we are taking in an effort to make this choice more informed. The first explores the dynamics of consumers’ click decisions as they are determined by the belief consumers hold about ad placements. The second explores the advertisers’ dynamic decision-making process under uncertainty as advertisers are trying to optimize their profits. The third one explores the strategic interactions involving a buyer, sellers (advertisers), and an ad provider (search engine or platform) when there is value uncertainty and the ad provider has better information about product values than either the buyer or the sellers. We are hoping these paths lead to new theoretical models and experimental designs that foster our understanding of this important aspect of online and mobile commerce.

Optimal bundles for sponsored search auctions via bracketing scheme

Sponsored search auction has been recently studied and auctioneer’s revenue is an important consideration in probabilistic single-item second-price auctions. Some papers have analyzed the revenue maximization problem on different methods to bundle contexts. In this paper, we propose a more flexible and natural method which is called the bracketing method.We prove that finding a bracketing scheme that maximizes the auctioneer’s revenue is strongly NP-hard. Then, a heuristic algorithm is given. Experiments on three test cases show that the revenue of the optimal bracketing scheme is very close to the optimal revenue without any bundling constraint, and the heuristic algorithm performs very well. Finally, we consider a simpler model that for each row in the valuation matrix, the non-zero cells have the same value. We prove that the revenue maximization problem with K-anonymous signaling scheme and cardinality constrained signaling scheme in this simpler model are both NP-hard.

Let’s play the search game: Strategic and behavioral properties of sponsored search auction mechanisms

Abstract This paper develops a model of buyer (consumer) and seller (advertiser) strategic behavior in sponsored search auctions. Through its various extensions, the model is able to predict the effect of different ranking and payment policies on auction performance in terms of consumer surplus as well as advertisers’ and search engine profits. The theoretical model demonstrates that 1) payment policies should not have an effect on auction performance, 2) ranking policies that do not include the advertiser’s bid as a ranking criterion should not have an effect on auction performance 3) policies that use advertisers’ bids as the single ranking criterion should be just as good as using advertisers’ bids combined with other criteria over which advertisers have no short-term influence (e.g., relevance and click-through rate), 4) some sort of a coordination mechanism is necessary in order for buyers and sellers to achieve a fully efficient outcome, however, coordination should prove much more difficult when the bid is not used as at least one of the ranking criteria, 5) decreasing the number of sponsored links shown to consumers and displaying product prices should improve coordination and increase the likelihood of a fully efficient outcome. We use economically-motivated human subjects and the methods of experimental economics to test the first two of the model’s predictions. We investigate six different sponsored search auction formats: two different pricing policies (pay-per-transaction and pay-per-click) and three different ranking policies (rank by relevance, rank by click-through rate, & rank by both relevance and click-through rate). As predicted by the model we find no significant effect due to difference in ranking policies. However, contrary to the model’s predictions, we find significant differences in total surplus due to payment policies. We discuss two behavioral effects that could be the reason for these differences.

Towards better models of externalities in sponsored search auctions

Sponsored Search Auctions (SSAs) arguably represent the problem at the intersection of computer science and economics with the deepest applications in real life. Within the realm of SSAs, the study of the effects that showing one ad has on the other ads, a.k.a. externalities in economics, is of utmost importance and has so far attracted the attention of much research. However, even the basic question of modeling the problem has so far escaped a definitive answer. The popular cascade model is arguably too idealized to really describe the phenomenon yet it allows a good comprehension of the problem. Other models, instead, describe the setting more adequately but are too complex to permit a satisfactory theoretical analysis. In this work, we attempt to get the best of both approaches: firstly, we define a number of general mathematical formulations for the problem in the attempt to have a rich description of externalities in SSAs and, secondly, prove a host of results drawing a nearly complete picture about the computational complexity of the problem. We complement these approximability results with some considerations about mechanism design in our context.

Collusion With a Rent-Seeking Agency in Sponsored Search Auctions

Collusion With a Rent-Seeking Agency in Sponsored Search Auctions

A Matching Approach to Factor Scores Based on Online Sponsored Search Auction

How does a search engine company decide which advertisements to display for each query to maximize its revenue? This turns out to be a generalization of the online bipartite matching problem. In this paper, search engines decide the strategy to allocate resources with an advertiser credibility factor under budget constraints. Based on the optimal algorithm, which is the notion of a trade-off revealing LP, this paper remains the competitive ratio as 1-1/e with an advertiser credibility factor. During the ranking in the keywords auctions, the authors calculate factor scores. The new arrival keywords should match to the advertisement which has the maximum factor scores. The amount of spent money, budget constraints, CTR (Click Through Rate) and a credibility factor are added to the trade-off function using in the authors' algorithms. In the long term, users tend to use the search engine with high credibility, which will bring greater revenues. The search engine will attract advertisers to bid on keywords and improve their credibility actively.

Competitive Analysis of Bidding Behavior on Sponsored Search Advertising Markets

Online advertisers bidding in sponsored search auctions through Web search engines are experiencing unprecedentedly fierce competition in recent years, resulting in an obvious upward trend in bids submitted by advertisers. This bid inflation phenomenon poses significant threat to the overall stability and the effectiveness of sponsored search markets. Existing research efforts, however, do not yet provide directly relevant theoretical or managerial insights to help understand this kind of real-world competitive bidding behavior in sponsored search. Our research is targeted at filling in this important research gap. Based on a model of advertisers' rational competitive preference, we propose a novel equilibrium solution concept called the upper bound Nash equilibrium (UBNE), which can be viewed as the upper bound of the output-truthful subset in the NE continuum of sponsored search auctions. We show that the UBNE can better characterize advertisers' competitive bidding behavior than other solution concepts, offering a viable theory driven behavioral explanation for bid inflation. We also show that the UBNE is a stable outcome of competitions among advertisers in repeated game settings and yields the optimal outcome for Web search engines.

Adopting the Cascade Model in Ad Auctions: Efficiency Bounds and Truthful Algorithmic Mechanisms

Sponsored Search Auctions (SSAs) are one of the most successful applications of microeconomic mechanisms, with a revenue of about $72 billion in the US alone in 2016. However, the problem of designing the best economic mechanism for sponsored search auctions is far from being solved, and, given the amount at stake, it is no surprise that it has received growing attention over the past few years. The most common auction mechanism for SSAs is the Generalized Second Price (GSP). However, the GSP is known not to be truthful: the agents participating in the auction might have an incentive to report false values, generating economic inefficiency and suboptimal revenues in turn. Superior, efficient truthful mechanisms, such as the Vickrey-Clarke-Groves (VCG) auction, are well known in the literature. However, while the VCG auction is currently adopted for the strictly related scenario of contextual advertising, e.g., by Google and Facebook, companies are reluctant to extend it to SSAs, fearing prohibitive switching costs. Other than truthfulness, two issues are of paramount importance in designing effective SSAs. First, the choice of the user model; not only does an accurate user model better target ads to users, it also is a critical factor in reducing the inefficiency of the mechanism. Often an antagonist to this, the second issue is the running time of the mechanism, given the performance pressure these mechanisms undertake in real-world applications. In our work, we argue in favor of adopting the VCG mechanism based on the cascade model with ad/position externalities (APDC-VCG). Our study includes both the derivation of inefficiency bounds and the design and the experimental evaluation of exact and approximate algorithms.

Sponsored Search Auction Considering Combinational Bids with Externalities

Sponsored search is a mechanism that shows the appropriate ads according to search queries. The orders and payments of ads are determined by the auction. Recently, the allocations are determined by the click through rate(CTR) and the externalities which give effects to CTR and haven't been considered in existing works. However, the mechanism with externalities has high computational cost. The algorithm which can calculate the approximated solution considering the externalities within the polynomial-time are proposed, however, it assumed that one bidder can propose only a single ad. In this paper, we propose the approximation allocation algorithm that one bidder can offer many ads considering externalities. The proposed algorithm employs the concept of the combinatorial auction in order to consider the combinational bids. In addition, the proposed algorithm can find the approximated allocation by the dynamic programming. Moreover, we prove the computational complexity and the monotonicity of the proposed mechanism, and demonstrate computational costs and efficiency ratios by changing the number of ads, slots and maximum bids. The experimental results show that the proposed algorithm can calculate 0.7-approximation solution even though the full search can't find solutions in the limited times.