Research Selected Survey of Sponsored Search Research at

Research Selected Survey of Sponsored Search Research at Yahoo! Research & the 1 st & 2 nd Workshops on Sponsored Search Auctions David Pennock, Yahoo! Research - New York Contributed slides: K. Asdemir, H. Bhargava, J. Feng, S. Lahaie, M. Schwarz

Sponsored search auctions Space next to search results is sold at auction search “las vegas travel”, Yahoo! “las vegas travel” auction

Outline • Yahoo! Research & microeconomics group • Motivation: Industry facts & figures • Introduction to sponsored search – Brief and biased history – Allocation and pricing: Google vs Yahoo! – Incentives and equilibrium • Selected survey of research at Yahoo! – Mechanism design • Analytic comparison of mechanisms [Lahaie]

Outline • Selected survey of research at Yahoo! – Mechanism Design (cont’d) • Learning click rates: N-armed bandit formulation [Pandey & Olsten] • Simulation I: Static [Feng, Bhargava, Pennock] • Simulation II: Equilibrium [Lahaie, Pennock] – Bidding agent design • Pragmatic robot [Schwarz, Edelman]

Outline • Brief summaries of the 1 st & 2 nd Workshops on Sponsored Search • Yahoo!/O’Reilly Tech Buzz Game • Not covered – Sponsored search: budget optimization, click rate prediction, content match, engine switching, expressive bidding, intelligent match, interactivity, inventory prediction, keyword-advertiser graph clustering/recommendation, long-run effects, pricing, query classification, & more. . . – General ad systems, algorithmic search, machine learning, other mechanism design problems

Research Yahoo! Research • New, growing, world-class researchers in search, machine learning, systems, UI, & microeconomics • Relatively open, connected to academia, yet grounded in real problems • Y!R-NYC in Manhattan: 9 scientists & growing Sub-concentrations: ML & microeconomics • Hiring interns & scientists • Academic outreach, visitors, collaborations Come visit us!

Auctions: 2000 View • Yesterday Going once, … going twice, . . . • “Today” (~2000) – e. Bay: 4 million; 450 k new/day

Auctions: 2000 View • Yesterday • “Today” (~2000)

Auctions: 2000 View • Yesterday • “Today” (~2000)

Auctions: 2006 View • Yesterday • Today – e. Bay – Google / Yahoo! – 200 million/month – 6 billion/month (US)

Auctions: 2006 View • Yesterday • Today

Auctions: 2006 View • Yesterday • Today

Newsweek June 17, 2002 “The United States of EBAY” • In 2001: 170 million transactions worth $9. 3 billion in 18, 000 categories “that together cover virtually the entire universe of human artifacts—Ferraris, Plymouths and Yugos; desk, floor, wall and ceiling lamps; 11 different varieties of pockets watches; contemporary Barbies, vintage Barbies, and replica Barbies. ” • “Since everything that transpires on Ebay is recorded, and most of it is public, the site constitutes a gold mine of data on American tastes and preoccupations. ”

“The United States of Search” • • • 6 billion searches/month 50% of web users search every day 13% of traffic to commercial sites 40% of product searches $5 billion 2005 US ad revenue (41% of US online ads; 2% of all US ads) • Doubling every year four years • Search data: Covers nearly everything that people think about: intensions, desires, diversions, interests, buying habits, . . .

Research Introduction to sponsored search • • What is it? Brief and biased history Allocation and pricing: Google vs Yahoo! Incentives and equilibrium

Sponsored search auctions Space next to search results is sold at auction search “las vegas travel”, Yahoo! “las vegas travel” auction

Sponsored search auctions • Search engines auction off space next to search results, e. g. “digital camera” • Higher bidders get higher placement on screen • Advertisers pay per click: Only pay when users click through to their site; don’t pay for uncliked view (“impression”)

Sponsored search auctions • Sponsored search auctions are dynamic and continuous: In principle a new “auction” clears for each new search query • Prices can change minute to minute; React to external effects, cyclical & non-cyc – “flowers” before Valentines Day – Fantasy football – People browse during day, buy in evening – Vioxx

Example price volatility: Vioxx

Sponsored search today • 2005: ~ $7 billion industry – 2004: ~ $4 B; 2003: ~ $2. 5 B; 2002: ~ $1 B • $5 billion 2005 US ad revenue (41% of US online ads; 2% of all US ads) • Resurgence in web search, web advertising • Online advertising spending still trailing consumer movement online • For many businesses, substitute for e. Bay • Like e. Bay, mini economy of 3 rd party products & services: SEO, SEM

Sponsored Search A Brief & Biased History • Idealab Go. To. com (no relation to Go. com) – Crazy (terrible? ) idea, meant to combat search spam – Search engine “destination” that ranks results based on who is willing to pay the most – With algorithmic SEs out there, who would use it? • Go. To Yahoo! Search Marketing – Team w/ algorithmic SE’s, provide “sponsored results” – Key: For commercial topics (“LV travel”, “digital camera”) actively searched for, people don’t mind (like? ) it – Editorial control, “invisible hand” keep results relevant • Enter Google – Innovative, nimble, fast, effective – Licensed Overture patent (one reason for Y!s ~5% stake in G)

Thanks: S. Lahaie Sponsored Search A Brief & Biased History • Overture introduced the first design in 1997: first price, rank by bid • Google then began running slot • auctions in 2000: second price, rank by revenue (bid * CTR) In 2002, Overture (at this point acquired by Yahoo!) then switched to second-price. Still uses rank by bid; Moving toward rank by revenue

Sponsored Search A Brief & Biased History • In the beginning: – Exact match, rank by bid, pay per click, human editors – Mechanism simple, easy to understand, worked, somewhat ad hoc • Today & tomorrow: – “AI” match, rank by expected revenue (Google), pay per click/impression/conversion, auto editorial, contextual (Ad. Sense, YPN), local, 2 nd price (proxy bid), 3 rd party optimizers, budgeting optimization, exploration exploitation, fraud, collusion, more attributes and expressiveness, more automation, personalization/targeting, better understanding (economists, computer scientists)

Sponsored Search Research A Brief & Biased History • • Weber & Zeng, A model of search intermediaries and paid referrals Bhargava & Feng, Preferential placement in Internet search engines Feng, Bhargava, & Pennock Implementing sponsored search in web search engines: Computational evaluation of alternative mechanisms Feng, Optimal allocation mech’s when bidders’ ranking for objects is common Asdemir, Internet advertising pricing models Asdemir, A theory of bidding in search phrase auctions: Can bidding wars be collusive? Mehta, Saberi, Vazirani, & Vaziran Ad. Words and generalized on-line matching 1 st & 2 nd Workshop on Sponsored Search Auctions at ACM Electronic Commerce Conference

Allocation and pricing • Allocation – Yahoo!: Rank by decreasing bid – Google: Rank by decreasing bid * E[CTR] (Rank by decreasing “revenue”) • Pricing – Pay “next price”: Min price to keep you in current position

Research Yahoo Allocation: Bid Ranking “las vegas travel” auction search “las vegas travel”, Yahoo! pays $2. 95 per click pays $2. 94 pays $1. 02. . . bidder i pays bidi+1+. 01

Research Google Allocation: $ Ranking “las vegas travel” auction x E[CTR] = E[RPS] x E[CTR] = E[RPS]

Research Google Allocation: $ Ranking “las vegas travel” auction search “las vegas travel”, Google Trip. Reservations x . 1 =. 301 pays 3. 01*. 1/. 2+. 01 = 1. 51 per click Expedia x . 2 =. 588 pays 2. 93*. 1/. 1+. 01 = 2. 94 LVGravity. Zone x . 1 = . 293 etc. . . x E[CTR] = E[RPS] pays bidi+1*CTRi+1/CTRi+. 01

Aside: Second price auction (Vickrey auction) • All buyers submit their bids privately • buyer with the highest bid wins; pays the price of the second highest bid Only pays $120 $150 $120 $90 $50

Incentive Compatibility (Truthfulness) • Telling the truth is optimal in second-price (Vickrey) auction • Suppose your value for the item is $100; if you win, your net gain (loss) is $100 - price • If you bid more than $100: – you increase your chances of winning at price >$100 – you do not improve your chance of winning for < $100 • If you bid less than $100: – you reduce your chances of winning at price < $100 – there is no effect on the price you pay if you do win • Dominant optimal strategy: bid $100 – Key: the price you pay is out of your control • Vickrey’s Nobel Prize due in large part to this result

Vickrey-Clark-Groves (VCG) • Generalization of 2 nd price auction • Works for arbitrary number of goods, including allowing combination bids • Auction procedure: – Collect bids – Allocate goods to maximize total reported value (goods go to those who claim to value them most) – Payments: Each bidder pays her externality; Pays: (sum of everyone else’s value without bidder) (sum of everyone else’s value with bidder) • Incentive compatible (truthful)

Is Google pricing = VCG? Well, not really … Put Nobel Prize-winning theories to work. Google’s unique auction model uses Nobel Prize-winning economic theory to eliminate the winner’s curse – that feeling that you’ve paid too much. While the auction model lets advertisers bid on keywords, the Ad. Words™ Discounter makes sure that they only pay what they need in order to stay ahead of their nearest competitor. https: //google. com/adsense/afs. pdf Yahoo! Confidential

VCG pricing • (sum of everyone else’s value w/o bidder) - (sum of everyone else’s value with bidder) • CTRi = advi * posi (key “separability” assumption) • pricei = 1/advi*(∑jibidj*advj*posj-1 -∑j≠ibidj*CTRj ) = 1/advi*(∑j>ibidj*advj*posj-1 - ∑j>ibidj*CTRj ) • Notes – For truthful Y! ranking set advi = 1. But Y! ranking technically not VCG because not efficient allocation. – Last position may require special handling Yahoo! Confidential

Next-price equilibrium • • Next-price auction: Not truthful: no dominant strategy What are Nash equilibrium strategies? There are many! Which Nash equilibrium seems “focal” ? Locally envy-free equilibrium [Edelman, Ostrovsky, Schwarz 2005] Symmetric equilibrium [Varian 2006] Fixed point where bidders don’t want to move or – Bidders first choose the optimal position for them: position i – Within range of bids that land them in position i, bidder chooses point of indifference between staying in current position and swapping up with bidder in position i-1 • Pure strategy (symmetric) Nash equilibrium • Intuitive: Squeeze bidder above, but not enough to risk “punishment” from bidder above Yahoo! Confidential

Next-price equilibrium • Recursive solution: posi-1*advi*bi = (posi-1 -posi)*advi*vi+posi*advi+1*bi+1 posi-1*advi • Nomenclature: Next price = “generalized second price” (GSP) Yahoo! Confidential

Research Selected survey of sponsored search research at Yahoo! • Analytic comparison of mechanisms [Lahaie] • Learning click rates: N-armed bandit formulation [Pandey & Olsten] • Simulation I: Static [Feng, Bhargava, Pennock] • Simulation II: Equilibrium [Lahaie, Pennock] • Pragmatic robot [Schwarz, Edelman]

Source: S. Lahaie An Analysis of Alternative Slot Auction Designs for Sponsored Search Sebastien Lahaie, Harvard University* *work partially conducted at Yahoo! Research ACM Conference on Electronic Commerce, 2006

Source: S. Lahaie Objective • Initiate a systematic study of Yahoo! and Google slot auctions designs. • Look at both “short-run” incomplete information case, and “long-run” complete information case.

Source: S. Lahaie Outline • Incomplete information (one shot game) • • Incentives Efficiency Informational requirements Revenue • Complete Information (long-run equilibrium) • • • Existence of equilibria Characterization of equilibria Efficiency of equilibria (“price of anarchy”)

Source: S. Lahaie The Model • slots, bidders • The type of bidder i consists of • a value per click of , realization • a relevance , realization • is bidder i’s revenue, realization • Ad in slot So CTRi, k = is viewed with probability • Bidder i’s utility function is quasi-linear:

Source: S. Lahaie The Model (cont’d) • is i. i. d on according to • is continuous and has full support • is common knowledge • Probabilities are common knowledge. • Only bidder i knows realization • Both seller and bidder i know other bidders do not , but

Source: S. Lahaie Auction Formats • • • Rank-by-bid (RBB): bidders are ranked according to their declared values ( ) Rank-by-revenue (RBR): bidders are ranked according to their declared revenues ( ) First-price: a bidder pays his declared value Second-price (next-price): For RBB, pays next highest price. For RBR, pays All payments are per click

Source: S. Lahaie Incentives • First-price: neither RBB nor RBR is truthful • Second-price: being truthful is not a dominant strategy, nor is it an ex post Nash equilibrium (by example): 1 6 1 4 • Use Holmstrom’s lemma to derive truthful payment rules for RBB and RBR: • RBR with truthful payment rule is VCG

Source: S. Lahaie Efficiency • Lemma: In a RBB auction with either a first- or second-price payment rule, the symmetric Bayes-Nash equilibrium bid is strictly increasing with value. For RBR it is strictly increasing with product. • RBB is not efficient (by example). 0. 5 6 1 4 • Proposition: RBR is efficient (proof).

Source: S. Lahaie First-Price Bidding Equilibria • • is the expected resulting clickthrough rate, in a symmetric equilibrium of the RBB auction, to a bidder with value y and relevance 1. is defined similarly for bidder with product y and relevance 1. • Proposition: Symmetric Bayes-Nash equilibrium strategies in a first-price RBB and RBR auction are given by, respectively:

Source: S. Lahaie Informational Requirements • RBB: bidder need not know his own relevance, or the distribution over relevance. • RBR: must know own relevance and joint distribution over value and relevance.

Source: S. Lahaie Revenue Ranking • Revenue equivalence principle: auctions that lead to the same allocations in equilibrium have the same expected revenue. • Neither RBB nor RBR dominates in terms of revenue, for a fixed number of agents, slots, and a fixed.

Source: S. Lahaie Complete Information Nash Equilibria Argument: a bidder always tries to match the nextlowest bid to minimize costs. But it is not an equilibrium for all to bid 0. Argument: corollary of characterization lemma.

Source: S. Lahaie Characterization of Equilibria • RBB: same characterization with replacing

Source: S. Lahaie Price of Anarchy Define:

Source: S. Lahaie Exponential Decay • • • Typical model of decaying clickthrough rate: [Feng et al. ’ 05] find that their actual clickthrough data is fit well by such a model with In this case

Source: S. Lahaie • • Conclusion Incomplete information (on-shot game): • • Neither first- nor second-pricing leads to truthfulness. RBR is efficient, RBB is not RBB has weaker informational requirements Neither RBB nor RBR is revenue-dominant Complete information (long-run equilibrium): • • First-price leads to no pure strategy Nash equilibria, but second-price has many. Value in equilibrium is constant factor away from “standard” value.

Source: S. Lahaie Future Work • Better characterization of revenue properties: under what conditions on does either RBB or RBR dominate? • Revenue results for complete information case (relation to Edelman et al. ’s “locally envy-free equilibria”).

Source: S. Lahaie Research Problem: Online Estimation of Clickrates • • Make virtually no assumptions on clickrates. Each different ranking yields (1) information on clickrates and (2) revenue. Tension between optimizing current revenue based on current information, and gaining more info on clickrates to optimize future revenue (multi-armed bandit problem. . . ) Twist: chosen policy determines rankings, which will affect agent’s equilibrium behavior.

Research Handling Advertisements of Unknown Quality in Search Advertising Sandeep Pandey, Carnegie Mellon University Christopher Olston, Yahoo! Research and CMU Neural Information Processing Systems, 2006

Research CTR estimation • Explore/exploit tradeoff • Exploit: Use current CTR est’s to rank • Explore: Try new or low rank advertisers in higher positions to improve CTR est’s

Research Analytic results • Unbudgeted: Cast as independent multi-armed bandits, propose “MIX” policy • Budgeted: New budgeted multi-armed multi-bandit formulation (BMMP) bpol(N) >= opt(N)/2 + O(ln N)

Research Experiments: Real Y! data

Research Extensions • Using prior information e. g. algorithmic relevance of listing • Allowing ads to come and go at any time • Additional performance bounds

Research Implementing sponsored search in web search engines: Computational evaluation of alternative mechanisms Jane Feng, University of Florida Hemant Bhargava, University of California Davis David Pennock, Yahoo! Research Informs Journal on Computing, forthcoming

Research Simulation model Source: J. Feng

Research Simulation model • = relevance click rate (CTR) • v = advertiser value • ( , v) = bivariate normal • Revenue:

Research Allocation Rules Tested • • Bid ranking Revenue ranking Relevance ranking Posted price

Research Simulation Results

Research Simulation Results Number of paid slots

Research Simulation Results Effect of editorial control

Research Simulation Results Effect of naive learning of

Research

Research Equilibrium revenue simulations of hybrid sponsored search mechanisms Sebastien Lahaie, Harvard University* *work conducted at Yahoo! Research David Pennock, Yahoo! Research

Source: S. Lahaie Monte-Carlo simulations • 10 bidders, 10 positions • Value and relevance are i. i. d. and have lognormal marginals with mean and variance (1, 0. 2) and (1, 0. 5) resp. • Spearman correlation between value and relevance is varied between -1 and 1. • Standard errors are within 2% of plotted estimates. Yahoo! Confidential

Revenue effects Y! today Highest bid wins Google/Panama Highest bid*CTR wins Hybrid Highest bid*(CTR)s wins s=0 s=1/2 ? s=3/4 ? • What gives most revenue? – Key: If rules change, advertiser bids will change – Use Edelman et al. envy-free equilibrium solution Yahoo! Confidential

Source: S. Lahaie Yahoo! Confidential

Source: S. Lahaie Yahoo! Confidential

Source: S. Lahaie Yahoo! Confidential

Source: S. Lahaie Preliminary Conclusions • With perfectly negative correlation (-1), revenue, efficiency, and relevance exhibits threshold behavior • Squashing up to this threshold can improve revenue without too much sacrifice in efficiency or relevance • Squashing can significantly improve revenue with positive correlation Yahoo! Confidential

Source: M. Schwarz Pragmatic Robots and Equilibrium Bidding in GSP Auctions • Michael Schwarz, Yahoo! Research • Ben Edelman, Harvard University

Thanks: M. Schwarz Testing game theory • Empirical game theory – Analytic solutions intractable in all but simplest settings – Laboratory experiments cumbersome, costly – Agent-based simulation: easy, cheap, allow massive exploration; Key: modeling realistic strategies • Ideal for agent-based simulation: when real economic decisions are already delegated to software “If pay-per-click marketing is so strategic, how can it be automated? That’s why we developed Rules-Based Bidding. Rules -Based Bidding allows you to apply the kind of rules you would use if you were managing your bids manually. ” Atlas http: //www. atlasonepoint. com/products/bidmanager/rulesbased Yahoo! Confidential

Source: M. Schwarz Bidders’ actual strategies Yahoo! Confidential

Source: M. Schwarz Models of GSP 1. Static game of complete information 2. Generalized English Auction (simple dynamic model) More realistic model • Each period one random bidder can change his bid • Before the move a bidder observes all standing bids Yahoo! Confidential

Source: M. Schwarz Pragmatic Robot (PR) • Find current optimal position i Implies range of possible bids: Static best response (BR set) • Choose envy-free point inside BR set: Bid up to point of indifference between position i and position i-1 • If start in equilibrium PRs stay in equilibrium Yahoo! Confidential

Convergence of PR Simulation Yahoo! Confidential Source: M. Schwarz

Source: M. Schwarz Convergence of PR Yahoo! Confidential

Source: M. Schwarz Convergence of PR • The fact that PR converges supports the assertion that the equilibrium of a simple model informs us about the outcome of intractable dynamic game that inspired it Complex game that we can not solve Yahoo! Confidential ? Simple model inspired by a complex game

Source: M. Schwarz Playing with Ideal Subjects Largest Gap (commercially available strategy) Moves your keyword listing to the largest bid gap within a specified set of positions Regime One: 15 robots all play Largest Gap Regime Two: one robot becomes pragmatic By becoming Pragmatic pay off is up 16% Other assumptions: values are log normal, mean valuation 1, std dev 0. 7 of the underlying normal, bidders move sequentially in random order Yahoo! Confidential

Source: M. Schwarz ROI • Setting ROI target is a popular strategy • For any ROI goal the advertiser who switches to pragmatic gets higher payoff Yahoo! Confidential

Source: M. Schwarz If others play ROI targeter • Bidders 1, . . . , K-1 bid according to the ROI targeting strategy • What is K’s best response? bidder payoffs if bidder K plays ROI bidder targeting 1 … K-1 K Yahoo! Confidential 0. 0387 PR 0. 0457

Reinforcement Learner vs Pragmatic Robot • Pragmatic learner outperforms reinforcement learner (that we tried) • Remark: reinforcement learning does not converge in a problem with big BR set Yahoo! Confidential Source: M. Schwarz

Thanks: M. Schwarz Conclusion • A strategy inspired by theory seems useful in practice: PR beats commercially available strategies and other reasonable baselines • Since PR converges and performs well, the equilibrium concept is sound in spite the fact that some theoretical assumptions are violated and there are plenty of players who are “irrational” • When bidding agents are used for real economic decisions (e. g. , search engine optimization), we have an ideal playground for empirical game theory simulations Yahoo! Confidential

Research First Workshop on Sponsored Search Auctions at ACM Electronic Commerce, 2005 Organizers: Kursad Asdemir, University of Alberta Hemant Bharghava, University of California Davis Jane Feng, University of Florida Gary Flake, Microsoft David Pennock, Yahoo! Research

Research Papers • Mechanism Design • Pay-Percentage of Impressions: An Advertising Method that is Highly Robust to Fraud, J. Goodman • Stochastic and Contingent-Payment Auctions, C. Meek, D. M. Chickering, D. B. Wilson • Optimize-and-Dispatch Architecture for Expressive Ad Auctions, D. Parkes, T. Sandholm • Sponsored Search Auction Design via Machine Learning, M. -F. Balcan, A. Blum, J. D. Hartline, Y. Mansour • Knapsack Auctions, G. Aggarwal, J. D. Hartline • Designing Share Structure in Auctions of Divisible

Research Papers • Bidding Strategies • Strategic Bidder Behavior in Sponsored Search Auctions, Benjamin Edelman, Michael Ostrovsky • A Formal Analysis of Search Auctions Including Predictions on Click Fraud and Bidding Tactics, B. Kitts, P. Laxminarayan, B. Le. Blanc, R. Meech • User experience • Examining Searcher Perceptions of and Interactions with Sponsored Results, B. J. Jansen, M. Resnick • Online Advertisers' Bidding Strategies for Search, Experience, and Credence Goods: An Empirical Investigation, A. Animesh, V. Ramachandran, • S. Vaswanathan

Research Stochastic Auctions C. Meek, D. M. Chickering, D. B. Wilson • Ad ranking allocation rule is stochastic • Why? • Reduces incentive for “bid jamming” • Naturally incorporates explore/exploit mix • Incentive for low value bidders to join/stay? • Derive truthful pricing rule • Investigate contingent-payment auctions: Pay per click, pay per action, etc. • Investigate bid jamming, exploration strategies

Research Expressive Ad Auctions D. Parkes, T. Sandholm • Propose expressive bidding semantics for ad auctions (examples next) • Good: Incr. economic efficiency, incr. revenue • Bad: Requires combinatorial optimization; Ads need to be displayed within milliseconds • To address computational complexity, propose “optimize and dispatch” architecture: Offline scheduler “tunes” an online (real-time) dispatcher

Research Expressive bidding I • Multi-attribute bidding Advertiser 1 Advertiser 2 Male users (50%) $1 $2 Pre-qualified (50%) $2 $2 Female users (50%) $2 $1 Other (50%) $1 $1 Undifferentiated $1. 50

Research Expressive bidding II • Competition constraints b x. CTR = RPS 3 x. 05 =. 15 1 x. 05 =. 05

Research Expressive bidding II • Competition constraints monopoly bid b x. CTR = RPS 4 x. 07 =. 28

Research Expressive bidding III • • • Guaranteed future delivery Decreasing/increasing marginal value All or nothing bids Pay per: impression, click, action, . . . Type/id of distribution site (content match) Complex search query properties Algo results properties (“piggyback bid”) Ad infinitum Keys: What advertisers want; what advertisers value differently; controlling cognitive burden; computational complexity

Source: K. Asdemir Second Workshop on Sponsored Search Auctions Organizing Committee Kursad Asdemir, University of Alberta Jason Hartline, Microsoft Research Brendan Kitts, Microsoft Chris Meek, Microsoft Research

Objectives n Diversity q Participants q q q Industry: Search engines and search engine marketers Academia: Engineering, business, economics schools Approaches q q q n Source: K. Asdemir Mechanism Design Empirical Data mining / machine learning New Ideas

History & Overview n First Workshop on S. S. A. q q n Vancouver, BC 2005 ~25 participants 10 papers + Open discussion 4 papers from Microsoft Research Second Workshop on S. S. A. q q q ~40 -50 participants 10 papers + Panel 3 papers from Yahoo! Research Source: K. Asdemir

Participants n Industry q q n Source: K. Asdemir Yahoo!, Microsoft, Google Iprospect (Isobar), Efficient Frontier, HP Labs, Bell Labs, Commerce. Net Academia q Several schools

Papers n Mechanism design q q n Bidding behavior q q n Edelman, Ostrovsky, and Schwarz Iyengar and Kumar Liu, Chen, and Whinston Borgs et al. Zhou and Lukose Szymanski and Lee Asdemir Borgs et al. Data mining q q Regelson and Fain Sebastian, Bartz, and Murthy Source: K. Asdemir

Source: K. Asdemir Panel: Models of Sponsored Search: What are the Right Questions? n Proposed by q n Lance Fortnow and Rakesh Vohra Panel members q q Kamal Jain, Microsoft Research Rakesh Vohra, Northwestern University Michael Schwarz, Yahoo! Inc David Pennock, Yahoo! Inc

Panel Discussions n Mechanisms q q q n q Hard or a soft constraint Flighting (How to spend the budget over time? ) Pay-per-what? CPM, CPC, CPS q q n Competition between mechanisms Ambiguity vs Transparency: “Pricing” versus “auctions” Involving searchers Budget q n Source: K. Asdemir Risk sharing Fraud resistance Transcript available!

Research Web resources • 1 st Workshop website & papers: http: //research. yahoo. com/workshops/ssa 2005/ • 1 st Workshop notes (by Rohit Khare): http: //wiki. commerce. net/wiki/RK_SSA_WS_Notes • 2 nd Workshop website & papers: http: //www. bus. ualberta. ca/kasdemir/ssa 2/ • 2 nd Workshop panel transcript: (thanks Hartline & friends!) http: //research. microsoft. com/~hartline/papers/ panel-SSA-06. pdf

Research http: //buzz. research. yahoo. com • • • Yahoo!, O’Reilly launched Buzz Game 3/05 @ETech Research testbed for investigating prediction markets Buy “stock” in hundreds of technologies • Earn dividends based on actual search “buzz” • • • API interface Exchange mechanism is Yahoo! invention (dynamic parimutuel) Cross btw stock market and horse race betting

Research Technology forecasts • i. Pod phone • What’s next? Google Calendar? price search buzz 8/28: buzz gamers begin bidding up i. Pod phone 8/29: Apple invites press to “secret” unveiling 9/7: Apple announces Rokr 9/8 -9/18: searches for i. Pod phone soar; early buyers profit • Another Apple unveiling 10/12; i. Pod Video? 9 am 10/5

Research Forecast accuracy Early lessons learned • Average forecast error across 352 stocks • Market closing deadline focuses traders • Dividend levels matter • Intelligent strategies work forecast error rapidly declines as traders zero in on correct predictions end of phase 1 contest period • Randomized bots lost money to real traders • Contest winner followed optimal buzz trading strategy (prices buzz); Went from 4 th to 1 st place in final days • Forecast error does decrease over time

Research Forecast accuracy • Stocks categorized by Day 0 implied buzz / actual buzz • Graph shows movement of actual buzz for each category

Research Tech Buzz Game

Research Pari-mutuel market Basic idea 1 1 1

Research Dynamic pari-mutuel market Basic idea 1 1 0. 2 0. 4 1. 3 1. 6 0. 9 2 5 2. 3

Research How are prices set? • A price function pi(n) gives the instantaneous price of an infinitesimal additional share beyond the nth • Cost of buying n shares: 0 pi(n) dn • Different reasonable assumptions lead to different price functions n

Research Share-ratio price function • One can view DPM as a market maker • Shares pay equal portion of total $$: C(Qfinal)/qo >= $1 • Ratio of shares qi/qj = ratio of prices pi/pj • Cost Function: • Price Function:

More Challenges • Predicting click through rates • Detecting click spam • Pay per “action” / conversion • Number of ad slots • Improved targeting / expressiveness • Content match