Rounding Up Three Recent Keyword Advertising Cases–Comphy v. Amazon & More

March 29, 2019 · by Eric Goldman · in E-Commerce, Marketing, Search Engines, Trademark

Three interesting recent keyword advertising cases:

Comphy Co. v. Amazon.com, Inc., 2019 WL 1128519 (W.D. Wash. March 12, 2019)

Comphy makes high-end linens. It chooses not to sell directly on Amazon. Nevertheless, Amazon has purchased search engine keyword ads triggered by the term “Comphy.” Amazon also auto-completes the characters “comph” with “comphy sheets,” “comphy company sheets,” “comphysheets,” and other unrelated terms. Searching for “Comphy” sheets on Amazon leads to competitive products, some of which are labeled “Comfy” or “Amazon’s Choice for ‘comphy sheets.’”

The court denies Comphy’s request for a preliminary injunction.

Trademark Scope. Comphy has a trademark registration for selling to “health spas.” Amazon sells direct to consumers, which is outside the scope of the trademark registration. Furthermore, the registration only applies to Comphy’s logo, instead of the word mark, which Amazon doesn’t use. Worse for Comphy, the court characterizes the Comphy word mark as descriptive and not commercially strong because the word mark isn’t heavily advertised to consumers.

Multi-Factor Likelihood of Consumer Confusion Test. The court engages the seminal question: are people searching for “Comphy” exclusively seeking the manufacturer? This has been the explicit or implicit premise of most keyword ad lawsuits, but it has numerous problems, including a lack of empirical support and the embedded inconsistency that brand-loyal searchers lose all brand loyalty when presented with competitive options.

The court says: “use of COMPHY as a search term does not necessarily indicate an intent to search for Plaintiff’s products. This is true even in the more limited situation where COMPHY is used in reference to bedding.” Amazon produced a survey showing that “thirteen-percent of consumers regarded COMPHY as referring to a particular source for bedding, but that more than twelve-percent of those consumers also regarded ‘comfort’—a non-existent company also presented—as also referring to a particular source for bedding.” Let’s hear it for survey controls!

The court gets at the issue another way, while discussing proximity. The court gives several hypotheticals of how consumers might use the Comphy mark to find linens on Amazon and still not be confused. The only countervailing example depends on whether consumers knew the Comphy brand before arriving: “If Defendant uses COMPHY to sell ‘Comfy Sheets’ to a customer who intends to buy ‘Comphy Sheets’ and has independently chosen to shop on Defendant’s site, there may or may not be confusion. Such a customer may have intended to buy ‘Comphy Sheets’ and elected instead to buy “Comfy Sheets’ because Defendant—the retailer the customer chose—only sold ‘Comfy Sheets.’ On the other hand, the customer may have intended to buy ‘Comphy Sheets,’ seen ‘Comfy Sheets’ available, and bought “Comfy Sheets” intending to receive ‘Comphy Sheets.’” The weakness in Comphy’s mark makes the latter scenario unlikely.

Comphy provided some compelling examples of actual confusion, but the court responds “some actual confusion may exist even if a reasonably prudent consumer would not be confused.”

A Troubling Footnote: “The Court notes again that its analysis may differ profoundly if the Court were to consider independently (1) Defendant’s use of keyword advertising to direct consumers to its website by placing ads for “Comphy Co Sheets” at Amazon or (2) Defendant’s use of the search term “comphy sheets” within its own website to offer consumers “comfy” brand sheets or other sheets marketed as comfy or comfortable. As noted previously, Plaintiff has elected to lump the contexts and uses together, and the Court declines the invitation to scour the limited record to craft a legally defensible injunction.”

Conclusion: The court concludes: “This is not an easy decision. Many of Defendant’s actions toe exceedingly narrow legal and equitable lines. But, on the record presented by Plaintiff, there are simply too many pitfalls for the Court to map an equitable course.”

Tichy v. Hyatt Hotels Corp., 2019 WL 1318674 (N.D. Ill. March 22, 2019)

This is an antitrust claim against major hotel chains. After 2015, allegedly:

The major hotel chains agreed to stop bidding on each other’s trademarks and in fact did so;
The major hotel chains agreed to impose restrictions in their online travel agency (OTA) agreements to stop bidding on their trademarks;
The hotels actually imposed these restrictions in their OTA agreements. As just one example, the complaint cites a November 2015 agreement between Marriott and Expedia with the following provision:

No Party will knowingly use in any way, including a bid in paid search engine advertising, any other Parties [sic] intellectual property, including the trademarks that are listed in Exhibit G and are owned by or licensed to Expedia or Marriott, as applicable (whether or not registered) for use in connection with hotel services. Expedia will list Marriott Marks listed in Exhibit G as well as its common misspelling as negative keywords as indicated in order to prevent its ads from appearing as a result of a search for those terms.

Allegedly, competitive keyword advertising effectively stopped in the industry after these developments. All of these outcomes, if true, sound like trouble after the FTC v. 1-800 Contacts ruling.

The plaintiffs don’t have direct evidence of a conspiracy, so they attempt to make a circumstantial showing. The changes in pre-2015/post-2015 keyword ad bidding, plus the proliferation of anti-competitive keyword advertising contract provisions in OTA contracts, is sufficient. In particular, the court notes the Prisoner’s Dilemma of competitive keyword advertising. Each individual hotel chain will get differential benefits if they are the only competitive keyword advertiser; and each individual hotel chain would lose business if they don’t use competitive keyword ads but their competitors do. This makes it fishy if all of them individually disarm. However, if all industry players engage in competitive keyword advertising, none get differential benefits (unless they are smarter about placing ads than their peers) and everyone ends up paying the arms dealer/search engine. Avoiding this “tax” gives financial incentive to collude.

The court also notes that it’s economically irrational to restrict OTAs from competitive keyword ads because they are a “second chance” to win the customer’s business. However, the court acknowledges that vertical ad restrictions aren’t as inherently anti-competitive as horizontal restrictions.

The defendants tried to offer some reasons why they would all independently decide to voluntarily restrict competitive keyword ads, including the fact that the ads cost money and that they would prefer to channel brand-loyal searchers away from OTAs and to their direct websites. The court responds:

the court finds Defendants’ alternative explanations unconvincing. First, Defendants examine each form of keyword advertising in isolation, ignoring how any Defendant’s unilateral choice to restrict only one form would leave it more vulnerable to competition from the other Defendants. Defendants also overemphasize consumers’ brand loyalty. Indeed, viewing the allegations in the light most favorable to Plaintiff, hotels’ willingness to spend money bidding on branded keywords—and search engines’ common practice of auctioning them—reasonably suggest that branded keyword advertising works. That is, even brand-interested hotel consumers are willing to (and do) book rooms from other hotels, depending on the advertisements they see.

Even if Defendants’ alternative explanations are plausible, they do not negate the plausibility of Plaintiff’s competing explanation: that Defendants would not have unilaterally restricted branded keyword advertising as alleged without a preceding agreement, because doing so would harm their economic interests. At the pleading stage, the court is not “to stack up inferences side by side and allow the case to go forward only if the plaintiff’s inferences seem more compelling than the opposing inferences.”…

Plaintiff does plausibly allege that each Defendant would have acted against its economic interest by unilaterally restricting branded advertising as alleged. Specifically, by ceasing to bid on other Defendants’ (or OTAs’) branded keywords, the independently-acting Defendant would sacrifice opportunities to win direct bookings and decrease the chance that a consumer searching for Defendant would visit its website. And by prohibiting OTAs from bidding on its keywords, the independently-acting Defendant would decrease the likelihood of winning indirect bookings, yet continue to face just as much advertising competition from other Defendants. Plaintiff’s allegations in this regard are, if not overwhelming, certainly sufficient to support an inference that Defendants reached a preceding agreement to restrict branded advertising. So, too, are her allegations that Defendants departed from an established advertising pattern, and that search results reflect the departure.

Citing Thompson v. 1800-Contacts, the court finds antitrust standing among consumers who possibly paid higher prices due to the restrictions on keyword ads:

Plaintiff has alleged that Defendants agreed to manipulate paid search results so that when a consumer queries an Internet search engine using a Defendant’s hotel name, only advertisements for that Defendant’s hotel appear at the top of the search results. In other words, according to Plaintiff, Defendants “interfer[ed] with the free flow of information from sellers to buyers” in the market for online hotel room bookings. Plaintiff alleges that the effect of Defendants’ conspiracy was to reduce “downward pricing pressure” from OTAs and competing hotel advertisements. According to Plaintiff, she and the putative class members paid more to find and book hotel rooms online “than they would [have] in the absence of [Defendants’] agreements.”

The court also finds proximate causation:

Plaintiff’s allegations reasonably permit the inference that hotel consumers’ brand loyalty is not absolute: consumers are willing to book hotels from different Defendants depending on the comparative information they see. Plaintiff’s allegations—including that many consumers begin their hotel searches by querying Internet search engines with branded keywords—also permit the inference that before 2015, a major form of competition among Defendants was the battle for advertising space associated with those searches. So although Plaintiff does not allege that OTAs “offered identical rooms at lower prices than” those on Defendants’ websites, her allegations nonetheless permit the inference that Defendants feel less pressure to price competitively now than they did before 2015, when consumers were more likely to see competing advertisements at the top of their branded searches. Just as in Thompson, therefore, the alleged advertising restrictions in this case could plausibly have caused consumers to pay higher prices for hotel rooms purchased online….

The court declines to credit the competing inference that Defendants offer: Plaintiff caused her own injuries by using ineffective search methods. Likewise, Plaintiff’s allegation that she navigated directly to Marriott.com to “save time” does not necessarily mean she did not incur higher search costs.

Just a reminder: the FTC v. 1-800 Contacts ruling put a lot of pressure on any attempts to restrict competitive keyword advertising. This ruling highlights those risks.

Snow Systems, Inc. v. Sneller’s Landscaping, LLC, 2019 WL 1317746 (N.D. Ill. March 22, 2019)

Sigh. This is a trademark battle over the putative trademark “SNOW SYSTEMS.” Really? The plaintiff handles snow and ice removal in Illinois, Indiana, Ohio, Wisconsin, Pennsylvania, Colorado, Massachusetts and Tennessee. The defendant, doing business as Sneller’s Snow Systems, provides similar services in Lansing and Grand Rapids, Michigan. Among other things, the plaintiff objects to the use of the term “snow systems” for keyword advertising. The plaintiff sued the defendant in Illinois.

This is an easy personal jurisdiction dismissal. The defendant had no clients outside Michigan and doesn’t market outside the state. The defendant’s “passive” (ugh) website doesn’t change the result; “the website is geared toward clients and potential clients in Michigan, not Illinois; it expressly indicates that it provides services in Grand Rapids and Lansing, Michigan, provides weather reports local to those areas alone, and includes company contact information local to the area.” The other Internet activities, such as a Facebook page and the keyword advertising, are immaterial because they were not directed towards Illinois residents.

More Posts About Keyword Advertising

Rounding Up Three Recent Keyword Advertising Cases–Comphy v. Amazon & More

Comments and Pings