How Consumer Sophistication and Niche IDs Can Overcome a 2(d) Refusal

MERMAID // MERMADE

How Consumer Sophistication and Niche IDs Can Overcome a 2(d) Refusal
Photo by Daniel Bernard / Unsplash
• In re TDI Novus Inc., No. 98593408 (TTAB May 12, 2026) • [Not a Precedent ] • case link
Core Issue: Likelihood of confusion analysis under DuPont factors, specifically focusing on the distinction between broad "apparatus" and specialized "systems" and the impact on consumer sophistication and trade channels.

In this appeal, we see a showdown between an Examining Attorney’s reliance on a broadly worded description and an Applicant’s specialized niche product. The Board’s decision to reverse the Section 2(d) refusal reminds us that while similarity of the marks is a heavyweight DuPont factor, it is not an insurmountable hurdle. Success is entirely possible if you can show that the niche goods not only differ from the broader definition, but also demand a high degree of care during the selection process.

Background

The Applicant, TDI Novus Inc., filed Application Serial No. 98593408, seeking registration on the Principal Register for its stylized word and design mark.

The mark consists of the word "MERMAID" in a specific stylized format, featuring a capitalized letter "A" amidst lowercase letters and a wireless signal icon above the "I." The application identifies the goods in International Class 9 as a voice communication system specifically designed for autonomous maritime vehicles. This system is comprised of an apparatus to record, transcribe, translate, and parse voice communications over the radio frequency spectrum, while autonomously providing system-generated voice responses appropriate to the context of the transmission.

The Trademark Examining Attorney refused registration under Section 2(d) of the Trademark Act, citing a likelihood of confusion with MERMADE (Registration No. 6285548). The cited registration covers a broad range of goods and services in International Classes 9, 35, and 41, including "apparatus for recording, transmission or reproduction of sound and images."

Following a final refusal and the denial of its request for reconsideration, the Applicant appealed.

How the Board Analyzed the DuPont Factors

Before diving into the DuPont factors, the Board addressed a procedural argument raised by the Applicant. The Applicant submitted that its companion application for the standard character mark "MERMAID" (Serial No. 98593388) had already been approved and published for opposition despite the existence of the cited MERMADE registration. Consequently, the Applicant argued that consistency in examination should dictate the approval of this stylized version of the mark.

The Examining Attorney countered that prior decisions by other examiners carry little evidentiary value and are not binding. He also pointed out that the Applicant had failed to properly introduce the companion application into the record, merely listing the serial number and an email notification in its appeal brief.

Interestingly, the Board noted that:

"... TMEP §702.03(a) discusses companion applications and provides that generally a single examining attorney should be responsible for their examination for registration. The TMEP further provides that that in cases where multiple examining attorneys are involved each “examining attorney must review the electronic record of the earlier companion application before taking action in a later companion case, and should act consistently, unless it would be clear error to do so.” TMEP §702.03(a)(ii)."

Nevertheless, the Board refused to consider the companion application's allowance, strictly adhering to Trademark Rule 2.142(d). Because the Applicant did not submit an official copy of the companion application or request a remand to introduce this evidence during the prosecution phase, the companion application and its prosecution history were not part of the evidentiary record on appeal.

DuPont #1: Similarity of the Marks

The Board analyzed the marks in their entireties, considering appearance, sound, connotation, and commercial impression.

Sound & Appearance

Regarding sound, the Board determined that the marks are likely to be pronounced in the same way. In terms of appearance, the Board acknowledged notable differences, such as the Applicant's varied spelling and the inclusion of the wireless signal icon. However, the Board ultimately found that the marks remain visually close because they share their first five letters, and consumers are likely to perceive the registered mark MERMADE as a novel spelling of MERMAID.

Connotation

The Applicant argued that the stylized "A" and the wireless signal icon were integral communicative elements suggesting technological advancement and artificial intelligence. While the Board agreed that the emphasized "AI" and the icon might suggest artificial intelligence and internet connectivity, it nonetheless concluded that consumers would still primarily understand the mark as conjuring the mythical sea creature. Because both marks ultimately project a highly similar commercial impression centered on a "mermaid," the Board found them to be more similar than dissimilar, weighing this factor in favor of a likelihood of confusion.

DuPont #2: Relatedness of the Goods

The Examining Attorney argued that the goods were legally identical. His position was that the registrant's broad description, "apparatus for recording and transmission of sound", presumably encompassed the Applicant's narrower description, "apparatus for recording and transmission of voice communications to maritime vehicles," electing to ignore the critical fact that this apparatus is just one component of a much larger voice communication system.

The Board disagreed with this overly broad interpretation. While acknowledging the Examining Attorney's logic, the Board highlighted that the Applicant's goods are defined as an entire "voice communication system" that does significantly more than a mere recording and transmission apparatus. Specifically, the application describes an AI-driven system that records, transcribes, translates, and parses voice communications, and then autonomously provides a system-generated response tailored to the context of the transmission.

To better understand the technical nature of these goods, the Board relied on extrinsic evidence submitted by the Applicant. This evidence corroborated the Applicant's claim that its goods are specialized hardware systems designed for autonomous maritime vehicles, targeting a highly niche market of maritime industry professionals. Ultimately, the Board found that the Applicant's sophisticated technology is distinct from the general audio apparatus covered by the cited registration. Consequently, the Board determined that the goods were not related, weighing this DuPont factor against a finding of a likelihood of confusion.

DuPont #4: Consumer Care

The Applicant argued that its goods are sold exclusively to specialized entities, such as maritime engineering firms, defense contractors, and autonomous vessel operators, primarily through highly controlled bidding and RFP processes.

While the Board noted that the Applicant failed to submit formal evidence such as declarations or affidavits to prove this specific purchasing process, the Board nevertheless found that the intrinsic nature of the goods spoke for itself. The Board reasoned that the consumers of such specialized maritime systems would be discriminating professionals or equally sophisticated purchasing agents who possess specialized knowledge of the autonomous marine vehicle industry.

The Board emphasized that this is a niche class of consumers who would exercise a far greater than ordinary degree of care, given the technical nature of the systems and the significant inherent safety concerns involved in operating autonomous marine vehicles. Because these purchasers are likely to be highly sophisticated and careful, the fourth DuPont factor was found to weigh against a likelihood of confusion.

Board’s Decision

The Trademark Trial and Appeal Board reversed the refusal to register the mark under Section 2(d). The Board concluded that while the first DuPont factor favored a finding of likelihood of confusion due to the similarity of the marks MERMAID and MERMADE in sound and connotation, this similarity was ultimately outweighed by other factors. Specifically, the Board determined that the second and fourth DuPont factors supported a finding that confusion is unlikely.

The Board reversed the refusal to register the mark under Section 2(d). It concluded that while the first DuPont factor favored a finding of a likelihood of confusion due to the similarity in appearance, sound and connotation between the marks MERMAID and MERMADE, this similarity was outweighed by other critical factors. Specifically, the Board determined that the second and fourth DuPont factors strongly supported a finding that confusion is unlikely.


This case is a fascinating case study in how the specific mechanics of a goods description can completely change the trajectory of an appeal. It also got me thinking about how to advise clients in a changing fee landscape. Here are my main takeaways from the Board’s decision, and a few things I’m wrestling with as I look ahead.

1. Functional Specificity vs. The $200 Surcharge: A New Tactical Calculus

The Applicant’s success here turned entirely on their highly specialized identification of goods. By explicitly describing an AI-driven, context-aware "voice communication system" for autonomous maritime vehicles, they gave the Board the ammunition it needed to dismantle the Examining Attorney's broad, categorical interpretation.

In the past, when writing identification of goods and services, I would include both broadly worded ones, as well as technical detail entries. But with the USPTO’s fee structure penalizing free-form, customized identifications with an extra $200 surcharge per class, applied across all active classes even if you customize just one, I think my strategy needs to be updated.

For clients with deep pockets, the investment in a custom, bulletproof ID at the outset is still the gold standard. But for budget-conscious clients, we have to get creative. One option is the "wait and see" approach. You can file the application using standard, pre-approved terms from the TM ID Manual to lock in the lower base fee. If a Section 2(d) refusal arises during prosecution, you then amend the standard language to a highly tailored, custom description to narrow the scope and differentiate. This keeps initial government fees low, though you still need to pay the fee if you switch to free form later during amendment.

Alternatively, you can choose the division route by isolating the cutting-edge, custom-ID goods into their own single-class application, while grouping the more standard goods into a separate filing. This protects the standard classes from being hit with the blanket $200 surcharge. The trade-off here is purely mathematical: you are trading the cost of the custom-text surcharge for the cost of a completely separate application fee (attorney's fee), alongside the administrative reality of managing multiple applications. Ultimately, there is no single right answer. It comes down to each client's specific needs.

2. When the Board Bails You Out: The Danger of relying on "Intrinsic Nature"

I think the Applicant caught a lucky break on the fourth DuPont factor (conditions of sale). They failed to submit formal evidence, like declarations or affidavits, detailing their specific RFP and bidding processes.

Instead, the Board threw them a lifeline, ruling that the "intrinsic nature" of autonomous marine vehicle systems naturally demands a hyper-sophisticated buyer exercising extreme care. It’s a great reminder that common sense can still win the day, but it’s a massive gamble. Relying on the Board to make a unilateral finding on market reality is a luxury we shouldn't count on. We should all take the time to build the record with formal evidence during prosecution.


What about you? How are you navigating the custom ID surcharge with your clients? Are you leaning more on the ID Manual, or are you finding that custom text is still worth the premium up front?