Second Circuit Says Google’s Keyword Ad Sales May Be Use in Commerce–Rescuecom v. Google
By Eric Goldman
Rescuecom Corp. v. Google Inc., 562 F.3d 123 (2d Cir. April 3, 2009)
The Second Circuit has issued its long-anticipated opinion in Rescuecom v. Google over Google’s sale of trademarked keywords as ad triggers. In a disappointing but not surprising conclusion, the Second Circuit reversed the lower court and says that Rescuecom properly alleged that Google’s keyword ad practices constituted a “use in commerce.” This ruling merely reverses the 12b6 dismissal for Google, but it raises some important questions–including whether this ruling effectively eliminates any future “use in commerce” defense in keyword advertising cases and whether Google and other search engines could reform their practices so that they are no longer deemed uses in commerce.
1-800 Contacts v. WhenU Distinguished
The most interesting part of the opinion is how this panel distinguishes its 2005 1-800 Contacts v. WhenU precedent, which held that an adware vendor did not make a use in commerce through its keyword ad triggering processes. The court says that Google is different in two main respects:
“First, in contrast to 1-800, where we emphasized that the defendant made no use whatsoever of the plaintiff’s trademark, here what Google is recommending and selling to its advertisers is Rescuecom’s trademark. Second, in contrast with the facts of 1-800 where the defendant did not “use or display,” much less sell, trademarks as search terms to its advertisers, here Google displays, offers, and sells Rescuecom’s mark to Google’s advertising customers when selling its advertising services. In addition, Google encourages the purchase of Rescuecom’s mark through its Keyword Suggestion Tool.”
The court appears to be making two distinctions. First, WhenU didn’t sell trademarked keywords directly but instead rolled up search queries into product categories that didn’t contain the trademark anywhere but in an internal database table, so there was an additional layer of abstraction away from trademarks built into WhenU’s matching process. Second, the court clearly doesn’t like Google’s Keyword Suggestion Tool, which I think has also frustrated trademark owners and been repeatedly cited against Google in pleadings.
In theory, then, Google could eliminate its trademark use in commerce by adding a product category abstraction–although this may not be a good idea, as it would not work with long-tail queries–and by modifying or dropping the Keyword Suggestion Tool.
The case also discusses Google’s “sponsored link” label and distinguishes it from WhenU’s labeling of its pop-up ads. The court gives credence (as it must on a 12b6) to Rescuecom’s allegations that Google’s placement of ads above the organic results might confuse consumers into thinking those ads were organic. In contrast, in WhenU, the “pop-up ad appeared in a separate browser window from the website the user accessed, and the defendant’s brand was displayed in the window frame surrounding the ad, so that there was no confusion as to the nature of the pop-up as an advertisement, nor as to the fact that the defendant, not the trademark owner, was responsible for displaying the ad, in response to the particular term searched.” Personally, I think Google’s interface is sufficiently clear to consumers, but this is a factual assertion not ready for judicial review in this case yet.
One oddity: the court repeatedly says that WhenU displayed ads “randomly” chosen in response to searcher behavior. I’m not sure what the court was trying to say, but the ads were hardly chosen at random, and this is a pretty significant factual error on the court’s part.
Finally, the court discusses the analogies to shelf-space adjacency in the retail context. This is a topic of special interest because I’ve parsed this issue in gory detail in my Brand Spillovers paper. The court, without any citations, reaches the conclusion that
It is not by reason of absence of a use of a mark in commerce that benign product placement escapes liability; it escapes liability because it is a benign practice which does not cause a likelihood of consumer confusion. In contrast, if a retail seller were to be paid by an off-brand purveyor to arrange product display and delivery in such a way that customers seeking to purchase a famous brand would receive the off-brand, believing they had gotten the brand they were seeking, we see no reason to believe the practice would escape liability merely because it could claim the mantle of “product placement.”
Fair enough—if consumers purchase a passed-off good, that would be actionable. However, the court sidesteps all of the nuance in concluding that shelf-space adjacency is a “benign practice that does not cause…consumer confusion.” Retailers are hardly “benign” in their practices; see my Brand Spillovers paper for more on that. Further, and perhaps more importantly, it’s unclear how Google’s ads misdirect anyone. The court had to accept Rescuecom’s allegations of diversion as true, but I think those bear very close scrutiny on remand.
What Is a Use in Commerce?
The opinion also contains a scholarly appendix, expressly labeled as dicta, explaining its statutory analysis of the Lanham Act’s use in commerce phrase. Not surprisingly, at the end of the appendix it says “It would be helpful for Congress to study and clear up this ambiguity.” Although it is dicta, I expect many other courts will follow and embrace this appendix when discussing use in commerce. I also expect that this will put an end to the cottage industry of law review articles debating what the phrase means in the keyword context.
Implications of this Ruling
1) This opinion narrows the 1-800 Contacts v. WhenU opinion substantially to a very specific set of facts. I’m not sure how many courts will be favorably citing that precedent in the future.
2) This case jeopardizes the half-dozen or so district court cases (in Second Circuit-controlled jurisdictions) that have held that keyword advertising purchases aren’t a trademark use in commerce. This case involves Google’s sale of keyword advertising, not an advertiser’s purchase of keyword advertising, but I think those cases are now very shaky precedent. (The court particularly says that the Merck and S&L Vitamins cases “overread” the 1-800 Contacts precedent). The Second Circuit still could find a way to distinguish ad buys from ad sales, but I would be surprised if it did so.
3) This case also jeopardizes the rulings in those cases that keyword metatags aren’t a trademark use in commerce. The court says specifically “We did not imply in 1-800 that an alleged infringer’s use of a trademark in an internal software program insulates the alleged infringer from a charge of infringement, no matter how likely the use is to cause confusion in the marketplace.” I’m not sure how this applies to keyword metatags, which can’t cause consumer confusion under any circumstance. Nevertheless, if the keyword metatags don’t have the layer of abstraction that WhenU used, I don’t think the court would regard them favorably.
4) Although this is clearly a loss for Google because Google no longer has a reliable way to kick out cases on a 12b6, Google might still prevail in the case. Google had won on a 12b6, and the court merely said that Rescuecom alleged enough in its complaint to survive the 12b6. Google could still win on summary judgment or trial, or the parties might settle. Either way, Rescuecom merely lives to fight another day. (In theory, Google could also appeal this ruling to the Supreme Court; I would be surprised if they went that route or if the Supreme Court would take it).
5) Accordingly, I don’t expect this ruling to do much for cases like American Airlines v. Yahoo. Indeed, perhaps anticipating this loss, Yahoo didn’t try to get the case into the Second Circuit. I suspect that’s because Yahoo had already decided not to expect the use in commerce defense to go in its favor.
6) I’m interested to see what this ruling will do to state efforts to attack keyword advertising, such as Utah’s ill-fated forays in this area. In theory, this ruling might alleviate some of the pressure state legislators feel that they have to do something. However, I suspect state legislators are only mildly interested in legal proceedings elsewhere, so I doubt this will make state legislators second-guess their own brilliance.
7) As the court says, it would make a lot of sense for Congress to clean up the statutory drafting muddle over use in commerce in the Lanham Act. I don’t think this is likely because of the political gridlock that would emerge over the topic. As I discuss in my Deregulating Relevancy paper, a more pragmatic approach would be for Congress to expressly provide a safe harbor for search engines selling keywords analogous to the safe harbor for domain name registrars selling domain names, but I doubt Google has the muscle for that either. As a result, I don’t anticipate legislative intervention to overturn this ruling.
The case library:
* Commercial Referential Trademark Uses (Rescuecom v. Google Amicus Brief Outtakes)
* Law professors’ brief by Stacey Dogan and me
* Electronic Frontier Foundation amicus brief by Jason Schultz, Corynne McSherry and Fred von Lohmann
* Public Citizen amicus brief by Paul Levy
* eBay/Yahoo/AOL amicus brief by Celia Goldwag Barenholtz, Janet Cullum and others of Cooley Godward Kronish [now mooted]