Essays on Optimal Bidding Strategies in Sponsored Search Advertising Auctions PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Essays on Optimal Bidding Strategies in Sponsored Search Advertising Auctions PDF full book. Access full book title Essays on Optimal Bidding Strategies in Sponsored Search Advertising Auctions by Prabirendra Chatterjee. Download full books in PDF and EPUB format.
Author: Prabirendra Chatterjee
Publisher:
ISBN:
Category : Internet advertising
Languages : en
Pages : 102
Get Book Here
Book Description
In this dissertation, I model generalized second price (GSP) auction for keyword search to analyze the optimal bidding strategies of the participating advertisers. The results also apply to a more general setting where multiple goods are being auctioned off. The study in chapter 3 examines the bidding strategies of the advertisers in a complete information static GSP auction. The results show that unlike in standard second price auction, truthful bidding is never a dominant strategy in general second price auction. In chapter 4, I have developed a model of static incomplete information GSP auction. I characterize all possible pure strategy Bayes-Nash equilibrium of the game and show that the consideration of the click through rates ratio plays a key role in determining the equilibrium bidding strategies for the advertisers. Specifically, I find that when the click through rates ratio exceeds a critical value, there will be no pure strategy Bayes-Nash equilibrium. The analysis also reveals that in a game of static incomplete information no asymmetric bidding equilibrium would prevail. The study in chapter 5 analyzes a model of incomplete information dynamic GSP auction. I find that in a dynamic game, the existence of both separating strategy equilibrium and pooling strategy equilibrium would depend upon critical values of click through rates ratio. I also prove that the advertisers with high valuation for a keyword will either reveal their identities at the very beginning or at the very end of this dynamic game. The results also show that when search engines do not publish the bidding history (i.e. there is 'minimum disclosure of information'), the advertisers will never try to mimic each other or in other words, there will be no pooling strategy equilibrium.
Author: Prabirendra Chatterjee
Publisher:
ISBN:
Category : Internet advertising
Languages : en
Pages : 102
Get Book Here
Book Description
In this dissertation, I model generalized second price (GSP) auction for keyword search to analyze the optimal bidding strategies of the participating advertisers. The results also apply to a more general setting where multiple goods are being auctioned off. The study in chapter 3 examines the bidding strategies of the advertisers in a complete information static GSP auction. The results show that unlike in standard second price auction, truthful bidding is never a dominant strategy in general second price auction. In chapter 4, I have developed a model of static incomplete information GSP auction. I characterize all possible pure strategy Bayes-Nash equilibrium of the game and show that the consideration of the click through rates ratio plays a key role in determining the equilibrium bidding strategies for the advertisers. Specifically, I find that when the click through rates ratio exceeds a critical value, there will be no pure strategy Bayes-Nash equilibrium. The analysis also reveals that in a game of static incomplete information no asymmetric bidding equilibrium would prevail. The study in chapter 5 analyzes a model of incomplete information dynamic GSP auction. I find that in a dynamic game, the existence of both separating strategy equilibrium and pooling strategy equilibrium would depend upon critical values of click through rates ratio. I also prove that the advertisers with high valuation for a keyword will either reveal their identities at the very beginning or at the very end of this dynamic game. The results also show that when search engines do not publish the bidding history (i.e. there is 'minimum disclosure of information'), the advertisers will never try to mimic each other or in other words, there will be no pooling strategy equilibrium.
Author: Song Yao
Publisher: Now Publishers Inc
ISBN: 1601982283
Category : Brand loyalty
Languages : en
Pages : 66
Get Book Here
Book Description
Sponsored Search Auctions reviews current academic research on this nascent topic with a focus on future practical and research opportunities
Author: Emmanuel Lorenzon
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
This thesis is a collection of three essays in theoretical auction analysis. Chapter 1 considersbid delegation in the GSP auction mechanism. In a game involving side-contracts and a compensationpolicy set by an agency, the first-best collusive outcome is achieved. We offer a characterization of the implementablebid profiles for the two-position game with three players. Chapter 2 considers the sequentialsale of an object to two buyers: one knows his private information and the other buyer does not. Buyershave a multi-unit demand and private valuations for each unit are perfectly correlated. An asymmetricequilibrium exists when the uninformed player adopts an aggressive bidding strategy. Conversely, hisinformed opponent behaves more conservatively by using bid shading. The bidding behaviour of theuninformed bidder is driven by the opportunity to learn his private valuation for free. This dynamic is atthe root of the decline in the equilibrium price across both sales. In chapter 3, information is observableduring the first-stage auction in a sequential-move game in which the first-mover bidder is observed byhis opponent. A separating equilibrium exists in which the informed bidder bids aggressively when he isthe first-mover which entails a non-participation strategy from his uninformed competitor. Conversely,the latter adopts a conservative behaviour when he is the first-mover. A pooling equilibrium in which theinformed bidder blurs his valuation can only exist if his uninformed opponent adopts a non-participatingstrategy.
Author: Jianqing Chen
Publisher:
ISBN:
Category : Internet advertising
Languages : en
Pages : 298
Get Book Here
Book Description
Advances in information technology have created radically new business models, most notably the integration of advertising with keyword-based targeting, or "keyword advertising." Keyword advertising has two main variations: advertising based on keywords employed by users in search engines, often known as "sponsored links," and advertising based on keywords embedded in the content users view, often known as "contextual advertising." Keyword advertising providers such as Google and Yahoo! use auctions to allocate advertising slots. This dissertation examines the design of keyword auctions. It consists of three essays. The first essay "Ex-Ante Information and the Design of Keyword Auctions" focuses on how to incorporate available information into auction design. In our keyword auction model, advertisers bid their willingness-to-pay per click on their advertisements, and the advertising provider can weigh advertisers' bids differently and require different minimum bids based on advertisers' click-generating potential. We study the impact and design of such weighting schemes and minimum-bids policies. We find that weighting scheme determines how advertisers with different click-generating potential match in equilibrium. Minimum bids exclude low-valuation advertisers and at the same time may distort the equilibrium matching. The efficient design of keyword auctions requires weighting advertisers' bids by their expected click-through-rates, and requires the same minimum weighted bids. The revenue-maximizing weighting scheme may or may not favor advertisers with low click-generating potential. The revenue-maximizing minimum-bid policy differs from those prescribed in the standard auction design literature. Keyword auctions that employ the revenue-maximizing weighting scheme and differentiated minimum bid policy can generate higher revenue than standard fixed-payment auctions. The dynamics of bidders' performance is examined in the second essay, "Keyword Auctions, Unit-price Contracts, and the Role of Commitment." We extend earlier static models by allowing bidders with lower performance levels to improve their performance at a certain cost. We examine the impact of the weighting scheme on overall bidder performance, the auction efficiency, and the auctioneer's revenue, and derive the revenue-maximizing and efficient policy accordingly. Moreover, the possible upgrade in bidders' performance levels gives the auctioneer an incentive to modify the auction rules over time, as is confirmed by the practice of Yahoo! And Google. We thus compare the auctioneer's revenue-maximizing policies when she is fully committed to the auction rule and when not, and show that she should give less preferential treatment to low-performance advertisers when she is fully committed. In the third essay, "How to Slice the Pie? Optimal Share Structure Design in Keyword Auctions," we study the design of share structures in keyword auctions. Auctions for keyword advertising resources can be viewed as share auctions in which the highest bidder gets the largest share, the second highest bidder gets the second largest share, and so on. A share structure problem arises in such a setting regarding how much resources to set aside for the highest bidder, for the second highest bidder, etc. We address this problem under a general specification and derive implications on how the optimal share structure should change with bidders' price elasticity of demand for exposure, their valuation distribution, total resources, and minimum bids.
Author: Nian Si
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
The digital advertising industry heavily relies on online auctions, which are mostly of first-price type. For first-price auctions, the success of a good bidding algorithm crucially relies on accurately estimating the highest bid distribution based on historical data which is often censored. In practice, a sequence of first-price auctions often takes place through multiple layers, a feature that has been ignored in the literature on data-driven optimal bidding strategies. In this paper, we introduce a two-step algorithmic procedure specifically for this multi-layer first-price auction structure. Furthermore, to examine the performance of our procedure, we develop a novel inference scheme for A/B testing under budget interference (an experimental design which is also often used in practice). Our inference methodology uses a weighted local linear regression estimation to control for the interference incurred by the amount of spending in control/test groups given the budget constraint. Our bidding algorithm has been deployed in a major demand-side platform in the United States. Moreover, in such an industrial environment, our tests show that our bidding algorithm outperforms the benchmark algorithm by 0.5% to 1.5%.
Author: Yanwu Yang
Publisher: Elsevier
ISBN: 0124115047
Category : Computers
Languages : en
Pages : 213
Get Book Here
Book Description
The Intelligent Systems Series publishes reference works and handbooks in three core sub-topic areas: Intelligent Automation, Intelligent Transportation Systems, and Intelligent Computing. They include theoretical studies, design methods, and real-world implementations and applications. The series' readership is broad, but focuses on engineering, electronics, and computer science. Budget constraints and optimization in sponsored search auctions takes into account consideration of the entire life cycle of campaigns for researchers and developers working on search systems and ROI maximization. The highly experienced authors compiled their knowledge and experience to provide insight, algorithms and development techniques for successful optimized/constrained systems. The book presents a cutting-edge budget optimization approach that embraces three-level budget decisions in the life cycle of search auctions: allocation across markets at the system level, distribution over temporal slots at the campaign level, and real-time adjustment at the keyword level. Delivers a systematic overview and technique for understanding budget constraints and ROI optimization in sponsored search auction systems, including algorithms and developer guides for a range of scenarios Explores effects of constraints on mechanisms, bidding and keyword strategies, and the strategies for budget optimization that developers can employ An informative reference source for both software and systems developers working in the search auctions, marketing and sales strategy optimization, services development for online marketing and advertisement, e-commerce, social and economic networking
Author: Oliver J. Rutz
Publisher:
ISBN:
Category : Internet advertising
Languages : en
Pages : 356
Get Book Here
Book Description
Author: Kinshuk Jerath
Publisher:
ISBN:
Category :
Languages : en
Pages : 41
Get Book Here
Book Description
As sponsored search becomes increasingly important as an advertising medium for firms, search engines are exploring more advanced bidding and ranking mechanisms to increase their revenue from auctions for sponsored search advertising. For instance, Google, Yahoo! and Bing are experimenting with auction mechanisms in which each advertiser can bid to be displayed exclusively. In such an auction, each advertiser submits two bids: one bid for the standard display format in which multiple advertisers are displayed, and one bid for being shown exclusively. The search engine decides the outcome as non-exclusive or exclusive based on revenue generated -- if the exclusive-placement bid by an advertiser is high enough then only that advertiser is displayed, otherwise multiple advertisers are displayed and ranked based on their multiple-placement bids. We find that allowing advertisers to bid for exclusivity has two countervailing effects for search engine revenue. First, competition is heightened because bidders can express their display preferences more flexibly and compete not only for positions in the non-exclusive outcome but also compete for the outcome to be exclusive or non-exclusive; this is good for search engine revenue. Second, competition between non-exclusive and exclusive outcomes gives bidders the incentive to reduce their bids for their non- preferred outcome; this is bad for search engine revenue. Under different parametric conditions, either force can prevail and search engine revenue may increase or decrease accordingly. We also find counterintuitive results on the bidding strategies of advertisers; for instance, we find that, under certain conditions, advertisers have the incentive to bid above their true valuations.
Author: Jun Zhang
Publisher:
ISBN:
Category :
Languages : en
Pages : 122
Get Book Here
Book Description
This thesis studies issues in auctions and contests. The seller of an object and the organizer of a contest have many instruments to improve the revenue of the auction or the efficiency of the contest. The three essays in this dissertation shed light on these issues. Chapter 2 investigates how a refund policy affects a buyer's strategic behavior by characterizing the equilibria of a second-price auction with a linear refund policy. I find that a generous refund policy induces buyers to bid aggressively. I also examine the optimal mechanism design problem when buyers only have private initial estimates of their valuations and may privately learn of shocks that affect their valuations later. When all buyers are \emph{ex-ante} symmetric, this optimal selling mechanism can be implemented by a first-price or second-price auction with a refund policy. Chapter 3 investigates how information revelation rules affect the existence and the efficiency of equilibria in two-round elimination contests. I establish that there exists no symmetric separating equilibrium under the full revelation rule and find that the non-existence result is very robust. I then characterize a partially efficient separating equilibrium under the partial revelation rule when players' valuations are uniformly distributed. I finally investigate the no revelation rule and find that it is both most efficient and optimal in maximizing the total efforts from the contestants. Within my framework, more information revelation leads to less efficient outcomes. Chapter 4 analyzes the signaling effect of bidding in a two-round elimination contest. Before the final round, bids in the preliminary round are revealed and act as signals of the contestants' private valuations. Compared to the benchmark model, in which private valuations are revealed automatically before the final round and thus no signaling of bids takes place, I find that strong contestants bluff and weak contestants sandbag. In a separating equilibrium, bids in the preliminary round fully reveal the contestants' private valuations. However, this signaling effect makes the equilibrium bidding strategy in the preliminary round steeper for high valuations and flatter for low valuations compared to the benchmark model.
Author: Peerapong Dhangwatnotai
Publisher:
ISBN:
Category :
Languages : en
Pages :
Get Book Here
Book Description
In this dissertation, we design and analyze auctions that are more practical than those in the traditional auction theory in several settings. The first setting is the search advertising market, in which the multi-keyword sponsored search mechanism is the dominant platform. In this setting, a search engine sells impressions generated from various search terms to advertisers. The main challenge is the sheer diversity of the items for sale -- the number of distinct items that an advertiser wants is so large that he cannot possibly communicate all of them to the search engine. To alleviate this communication problem, the search engine introduces a bidding language called broad match. It allows an advertiser to submit a single bid for multiple items at once. Popular models such as the GSP auction do not capture this aspect of sponsored search. We propose a model, named the broad match mechanism, for the sponsored search platform with broad match keywords. The analysis of the broad match mechanism produces many insights into the performance of the sponsored search platform. First, we identify two properties of the broad match mechanism, namely expressiveness and homogeneity, that characterize the performance of the mechanism. Second, we show that, unlike the GSP auction, the broad match mechanism does not necessarily have a pure equilibrium. Third, we analyze two variants of the broad match mechanism, the pay-per-impression variant and the pay-per-click variant. Under a common model of advertiser valuation, we show that the pay-per-click variant is more economically efficient than the pay-per-impression variant. This result justifies the prevalent use of the pay-per-click scheme in search advertising. In addition, the broad match mechanism can be viewed as an auction of which the bidding language is crucial to its performance. In the second part, we design and analyze approximately revenue-maximizing auctions in single-parameter settings. Bidders have publicly observable attributes and we assume that the valuations of bidders with the same attribute are independent draws from a common distribution. Previous works in revenue-maximizing auctions assume that the auctioneer knows the distributions from which the bidder valuations are drawn \cite{M81}. In this dissertation, we assume that the distributions are a priori unknown to the auctioneer. We show that a simple auction which does not require any knowledge of the distributions can obtain revenue comparable to what could be obtained if the auctioneer had the distributional knowledge in advance. Our most general auction has expected revenue at least a constant fraction of that of the optimal distributional-dependent auction in two settings. The first setting concerns arbitrary downward-closed single-parameter environments and valuation distributions that satisfy a standard hazard rate condition, called monotone hazard rate. In this setting, the expected revenue of our auction is improved to a constant fraction of the expected optimal welfare. In the second setting, we allow a more general class of valuation distributions, called regular distributions, but require a more restrictive environment called the matroid environment. In our results, we assume that no bidder has a unique attribute value, which is obviously necessary with unknown and attribute-dependent valuation distributions. Our auction sets a reserve price for a bidder using the valuation of another bidder who has the same attribute. Conceptually, our analysis shows that even a single sample from a distribution -- another bidder's valuation -- is sufficient information to obtain near-optimal expected revenue, even in quite general settings. In the third part, we design and analyze auctions that approximately maximize residual surplus in single-parameter settings. Residual surplus is defined to be the surplus less the sum of the bidders' payments. The guarantee of our auction is of the same type as the auctions in the second part, i.e., its expected residual surplus is a fraction of that of the optimal distributional-dependent auction. Instead of the no-unique-attribute assumption made in the second setting, in this setting we assume that the distributions of bidder valuations can be ordered, that is, the distribution of the first bidder stochastically dominates that of the second bidder and the distribution of the second bidder stochastically dominates that of the third and so on. In every downward-closed stochastic-dominance environment where the distributions of bidder valuations satisfy the monotone hazard rate condition, our auction produces residual surplus that is a $\Omega(\tfrac{1}{\log n})$ fraction of the optimal residual surplus, without taking any bid (although it makes use of the ordering), where $n$ is the number of bidders.
