The Initial Interest Confusion Doctrine Refuses to Die

Continuing my recent critical coverage of the initial interest confusion doctrine, here are a few more rulings on the subject.

Hoffmann Brothers Heating and Air Conditioning, Inc. v. Hoffmann Air Conditioning & Heating, LLC, 2025 WL 2587109 (8th Cir. Sept. 8, 2025)

The Eighth Circuit revisits the initial interest confusion doctrine after embracing it in 2021. Oh joy.

A family fallout produced two rival businesses: Hoffmann Brothers Heating and Air Conditioning, Inc., and Hoffmann Air Conditioning & Heating, LLC. Incredibly, the jury found no confusion betwen these two entities. The court addresses the post-trial motions:

Customer confusion can occur at various times. One of them is before a sale, what the law calls initial-interest confusion. For us, the theory is of recent vintage, although courts around the country have been applying it for about half a century.

Our variation of the rule, however, is different. When customers are “sophisticated[,] and exercise a relatively high degree of care in making their purchasing decisions,” it is “less likely” that they will experience “initial confusion.” Even if there is reason to question whether our “sophistication exception” should exist, there is no doubt it does. [cite to Select Comfort v. Baxter]…

Since Sensient, we have recognized that initial-interest confusion does “not apply” when customers are sophisticated and exercise a “high degree of care in making their purchasing decisions.”

Better to reject the IIC doctrine when consumers are sophisticated, I guess, but even better would be to reject it always.

Loanstreet Inc. v. Troia, 2025 WL 2530884 (S.D.N.Y. Sept. 3, 2025)

Prior blog post. I previously summarized the case:

Troia was a LoanStreet employee. He was allegedly fired for cause. Troia posted disparaging comments about LoanStreet at Glassdoor.com, Reddit.com, and Teamblind.com. He then worked to boost the posts’ visibility [including buying Google keyword ads]

The court says:

plaintiffs allege the existence of “initial-interest confusion,” in which one party creates initial interest in its competing product, service, or website, “even though no actual sale is finally completed as a result of the confusion.” [1-800 Contacts v. JAND]…In order to sufficiently demonstrate internet-related initial-interest confusion, the Second Circuit also requires a plaintiff to prove intentional deception by the defendant.

In a footnote, the court adds:

Troia also contends that plaintiffs have failed to produce any screenshot or description of their home page that would permit the Court to compare landing pages of plaintiff’s and defendant’s Google ads, to assist the Court in determining whether the Reddit page linked by the Google advertisements bears any resemblance to LoanStreet’s website or brand “look and feel.” This is also irrelevant, as plaintiffs allege only initial interest confusion and do not base their unfair competition claims on Troia’s Reddit posts, only his advertisements.

If you’re a trademark plaintiff and you’re only alleging initial interest confusion, you should question your life choices.

With respect to Troia’s keyword ad buys, the court distinguishes the JAND case:

Troia used LoanStreet’s mark repeatedly: at least once in bold in the heading of the advertisement; sometimes in the body of the advertisement; and in the URL he linked. As an initial matter, plaintiffs assert that the question of whether these advertisements were “clearly labeled” as advertisements is fact-intensive and should be evaluated by the jury. Given defendant’s repeated and verbatim use of the LoanStreet mark, we agree. Despite the fact that defendant’s advertisements contained language critical of LoanStreet, the advertisements clearly displayed plaintiff’s mark, referring in some cases to “LoanStreet Careers” or “LoanStreet Software Engineer.” Moreover, as plaintiffs’ expert noted in his report, “anyone doing a search for LoanStreet’s website would expect that website to appear as the first search result and would be confused by [d]efendant’s ad appearing at the top of the page with the same branded, trademarked name as the company website.”…

I question the expertise of that so-called “expert” who claims that “anyone doing a search for LoanStreet’s website would expect that website to appear as the first search result,” especially given that we’re talking about ads and not organic search results.

The court also drops this jaw-dropper:

although Troia did not offer a competing good or service, he cannot truthfully assert that he did not intend to attract Google users searching for the real LoanStreet when he purchased his Google advertisements….These statements are sufficient to demonstrate that Troia sought to compete with LoanStreet for users within the same market – the internet and, in particular the Google search engine context

This is seriously stupid. Gripers and their targets are not commercially competing against each other. I don’t understand how judges can get so lost in the weeds that they miss this obvious fact. It reminds me of one of my least favorite trademark law cases of this century, the SMJ case from twenty years ago.

The judge also twists the nominative use defense against Troia. Crazy. That sends the case to an expensive jury trial that neither side should want. Then again, the company plaintiff might be OK with the lawfare of grinding its griping former employee into a financial pulp.

Kyjen Co. v. Schedule A Defendants, 2025 WL 2696439 (N.D. Ill. Sept. 22, 2025)

This case involves slow feeder pet bowls. Kyjen, also known as Outward Hound, has a registered trademark in “fun feeder.” Its rival, ThinkPet, included the words “fun feeder” in its Amazon product listings and otherwise calls its products “fun” for the pets. Kyjen sued ThinkPet for trademark infringement. The court rejects ThinkPet’s motion for summary judgment.

To understand this case, you have to revisit the Seventh Circuit’s Promatek v. Equitrac case from 2002, a keyword metatag case that turned on initial interest confusion, even though the Seventh Circuit couldn’t define the doctrine then and still can’t now. I still teach that ancient case in my Internet Law course….but its primary pedagogical payoff is to show my students how everyone involved in the case–the litigants, the lawyers, and the court–all got it embarrassing wrong. As a legal ruling, the case has aged terribly, and it’s incredible that it’s still an acceptable citation in 2025.

Fair Use. The court says “a reasonable juror could find that ThinkPet did not use “Fun Feeder” in good faith for only descriptive purposes.” If the court had stopped there, OK, but the court kept talking:

One way to demonstrate that a defendant is not using a trademarked term merely for a description is if the defendant’s use creates “initial interest confusion.” [Cite to Promatek v. Equitrac.]

What? This makes no sense. IIC adds nothing to the analysis of whether the defendant is using the trademark for descriptive purposes or to confuse consumers.

Following the Promatek “inspiration,” the court says:

Intending to confuse a consumer, even only temporarily, prevents a defendant from claiming the fair use defense because the defendant is using the mark for a purpose other than describing their product.

This might be true if in fact the court has proof of the defendant’s “intent” as distinguished from the defendant’s desire to engage in standard marketing, i.e., using the word “fun” to describe its “feeder” product (an outcome that the descriptive fair use doctrine fully permits). But if a court is citing Promatek in 2025, it’s going to make circumstantial inferences about the defendant’s “intent” to confuse consumers, thus bootstrapping plaintiffs from needing to show the requisite evidence.

The court continues:

Initial interest confusion is particularly relevant in the online space, as here.

FTFY: “Initial interest confusion is never relevant, in the online space or off.”

The court gets to its doctrinal payoff:

The facts in the instant case parallel Promatek. Kyjen alleges ThinkPet added “Fun Feeder” to its Amazon listings to increase its products’ visibility and, therefore, intended to benefit from Kyjen’s goodwill through initial interest confusion. Just as Equitrac could violate the Lanham Act by using the metatags to cause initial customer confusion and profit off the plaintiff’s goods will, ThinkPet could violate the Lanham Act by using Kyjen’s trademark for SEO that increased its visibility and benefit off Kyjen’s goodwill. If true, ThinkPet would be using “Fun Feeder” for non-descriptive purposes and could not rely on the fair use defense.

Cue the sad trombone whomp-whomp. I mean…what? Some of the obvious problems with the court’s discourse:

• The keyword metatags at issue in the Promatek case weren’t visible to consumers at all, unlike Amazon product descriptions.
• Trademark law may distinguish between Google search engine indexing and internal search listings in Amazon–see MTM v. Amazon.
• Being indexed in Google for a third-party trademark is in fact usually entirely permissible under trademark law.
• The initial interest confusion doctrine adds nothing to this equation–if the “fun feeder” reference didn’t qualify as descriptive fair use, then this is just a standard trademark infringement case.
• I question if ThinkPet was using the term “as a mark.” The court later addresses this issue by saying the jury could determine that the fun feeder term was “used to gain visibility off of Kyjen’s goodwill.”

In other words, this ruling is messed up on pretty much every dimension. Terrible.

The court says that the court can infer that the “fun feeder” mark is strong based on Kyjen’s high prominence in search results (sadly, no citation to Lisa Larrimore Ouellette’s paper):

Outward Hound appears in search results for “fun feeder bowl” across various search engines. On Google, it appears as the first sponsored product and its products are referenced in the second through seventh search results. On Bing, Outward Hound is the first search result. Finally, on Amazon, Kyjen products are the first two non-sponsored results

Um…saying a mark is strong because of prominent placement in SPONSORED listings…seriously? I guess trademark owners can truly buy their way to the top.

The Likelihood of Consumer Confusion Test

The court turns to the multi-factor consumer confusion test. With respect to purchaser care, the court says:

In Promatek, the Seventh Circuit found that lack of care exercised by consumers was the most important factor in an initial interest confusion case because the consumer’s lack of care leads to the initial interest confusion. Therefore, the fact that a reasonable jury may find consumers exercise a low degree of care in selecting dog bowls is particularly relevant.

Contrast the Lerner & Rowe case, where the court said that if consumers care enough to search by brand name, they will be careful about evaluating the results.

With respect to evidence of actual confusion, the court says per Promatek “a showing of actual confusion was not required for the court to find a likelihood of success on the merits in an initial interest confusion case.” In other word, pleading IIC is a cheat code for plaintiffs to win more factors in the multi-factor likelihood of consumer confusion test.

This is a bad ruling, but we can extend the blame to the antiquated and wrong-headed Promatek case. #OverturnPromatek.