Long Bow has a website with information critical of Jenzabar. These results showed up high in Google searches for “Jenzabar.” Jenzabar sued, though it quickly dropped its defamation claims, and pursued its meritless trademark claims instead. The court of appeals affirmed a grant of summary judgment in Long Bow’s favor. As others have noted, the decision is a mixed one: it affirms what is obvious, that critical uses aren’t infringing—even using the doctrine of initial interest confusion—but it fails to make clear that this is a case that shouldn’t have survived a motion to dismiss, even assuming that a multifactor confusion analysis (as opposed to a nominative fair use or straight-up First Amendment defense) applied. And there’s a dissent that seems to think that trademark is for controlling what other people say about you.
Jenzabar makes educational software for colleges and universities (who might well consider how much they want to contract with a business willing to sue those who speak ill of it for trademark infringement). Its founder, Ling Chai, also was involved in the 1989 Tiananmen Square protests, about which Long Bow made a documentary. Long Bow created a related website with background information, with over 2000 pages of text, an interactive map of Tiananmen Square, and a media library with photos, music, and video clips. One section contains information about 16 key characters, including Chai. There’s a short bio of her, noting that she declined to be interviewed for the film and containing exceprts from her criticism of the film. After Long Bow learned that she’d founded a software company, Long Bow added a link from the bio to a new page with information about Jenzabar, titled “Jenzabar.”
A Google search for “Jenzabar” would find Long Bow’s page on the first result page, in the form:
The information on these pages about Chai Ling and Jenzabar, the software company she runs with her husband, Robert Maginn, contains excerpts from and links ...
The title came from the title metatag, and the description came from the opening text on the page. The court treated as true Jenzabar’s claim that Long Bow could improve its page ranking through keyword metatags, even though Google and other sources say that’s not true. Long Bow used Jenzabar as a keyword metatag on its Jenzabar page and “several subsidiary pages linked to it,” though only the Jenzabar page showed up in the searches.
Jenzabar complained to Long Bow, alleging that some of the information on the site was false. Long Bow altered the site somewhat, including making clear that Jenzabar disputed the accuracy of some of the claims and adding prominent disclaimers of affiliation, but kept the URL, title, and metatags. Jenzabar sued for defamation, trade libel, trademark infringement, state and federal dilution, and unfair competition, but pretty soon all that was left was infringement (and equally ridiculous state dilution).
The court of appeals noted Long Bow’s argument that its use was not “in connection with the sale, offering for sale, distribution, or advertising of any goods or services” but decided to resolve the issue on other grounds, given that there was some support in the case law for the idea that, because there was contact information for distributors elsewhere on the Long Bow site, any use on the site was “in connection with” the sale of the films. (This strikes me as unnoticed web exceptionalism. Open the nearest mass-published book of your choice: can you find information about how to get more from the same publisher? By this logic, every mention of a trademark in the book's text is "in connection with" the sale of the publisher's books.)
Jenzabar didn’t even try to claim real confusion or confusion that would persist past the moment a visitor read the site (and the court noted that a confusion claim would be “untenable” even without the disclaimers), only initial interest confusion. Someone searching for Jenzabar allegedly might click Long Bow’s link “under the mistaken belief that the linked page is officially sponsored by Jenzabar,” and that was enough to be actionable.
Sadly, the court didn’t just say that this was bunk, or say that initial interest was limited to pure “bait and switch” by a competitor, though it noted that the doctrine has its foundation in bait and switch. (In a footnote, it also declined to rely on yet another rational limit on the doctrine, that its application should require financial benefit from the confusion.) Instead, it identified six relevant factors: (1) the relatedness of the parties' goods or services; (2) the level of care exercised by the relevant consumers; (3) the strength of the plaintiff's mark; (4) the defendant's intent in using the plaintiff's mark; (5) evidence of actual consumer confusion; and (6) the visual appearance of the link/ad in the context on the screen. No single factor was dispositive, but the first two were especially important. Without them, confusion would have little or no meaningful effect in the market.
Here, there was no similarity between the parties’ goods: complex software and documentaries about China. Some overlap in the “education” field was completely insufficient, strongly suggesting that there was no actionable confusion. Jenzabar claimed that it might expand its business to educational content, but provided no evidence and also conceded the contrary in deposition.
Also, Jenzabar’s consumers were quite sophisticated; the software was expensive and its purchase was a matter of careful consideration. “It is not sufficient for Jenzabar to suggest that a careless Web user might be confused by Long Bow's search result. To defeat summary judgment, Jenzabar must instead produce evidence from which a jury could conclude that Jenzabar's actual or potential customers were substantially likely to be confused.”
Next, Jenzabar’s mark was strong, but in the context of a referential use that factor was less important. “An Internet user who runs a search even for an extremely strong trademark may be searching for information about the mark's holder rather than for the mark holder's official Web site. A search result leading to a third-party Web site that contains such information thus may not be confusing at all.” (In a footnote, the court noted that having one factor in Jenzabar’s favor was insufficient to defeat summary judgment.)
On intent, Jenzabar had zip, zero, nada in the way of evidence of intent to confuse, which is the only intent that matters. Its claim (which the dissent, preposterously, accepted) was that the use of the mark in the title/metatags was itself evidence of an intent to confuse. But evidence of intent to use was not enough. Even assuming that Long Bow’s intent was to improve its Google rank for Jenzabar searches, that wasn’t intent to confuse. “Given that Long Bow's site actually contains information about Jenzabar and does not sell goods that compete with Jenzabar, the most natural inference arising from Long Bow's use of the word ‘Jenzabar’ is not of intent to confuse but rather of intent to accurately describe the page's contents, and to bring it to the attention of Internet users searching for information about Jenzabar, who would find it relevant to their search.”
The undisputed evidence also cut against intent in that Long Bow used only the word Jenzabar, which was necessary to describe the site’s contents, without imitating Jenzabar’s logo or website. Likewise, Long Bow only used Jenzabar as a keyword for Jenzabar-relevant pages. (This is an incorporation of nominative fair use concepts into the modified multifactor confusion test the court is using; courts did and do that without using nominative fair use, but nominative fair use seems a lot cleaner and cheaper to me.) And when Jenzabar complained, Long Bow conscientiously posted Jenzabar’s objections on its site so visitors could judge for themselves, along with adding prominent disclaimers. It also added a metatag so that the description in Google now reads “Jenzabar has tried to censor this web page because it carries critical information about the software company that Chai Ling started with her husband.” There was no legally sufficient evidence to create a factual dispute on Long Bow’s intent.
Nor was there any evidence of actual confusion. There was evidence that a number of visitors reached Long Bow’s site via a Jenzabar search, but that’s not evidence of confusion. Quoting McCarthy, himself quoting Eric Goldman’s wisdom: “[I]t is improper to assume that using a trademarked keyword means that the searcher wanted to find the trademark owner.... Finding searcher 'diversion' is not possible until one knows where searchers were heading in the first place.” The court commented that “[t]his is especially true because, as Jenzabar concedes, the types of consumers most likely to be running a Google search for Jenzabar would be inclined to do ‘extensive due diligence’ on the company.”
Finally, the court examined what the Google search results actually looked like. Jenzabar always held the top result, so a searcher would find Jenzabar “exactly where she expected to.” “There would be no reason to continue down the list of search results unless she was looking for further information about Jenzabar, in which case the Long Bow site would be relevant and helpful rather than confusing.” In a footnote, the court said that this situation wouldn’t insulate against all confusion claims; for example, the result might be different if the parties directly competed (though presumably one could easily make clear that one is engaged in comparative advertising). But if the defendant’s site were the top result, that “might be suggestive of a likelihood of confusion even in a case such as this one, where the balance of factors otherwise tilts so strongly against such a finding.” Sigh. So I guess if the NYT does a story—whether critical or laudatory—about a small company, and the NYT’s Google juice puts it ahead of the company’s own site, it now has to worry about a trademark infringement claim? Another reason that a clean defense would be better.
In addition, the court commented, the Long Bow site’s listing “did little affirmatively to suggest official endorsement by Jenzabar. The description was an accurate summary of the page's contents, and was phrased in a neutral manner that did not suggest official sponsorship by Jenzabar.” The prominently displayed URL was “visibly different” from the official Jenzabar URLs that also showed up. “The only aspect of the listing that was even potentially confusing was its single-word title: ‘Jenzabar.’ Although Jenzabar may be correct that it is unusual for a Web page's title to be the unadorned trade name of an entity unaffiliated with the page, the title does little affirmatively to suggest official sponsorship by or affiliation with Jenzabar.” By contrast, the court suggested, a title like “Jenzabar—Official” would actively suggest official sponsorship as well as an intent to deceive.
The court returned to the topic of sophistication: internet users, especially the relevant consumers here, “understand that a Google search returns a list of pages of varying relevance from a wide range of sources.” Tabari says that they expect some uncertainty, and don’t form firm expectations about sponsorship until they hit the landing page, if then. (If this is true, then this case should have been dismissed before summary judgment, because at most Jenzabar could plead initial interest uncertainty, not initial interest confusion.) Thus, “an Internet user searching for Jenzabar's products who came across the listing for Long Bow's Web site (after already having found Jenzabar's official site) perhaps might be uncertain about this additional site's relationship to Jenzabar. However, Jenzabar has not shown how Long Bow would capitalize on any such uncertainty in a way that trades on Jenzabar's good will.” Any uncertainty was the result of “sensible agnosticism, not consumer confusion,” Tabari. Given the other factors strongly favoring Long Bow, such uncertainty couldn’t create a material factual dispute.
There was also no dilution, since Long Bow’s use couldn’t dilute the mark’s link in consumers’ minds to the mark’s source. “Instead, Long Bow's use actually tends to reinforce that quality, since Long Bow uses the mark only to refer to Jenzabar itself.” Any harm to Jenzabar comes from the critical content (and Jenzabar's own attempt to suppress critical speech), which isn’t a trademark harm. And there was no separate violation of Chapter 93A, since Jenzabar failed to explain what that could be other than trademark infringement.
There’s a dissent; it’s physically painful to read. Long Bow’s use apparently might confuse internet users to believe that Long Bow’s website “is somehow connected to, or affiliated with, Jenzabar.” Even if the Lanham Act reaches “somehow connected to,” the dissent doesn’t engage with the point that the Long Bow pages are about Jenzabar. The confusion was supposedly “crafted” through the “negative” omission of a Long Bow title and a “positive” singular reference to Jenzabar. “The trade name Jenzabar is thus embedded within an Internet link, which actually belongs to Long Bow, not Jenzabar. The confusing and misleading link diverts the Internet user to the secondary Long Bow site by trading on the Jenzabar name.” Ignoring the other parts of the search result, the dissent said that the search result “devised by Long Bow” (also untrue, since Google created it) “bears only the trademark ‘Jenzabar.’” Well, if you’re wrong on the law, might as well be wrong on the facts, I guess. (Next the court says Jenzabar was the “dominant” term in the result, which edges closer to reality.) Anyway, the search result didn’t say “Long Bow.” Thus, the use “trades on the name Jenzabar and, thereby, achieves immediate placement behind the trademark for Jenzabar and the Jenzabar primary Internet pages.”
Plus, Long Bow deliberately used Jenzabar’s trademark! In the HTML, no less, so that search engines would notice! This use of the “unadorned” name was evidence of intent to confuse. (Or, perhaps, intent to communicate?) The majority therefore erred by failing to weigh the strength of the Jenzabar mark in its favor (um, no, it didn’t, even though other courts have pointed out that it’s particularly easy to recognize critical uses of strong marks), by failing to recognize that the Google result was confusing, and by failing to hold Long Bow’s intent against it.
Maybe a sophisticated internet user could figure this out, once s/he read the webpage. But there’s already been initial interest confusion! This was obviously confusing as a matter of law. (Fortunately, in the real world, there’s a remand for Long Bow to seek its fees, which it should receive, since the Lanham Act does recognize certain anti-SLAPP principles.) It doesn’t matter that Long Bow had a public, political, or nonprofit motive and didn’t compete with Jenzabar.
Indeed, the dissent thought, “this case implicates the usability of the Internet.” (And there we do agree.) Long Bow gets to “trade on any Internet good will earned by Jenzabar to divert Internet users, either directly or indirectly, to the Long Bow Internet site.” (Since I’ve never quite understood what good will is, I will just express my extra uncertainty about what Internet good will is. I must say, this supposed diversion scheme reminds me of the internet meme, “Step 1, collect underpants, Step 2 ?, Step 3, profit.” Only without the profit.)
We don’t want people using SEO “when ambiguous title tags or poorly labeled text describe the search result link or the displayed advertisement, as is the case here…. This is wrong.” Mark Lemley says that many trademark owners think that the phrase “unfair competition” is redundant. Here we don’t even have the competition—a grin without a cat, as it were—but an equally powerful sense that it’s just wrong to show up in search results about a trademark just because you have something to say about the trademark owner. I don’t really know what to say to that intuition, because it seems so unreasonable to me. The thought that a trademark owner suffers actionable harm because someone might, even for a short time, see criticism of that trademark owner, seems so contrary to our system of free expression that I can only think that the incantations of “good will” and “confusion” cause people like the dissenting judge to lose sight of what trademark is for.