ABA IP Section Quietly Considering Anti-Consumer Proposals to Regulate Keyword Advertising
By Corynne McSherry and Eric Goldman
[Eric's note: this is a bit of an unusual post for this blog, as it is being simultaneously posted both here and on the EFF Deep Links site]
The tussle over keyword advertising has spilled over into numerous arenas, including the courts, the legislatures (such as Utah’s ill-conceived attempt to ban keyword advertising), the private trademark policies of the search engines and the law review literature. Given the magnitude of the issue—and the billions of dollars associated with keyword advertising—it’s not surprising that new battlefields are cropping up all over the place. But the latest skirmish has an unexpected venue—the closed-door deliberations of the American Bar Association.
The Trademark Litigation subcommittee of the ABA’s IP Section is evaluating four resolutions relating to keyword advertising. The emergence of this battlefront is hardly good news, however. From a substantive standpoint, the resolutions reflect misguided and anti-consumer priorities. From a procedural standpoint, it’s hard to see why the ABA would take a side in these contentious matters, and it’s not clear how these resolutions add any value to this debate. We strongly urge the subcommittee members to reject the resolutions.
The first resolution offers a perplexing answer to a problem that doesn’t exist. It declares that the likelihood of consumer confusion is “purely an issue of fact.” The question is: who has argued otherwise? As noted by the leading treatise on trademarks, by Prof. McCarthy, this resolution restates the general rule that “Traditionally, the law has classified likelihood of confusion as an issue of fact.” See McCarthy on Trademarks and Unfair Competition Sec. 23:67.
Indeed, no keyword advertising case has suggested otherwise. Having said that, at least one case was resolved on summary judgment when the court concluded that there were no material disputes about the facts—and rightly so. See, e.g., J.G. Wentworth SSC Ltd v. Settlement Funding LLC, 2007 WL 30115 (E.D. Pa. Jan. 4, 2007), where the court granted summary judgment to the advertisers based on the lack of evidence of consumer confusion. The resolution doesn’t seem to affect the J.G. Wentworth situation, so it’s not clear what it’s supposed to accomplish.
At the moment, it appears this resolution doesn’t do anything more than restate current law. Surely the Section has better things to do than pass resolutions approving the status quo.
The resolution also proposes that, with respect to keywords, the analysis must consider how a keyword or trademark is used in an ad or to link to a separate site. Again, in most cases that will already happen. But there may be other circumstances to consider in any given case: e.g., whether links appear in such a way that consumers are highly unlikely to be confused. Judges already have ample guidance on this issue, and they are well capable of applying that precedent to specific factual situations.
The only conceivable benefit of this resolution is that it might discourage courts from using doctrinally deficient heuristics like the initial interest confusion doctrine, which some courts have used to bypass a more careful inquiry into consumer confusion. For more discussion about the problems of the initial interest confusion doctrine, see Eric Goldman, Deregulating Relevancy in Internet Trademark Law. We’d be thrilled if that were the goal, but it would be better accomplished by explicitly encouraging courts to stop using the initial interest confusion doctrine. However, given that the initial interest confusion doctrine is a favorite tool of plaintiffs with weak cases, and given the pro-plaintiff bias of the resolutions, we suspect that wasn’t the point here. So we’re left scratching our heads, wondering why this resolution is being proposed at all.
The second resolution seeks to declare that a trademark use in commerce always occurs (1)“if it appears directly in an advertisement on the resulting Internet web page,” and/or (2) when a trademark is (a) “hidden in a Metatag,” or (b) “used principally for its importance as a Keyword.”
The first part of this resolution appears to restate the law. Every court that has opined on the topic has concluded that a use in commerce occurs when an advertiser references a third party trademark in a keyword-triggered ad. See, e.g., Hamzik v. Zale Corp./Delaware, 2007 WL 1174863 (N.D.N.Y. Apr. 19, 2007).
Even so, the current state of the law is misdirected, and the Section should not encourage it further. If the ad copy says “our product is cheaper than [competitor’s trademark],” the advertiser isn’t trying to portray its goods as the competitor’s—yet, the advertiser has to defend an expensive lawsuit over whether its reference creates a likelihood of consumer confusion. Instead, trademark law should give more breathing room to advertisers making “referential uses” of third party trademarks. See Universal Communication Systems, Inc. v. Lycos, Inc., 2007 WL 549111 (1st Cir. Feb. 23, 2007); Eric Goldman, Online Word of Mouth and its Implications for Trademark Law.
Because of the misguided state of the current law, Google has adopted a trademark policy that allows trademark owners to prevent competitors from including the phrase “we’re cheaper than [trademark owner]” in the ad copy—or, for that matter, referencing the competition at all in the ad copy. This may make keyword ads more confusing to consumers, not less.
The second part of the resolution has at least two fundamental flaws. The discussion about metatags reflects an uneducated—or, at best, dated—view of the Internet. Back in the 1990s, some search engines considered “hidden text” such as metatags in their rankings algorithms. Today, many search engines—most notably Google—simply ignore hidden metatags when making ranking determinations. If a website includes a third party trademark in its metatags and the search engines simply ignore it, treating it as a use in commerce defies reason.
A blanket rule that a trademark is used in commerce whenever it is “used principally for its importance as a keyword” is equally specious. We don’t understand how this test is supposed to be applied, and the accompanying discussion does little to illuminate matters. For example, the resolution explains that the person making a “use in commerce” is the person “who was to gain click-through monies at the expense of the trademark owner,” which raises a set of impossible-to-determine factual questions, such as when does keyword advertising actually take money out of the pockets of trademark owners, rather than simply presenting consumers with increased choices that improve their marketplace decisions. Of course, some trademark owners mistakenly think that every consumer searching for their trademark is “their” customer and therefore is poached by anyone who seeks to educate that consumer about other marketplace alternatives. We hope the resolution drafters aren’t succumbing to such an empirically unsupportable and anti-consumer view.
Finally, in an ineffectual nod to fair use, the resolution makes a confused distinction between a purely nominative or “otherwise allowable” use and a use intended to trigger keyword ads. Of course, the one does not preclude the other.
The third resolution proposes that search engine liability for keyword advertising should always be a question of fact. Like the fourth resolution (discussed below), this appears to be a simple attack on 1-800 Contacts ruling—which held that the use of trademarks to display ads does not, by itself, trigger trademark liability—with no justification for why the Section should substitute its judgment for that of the Court of Appeals and allow trademark owner to effectively control what appears on consumers’ desktops.
Moreover, this resolution appears to be designed to ensure that keyword lawsuits will be as expensive as possible, to no good purpose. It rejects the numerous cases that have held that search engines don’t make a trademark use in commerce when selling trademarked keywords (see, e.g., Rescuecom v. Google), which have led to very quick and efficient victories for the search engines. In contrast, when courts have held that search engines do make a trademark use in commerce for their sales, the lawsuits can become needlessly protracted. For example, the American Blinds case went almost four years before American Blinds completely capitulated in a settlement; the 1-800 JR Cigar v. Yahoo case ran a remarkable six years before settling. Years of litigation may generate a lot of legal fees and help lawyers pay the bills, but these lawsuits do little to help most trademark owners or consumers.
Besides, search engines already do plenty to help trademark owners. Every major search engine has an internal policy that allows trademark owners to restrict advertising on their trademarks without needing to run to court. In practice, this means that search engines help censor ads that trademark owners don’t like, especially ads that may increase competitive pressures on trademark owners. If courts ever held search engines liable for selling keywords, search engines would be even more aggressive about squelching ads that might benefit consumers.
The fourth resolution seeks to declare a 2005 Second Circuit opinion, 1-800 Contacts v. WhenU, to be a “minority position” that is “flawed in its reasoning.” In support of this, the resolution notes that the case was “based on facts too unusual to apply to most other cases in this area.”
The 1-800 Contacts v. WhenU case involved the sale of trademarked keywords by an adware vendor. In that case, the Second Circuit flatly declared that such sales did not constitute a trademark use in commerce, handing a decisive and clean victory to the defense. This case has proven to be a major ruling, with at least a half-dozen cases favorably interpreting it to find for the defendants in keyword advertising cases.
The resolution is therefore factually wrong to characterize the 1-800 Contacts opinion as a minority view, as it is the case that is the most frequently cited favorably. The resolution is also normatively wrong when it says the case is flawed in its reasoning; nor does it offer any explanation for this conclusion.
Beyond the substantive deficiencies, these resolutions raise two interesting procedural issues.
First, why is the ABA IP Section attempting to intervene in this contentious area of the law in the first place? These resolutions are political in nature and reflect a deep trademark owner-favorable bias that does not necessarily reflect the views of American lawyers generally or IP lawyers specifically. It’s hard to see how these biased and divisive statements are an appropriate use of the ABA’s resources or authority.
Second, the ABA subcommittee proposing these resolutions has sought to keep them from public dissemination. An ABA committee representative emailed one of us to claim that we “cannot share committee or subcommittee work product with [non-committee members]” (although we received the resolutions from several sources, including the subcommittee chair himself), and we were asked to destroy “any ABA-IPL Section 208 Keyword Subcommittee work product that you may have received.” We don’t understand why an ABA committee seeks to conduct deliberations about resolutions–that are presumably intended to be promulgated to the legal community–behind closed doors, away from public scrutiny. As this critique of their content indicates, these resolutions would substantially benefit from the input of a wider range of voices. We hope the committee solicits that input or, better yet, squelches this ill-conceived effort to micromanage the courts.
UPDATE: Paul Levy offers more insights.