A proposed advertising deal between Internet competitors Google and Yahoo would reduce competitiveness in the Internet advertising market, likely resulting in higher advertising rates, according to Harvard Business School assistant professor Benjamin Edelman.
In testimony prepared for an appearance before the U.S. House of Representatives Committee on the Judiciary Task Force on Competition Policy and Antitrust Laws, Edelman wrote that the deal could have an anticompetitive effect on Internet advertising. "The proposed agreement creates a serious risk of foreclosing the future of Yahoo's ad platform, and reducing—from three to two—the number of major sellers of sponsored search advertising."
The deal, characterized as a nonexclusive partnership, would allow Google to sell some search ads that display next to Yahoo search results. Yahoo says it would see significant revenue increases from the arrangement, but critics claim Google would be the ultimate winner by consolidating more power in the market.
Edelman prepared the remarks for a June 27, 2008, hearing, which was later rescheduled. Here is the prepared testimony:
Chairman Conyers, Ranking Member Chabot, Members Of The Subcommittee:
My name is Benjamin Edelman. I am an assistant professor at the Harvard Business School, where my research focuses on the design of electronic marketplaces, including designing online marketplaces to assure safety, reliability, and efficiency. My full biography and publication list are at http://www.benedelman.org/bio and http://www.benedelman.org/publications. Relevant disclosures appear on the final page of my testimony.
Today the Committee considers the important question of Google's proposed purchase of certain advertising space from Yahoo.
My bottom line:
- Google sets reserve prices and other parameters that substantially determine prices. Contrary to Google's claims, Google's auction does not fully determine prices; Google's actions and policies importantly influence prices.
- Google's purchase of substantial advertising inventory from Yahoo would increase prices for many advertisers that currently buy ads from Yahoo.
- The proposed deal would substantially reduce Yahoo's ability to offer competitive payments to Web site publishers seeking to show pay-per-click ads. Without Yahoo bidding against Google to obtain publishers' inventory, publishers are likely to receive far lower payments.
- Other Google practices, particularly Google's restrictions on export and copying of advertisers' campaigns, further hinder competition in Internet advertising-without any countervailing benefit whatever.
Google Sets Prices
Google has publicly claimed that because it sells ads through an auction, it does not set prices. See, e.g., Larry Page's May 2008 remarks to C|NET: "AdWords is an auction. We're not setting prices. Auctions are determined by supply and demand." 1
I disagree. Google controls a variety of characteristics of its marketplace, and Google can use these characteristics to influence prices.
For example, consider an advertiser that is the sole bidder for a given term, e.g., a product that it alone sells. That advertiser's payment to Google does not reflect bidding to beat a competitor. Rather, that advertiser pays a reserve price set by Google at its discretion.
More generally, Google sets reserve prices for a variety of aspects of its marketplace-including the minimum price for an ad to appear at all, the minimum price for a position in the valuable space at the top of search results, and minimum bids for particular ads and particular keywords.
These minimum bids are not mere baseline requirements that apply equally to all. To the contrary, Google sets different minima for specific ads, specific keywords, and specific advertisers. By raising these minimum bids, Google can increase advertisers' payments.
Furthermore, minimum bids affect the bids of higher-ranked advertisers. Consider an advertiser ("A") who obtains the next-to-last slot. The fee paid by the lowest advertiser ("B") is generally set by Google's minimum bid. But A needs to set its bid high enough to prevent B from electing to raise its bid to take the next-to-last slot. In Internet Advertising and the Generalized Second Price Auction 2 my coauthors and I develop this envy-free condition. In Optimal Auction Design in a Multi-unit Environment 3 my coauthor and I demonstrate that when a search engine raises its minimum bid, the indirect revenue effect (increased payment from higher-ranked advertisers like A can exceed the direct effect (increased payment by the lowest-ranked advertiser, i.e., B. So minimum bids matter even more than intuition suggests.
Other Google decisions can also affect advertisers' payments to Google. Google can elect to increase or decrease the number of advertising positions it shows on a page. Google can change the placement and prominence of ads, or make some ads more prominent than others. Google can reward or penalize advertisers through secret adjustments that only it knows. By making these choices strategically, Google can increase advertisers' costs.
Increased Costs To Advertisers
Once Google controls a substantial share of inventory at Yahoo, advertisers' search advertising options will be severely limited: Save for Google, advertisers will have few viable choices for search advertising. Google will then obtain substantial power to increase prices: If Google increases prices, advertisers will be unable to shift to a competing provider to obtain a comparable quantity of traffic. Shifts to other kinds of advertising—be they banner ads, TV, or print—are similarly unavailing; for typical online advertisers, these are poor substitutes.4
The harm to advertisers will be particularly acute for those advertisers that currently advertise on Yahoo for average to obscure search terms—e.g., small businesses selling unusual products, or local businesses selling in a limited geographic region. At Yahoo, these advertisers currently enjoy prices that are relatively low, because they face relatively few competitors. Once Yahoo begins to sell this traffic through Google, these advertisers will face sharply increased advertising fees. In particular, these advertisers will either have to place their bids through Google, or increase their bids at Yahoo to pay the higher Google price.
Market Concentration And Future Innovation
The proposed agreement creates a serious risk of foreclosing the future of Yahoo's ad platform, and reducing-from three to two-the number of major sellers of sponsored search advertising.
But if Yahoo assigns much of its traffic to Google, advertisers will have little incentive to sign up with Yahoo in the first place. Yahoo will find itself with only the remnants of an ad platform—not a robust marketplace with a wide variety of listings. Yahoo itself has worried that a deal with Google would create "an effective monopoly."7
“Save for Google, advertisers will have few viable choices for search advertising.”
Serious harms occur to others if Yahoo's ad platform disintegrates. Consumers benefit from diversity in ad platforms so that no single platform can unduly restrict the information available to consumers. Such problems are more than hypothetical. Consider the August 2004 incident in which Google banned an ad critical of President Bush 8—leading to revelation of a variety of Google restrictions on ads about politics, religion, and abortion, among other topics. These are sensitive subjects, and perhaps Google's existing rules are exactly right. But if Google obtains excessive market share, consumers who prefer a different approach will have no viable choice.
The Effect Of Market Concentration On Payments To Publishers
The future viability of Yahoo's ad platform is particularly important for publishers seeking to show sponsored search ads in their Web sites or search results. Publishers' payments are set through competition between search providers: To win a publisher's business, Google must offer enough to outbid Yahoo. But if many advertisers cease advertising with Yahoo directly, Yahoo will be unable to offer high payments. Then publishers will have little real alternative to Google—letting Google win a publisher's business while paying far less than current rates.
Google may argue that publishers could show some other kind of advertising in lieu of sponsored search. But sponsored search ads tend to carry higher fees than graphical ads. Moreover, text ads are viewed more favorably by users, e.g., as potentially useful rather than distracting. Also, search ads can better match users' apparent interests. So for many publishers, display ads are not a viable alternative. Nor is any other choice readily apparent.
The Illusory "non-exclusive" Provision
The parties characterize the proposed relationship as "non-exclusive," which they claim reduces its anti-competitive effect. I disagree. Despite Google's willingness to let Yahoo sell advertising space to others, no other buyer could afford to pay as much as Google. Yahoo is unlikely to exercise its right to sell search ads through other providers, for no such sale is likely to yield revenues comparable to what Yahoo can obtain from Google.
Google's Opaque Approach Hinders Detection Of Abuses Of Market Power
Google does not disclose the methods by which it sets reserve prices or other details of its dealings with advertisers. Furthermore, Google does not generally disclose the terms of its contracts with publishers.
Due to Google's opacity, advertisers, publishers, and the public face substantial challenges in identifying trends in the terms Google offers. It is therefore unusually difficult to detect any Google abuse of market power; Google can use its market power without leaving a trail of publicly available documents.
Api Terms & Conditions As A Further Barrier To Search Engine Competition
I take the Committee's larger concern to be barriers impeding competition among search engines. Looking beyond the proposed Google-Yahoo transaction, I urge the Committee to examine the "may not … cop[y]" provision in Google's AdWords API Terms & Conditions. This provision lacks any plausible pro-competitive benefit, but impedes advertisers' signup with competing paid search providers—thereby hindering competition in search engine advertising.
Google has long offered an application programming interface ("API") by which advertisers can submit advertisements to Google, monitor their advertising campaigns, and update their bids. The API is an automated method for efficient computer-to-computer communications—letting an advertiser use software to manage its account automatically.
The AdWords API provides a function whereby an advertiser can retrieve its ads, and this function would be a natural way to let an advertiser copy ads from Google to a competing vendor. Consider a Google advertiser that wanted to begin to advertise with Yahoo or Microsoft adCenter too. Yahoo, Microsoft, or third-party software developers could help the advertiser use the API to directly download all ads, then reformat the ads for uploading to other providers as desired. In this way, the advertiser could quickly and easily signup-without unnecessary transition costs or delay.
Remarkably, Google's AdWords API Terms & Conditions9 exactly prohibit such copies. The T&C's require, in relevant part, that an "AdWords API Client may not offer a functionality that copies data … from an AdWords account to a non-AdWords account."
The clear effect-and explicitly stated intention-is to impede advertisers' efforts to efficiently copy their ad campaigns to other providers. Instead, advertisers must resort to convoluted procedures to copy their ad campaigns out of Google. For example, Microsoft offers a 17-step procedure to copy Google AdWords ads into Microsoft adCenter. 10
I see no plausible pro-competitive effect of the quoted provision. Moreover, Google's own statements reveal the benefits of data portability. See, e.g., Google CEO Eric Schmidt: "People should be able to move from place to place, and their data is available everywhere" and "open is best for the consumer." 11 In contrast, the quoted provision hinders advertisers considering expanding beyond Google-impeding growth of competing sponsored search providers.
My PPC Platform Competition and Google's "May Not Copy" Restriction 12 presents these API concerns in greater detail.
I appear on my own behalf, not on behalf of Harvard Business School or anyone else.
I serve as a consultant to Microsoft on subjects unrelated to that at issue here.
I serve as cocounsel in Vulcan Golf, LLC, et al., v. Google Inc., et al., Civ. Act. No. 07 CV 3371 (N. Dist. Ill.). I previously served as co-lead counsel in In Re Yahoo Litigation, Master File CV-06 2737 CAS (FMOx) (C. Dist. CA.). These matters are both unrelated to that at issue here.