SNAP INC v. VIDAL USPTO (2024)

MEMORANDUM OPINION AND ORDER

Snap Inc., owner of the social media application Snapchat, seeks to register SPECTACLES as a trademark for its eye-wearable technology product that it has been selling in evolving versions since 2016. The U.S. Patent and Trademark Office (PTO) denied Snap's registration applications, after which the Trademark Trial and Appeal Board (TTAB) affirmed the PTO's decision. The PTO found that “spectacles” is a generic product name for “smart glasses” and thus unprotectable as a trademark, much less registrable under the Lanham Act. Alternatively, the PTO found that “spectacles,” if not generic for smart glasses, is still unregistrable as a trademark because it is merely descriptive of the product with inadequate evidence of secondary meaning. And in finding the term to be descriptive, the PTO rejected Snap's claim that “spectacles” is suggestive and thus registrable as an inherently distinctive mark with no need to establish secondary meaning.

Rather than appeal the TTAB's ruling to the Federal Circuit based on the administrative record, Snap elected to pursue de novo civil proceedings against the PTO under 15 U.S.C. § 1071(b). As a result, the parties developed more evidence, including dueling expert opinions and consumer surveys. They presented that evidence at a bench trial where the parties’ experts and a former Snap employee testified. When proceeding under § 1071(b), the court “must make de novo factual findings that take account of both the new evidence and the administrative record before the PTO.” Kappos v. Hyatt, 566 U.S. 431, 446 (2012); see B&B Hardware, Inc. v. Hargis Indus., Inc., 575 U.S. 138, 144 (2015). Having now judged the credibility of the parties’ witnesses and reviewed all trial exhibits including the administrative record, the court details here its findings of fact and conclusions of law under Federal Rule of Civil Procedure 52(a).

I.

Snap began selling its eye-wearable technology under the mark SPECTACLES in 2016. (A0366).¹ That was the year it released Spectacles 1 with a single camera that allowed users to take photos or videos and later add filters in the Snapchat application. (Chan, Vol. I at 23-24, 63-64). Two years later, Snap released Spectacles 2, which added water resistance and different eyewear styles. (Id. at 25-26). In 2019, Snap added a second camera and released Spectacles 3. (Id. at 26-27). The most recent version of Snap's product, Spectacles 4, was released in 2021 and is offered only to content creators by invitation. (Id. at 27-28). This version has an augmented reality display permitting users to see virtual effects created with Snap's trademarked LENSES software. (Id.). Depicted in the images below are the Spectacles 3 (on the left) and the Spectacles 4 (on the right). (EX 75-4, 106-1).

Snap's registration applications claimed SPECTACLES, both as a word and stylized mark, for use with the following goods:

Computer hardware; computer peripherals; wearable computer hardware; wearable computer peripherals; computer hardware and peripherals for remotely accessing, capturing, transmitting and displaying pictures, video, audio and data; downloadable computer software, namely, software for setting up, configuring, and controlling wearable computer hardware and peripherals; downloadable computer software and software applications for use in uploading, downloading, capturing, editing, storing, distributing and sharing photographic and video content and other digital data via global and local computer networks and via mobile devices; downloadable multimedia files containing digital audio and video files featuring user generated images, videos, multimedia files, and other digital data, all in the fields of entertainment, photography and online social networking; computer software for accessing and transmitting data and content among consumer electronics devices and displays.

(A2099-2100; A4296-297). On its face, this recital of goods is not limited to only eye-wearable technology products but includes any wearable “computer hardware and peripherals for remotely accessing, capturing, transmitting and displaying pictures, video, audio and data.”

With its applications, however, Snap also submitted (as shown below) specimen photographs of its smart glasses product as sold in its typical eyewear form along with its protective case.² (A0012-13; A2205-06).

As a result, the parties agreed below—and continue to stipulate now—that the relevant product category based on Snap's applications and its product specimens is accurately and adequately captured as “smart glasses.” (A2110-111; ECF 143 at 40; ECF 144 at 24). That stipulation is intended to cover both the computer hardware features of Snap's product and its visibly prominent eyewear form factor. (ECF 143 at 40-41; ECF 144 at 24). The parties also agree that relevant consumers are average buyers of wearable technology goods, including smart glasses, sold by Snap and its competitors. (ECF 143 at 35, 58; ECF 144 at 24-25).

II.

The purpose of trademarks is to “indicate the source” of the goods or services to which they are applied, so that consumers can “identify and distinguish” among the goods or services of one producer from another.³ 15 U.S.C. § 1127. Trademark law thus classifies marks according to their source-identifying significance along a continuum of inherent and acquired distinctiveness. See Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768 (1992); In re Merrill Lynch, Pierce, Fenner, & Smith, 828 F.2d 1567, 1569 (Fed. Cir. 1987). At one end of that continuum, arbitrary or fanciful marks are considered the most inherently distinctive and enjoy the strongest trademark protections. See Elliott v. Google, Inc., 860 F.3d 1151, 1155 (9th Cir. 2017). At the other end, generic names “denominate the general category, type, or class of the goods, services, or business with which it is used.” Restatement (Third) of Unfair Competition § 15 (1995). By definition, a generic name for a product or service is incapable of indicating the source of—and thus helping consumers distinguish among—competing products or services in the market. See Park ‘N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189, 193-94 (1985). As a result, the “generic name of a product—what it is—can never serve as a trademark.” Official Airlines Guides, Inc. v. Goss, 6 F.3d 1385, 1391 (9th Cir. 1993) (citing Surgicenters of Am., Inc. v. Med. Dental Surgeries, Co., 601 F.2d 1011, 1014 (9th Cir. 1979)).

Descriptive and suggestive marks lie between generic names and arbitrary or fanciful marks on the distinctiveness spectrum. See Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1141 (9th Cir. 2002). A mark is descriptive if it “immediately conveys knowledge of a quality, feature, function, or characteristic of the goods or services with which it is used.” In re Bayer Aktiengesellschaft, 488 F.3d 960, 963 (Fed. Cir. 2007) (citing In re Gyulay, 820 F.2d 1216, 1217 (Fed. Cir. 1987)). Like generic names, descriptive marks “do not initially satisfy the distinctiveness element” required for trademark protection. Kendall-Jackson Winery, Ltd. v. E. & J. Gallo Winery, 150 F.3d 1042, 1047 (9th Cir. 1998). But unlike generic names, descriptive marks “can acquire distinctiveness if the public comes to associate the mark with a specific source.” Id. That acquired distinctiveness is called “secondary meaning.” AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 349 (9th Cir. 1979), abrogated on other grounds by Mattel, Inc. v. Walking Mountains Prods., 353 F.3d 792 (9th Cir. 2003). A suggestive mark, by contrast, is one for which a “consumer must use imagination or any type of multistage reasoning to understand the mark's significance.” Kendall-Jackson, 150 F.3d at 1047 n.8. Though weaker than arbitrary or fanciful marks, suggestive marks are still considered inherently distinctive and thus protectable as a trademark with no proof of secondary meaning. See Zobmondo Ent., LLC v. Falls Media, LLC, 602 F.3d 1108, 1113 (9th Cir. 2010).

III.

The mark SPECTACLES cannot be trademarked if the “primary significance” of that word “to the relevant public” is a “generic name for the goods or services” at issue. 15 U.S.C. § 1064(3). The primary significance inquiry is a question of fact typically answered in two steps. See Real Foods Pty Ltd. v. Frito-Lay N. Am., Inc., 906 F.3d 965, 981 (Fed. Cir. 2018). “The first step ․ is to determine with specificity the genus of goods and services for which the designation is sought to be registered.” 2 McCarthy § 12:57. In the second step, the factfinder must determine how “members of the relevant public primarily use or understand the term sought to be protected” as applied to the relevant genus of goods or services defined in the first step. H. Marvin Ginn Corp. v. Int'l Ass'n of Fire Chiefs, Inc., 782 F.2d 987, 989-90 (Fed. Cir. 1986). The PTO bears the burden of proving by clear and convincing evidence that relevant consumers of smart glasses perceive SPECTACLES as a generic name for the product.⁴ See In re Cordua Rests., Inc., 823 F.3d 594, 600 (Fed. Cir. 2016).

A. The Product Category of Smart Glasses Embodies Both Conventional Eyewear and Technological Hardware.

Whether a term is generic is not determined in the abstract but only as applied to a specific class of products or services. See Elliott, 860 F.3d at 1157 (“Trademark law recognizes that a term may be unprotectable with regard to one type of good, and protectable with regard to another type of good.”). The same word, after all, can be generic for one class of goods (e.g., “apple” for fruit) but a trademark for another (e.g., APPLE computers). See 2 McCarthy § 12:1. At the same time, a product or service “may fall within more than one ‘genus’ ” since “[t]here can be layers of genera in many fields.” Anthony L. Fletcher, Separating Descriptive Sheep from Generic Goats, 103 Trademark Rep. 487, 488 (2013).⁵ So just because a product in one category may be “related or similar” to that of another does not mean that a generic name of one is a generic name for the other. 2 McCarthy § 12:24; see 1 Gilson § 2.03[2][b] (A mark “may convey immediately that the goods or services are whatever the generic term is, without veering entirely into being generic itself.”). Thus, defining the relevant product genus with specificity includes precisely identifying the genus at the right level of generality.

Ostensibly, the parties agree that “smart glasses” accurately captures the relevant product genus. That stipulation is meant to hold an important concession from each side. For Snap, it concedes that while its applications recite a broad field of wearable computer hardware, the goods for which it seeks registration is narrower: only the eye-wearable computer peripherals Snap sells in commerce. For the PTO, the word “smart” added to glasses is (supposedly) intended to recognize the computer features of Snap's product, differentiating it from the higher-level category of general eyewear that would subsume smart glasses. According to the parties’ stipulation, then, “smart glasses” as the relevant genus embodies a product with two presumed essential aspects in the minds of relevant purchasers: the smart glasses’ prominently visible eyewear design and their less visible but instrumental embedded computer technology.

Even so, while professing to agree that eyewear is not the product type for which Snap seeks registration, the PTO makes a “subtle rhetorical move that attempts to abstract [the relevant product] to a higher level of generality” equivalent—in all but name—to the unclaimed category of eyewear. Nat'l Nonwovens, Inc. v. Consumer Prods. Enters., Inc., 397 F. Supp. 2d 245, 252 (D. Mass. 2005). Left unquestioned, that maneuver rigs the genus inquiry (at step one) to effectively preordain a finding of genericness (at step two). See A.J. Canfield Co. v. Honickman, 808 F.2d 291, 301 (3d Cir. 1986) (Once a relevant genus is defined at step one, the “resulting question” in step two “often decides itself.”). “By expanding or contracting the definition of a ‘genus’ of goods or services, a court can substantially affect the final determination of whether a term is a ‘generic name.’ ” 2 McCarthy § 12.23.⁶

According to the PTO, its rhetorical reengineering of the product genus is permitted by the Federal Circuit's decision in Cordua, which says that “a term is generic if the relevant public understands it to refer to a key aspect of the genus of goods or services in question.” 823 F.3d at 604 n.5 (emphasis added). By its reading of Cordua, the PTO says that the “key aspect” of a product can be virtually “any” distinctive attribute. (ECF 144 at 26; ECF 152 at 39). It then presupposes that a key aspect of smart glasses is its “glasses” form factor, draining away any “smart” attributes from the product with no principled explanation. And based on that assumption, the PTO claims that if consumers understand “spectacles” to be a generic reference to eyewear overall, then Cordua compels only one conclusion: SPECTACLES is a generic product name for smart glasses—no matter its distinctive computing features.

Cordua does not sanction that gambit. If it did, the case would erase any legal difference between descriptive terms and generic names. See 2 McCarthy §§ 12:1, 12:10, 12:20; see also 1 Gilson § 2.02[6][a] (questioning Federal Circuit's apparent test that a term need only “refer to” some aspect of goods to be generic since that “arguably sweeps in descriptive terms as well”). It would resurrect a trademark notion in circulation some two or three decades ago that has since been discredited. Compare 2 McCarthy § 12.02[5] (3d ed. 1996) (citing cases for the idea that “[t]o be generic, a term need not directly name the product, but may name some distinctive characteristic of that genus of products”), with 2 McCarthy § 12:10 (5th ed. 2017) (citing same cases from 1996 edition but interpreting them to say “that a word that identifies some distinctive characteristic of a product should be classified as a descriptive term that requires a secondary meaning”). And it would contravene the text of the Lanham Act, which “does not ask if [a] word ‘refers to’ a product or service,” but whether “the designation is a name of the product or service.” 2 McCarthy §§ 12:1, 12:20; see U.S. Pat. & Trademark Off. v. Booking.com B.V., 591 U.S. 549, 556 (2020) (“[A] ‘generic’ term names a ‘class’ of goods or services, rather than any particular feature or exemplification of the class.”).

Properly read, Cordua applies only when an applicant seeking to register a mark for a broad category of goods or services tries to avoid a genericness finding by arguing either (i) that the mark names only a subcategory of the claimed goods or services, or (ii) that it names merely a product when registration is sought for services alone. In Cordua, the applicant wanted to register a common Spanish name for specialty barbequed meat—“churrascos”—for the “broad genus of all restaurant services.” 823 F.3d at 605. To avoid a finding of genericness, the applicant argued that “churrascos” referred only “to a style of grilling meat and not to restaurant services.” Id. at 603. And even if it did, the applicant argued it was not generic for “all” restaurant services, but only a subtype serving a specialty grilled meat. Id. at 604-05. Cordua rejected both arguments by applying settled trademark principles.⁷ See generally 2 McCarthy § 12:57 (“registration is properly refused if the word is the generic name of any of the goods or services for which registration is sought”); id. § 12.24 (“generic name of goods may also be a generic name of the service of selling or designing those goods”).

Cordua rejected the applicant's first argument because a “term which is the generic name of a particular category of goods is likewise generic for any services which are directed to or focused on that class of goods.” 823 F.3d at 604 (quoting In re CyberFinancial.Net, Inc., 65 U.S.P.Q.2d 1789, 1791, 2002 WL 1980117, at* 3 (T.T.A.B. 2002)). So when a “key good” is “an integral, if not the paramount, aspect” of the services for which registration is sought—as was the case in Cordua—the generic name of that good cannot be registered as a mark for the claimed services. Id. at 603-04 (quoting Reed Elsevier, 482 F.3d at 1379); see, e.g., In re Tires, Tires, Tires, Inc., 94 U.S.P.Q.2d 1153, 1157, 2009 WL 4075360, at *5 (T.T.A.B. 2009) (TIRES TIRES TIRES is generic for retail tire sales). That is why “churrascos” was unregistrable for the claimed restaurant services even though the term was technically just the generic name for the specialty food that would be served. Cordua, 823 F.3d at 604.

Cordua rejected the applicant's second argument because a proposed mark can still be generic if “the relevant public understands the term to refer to part of the claimed genus of goods or services, even if the public does not understand the term to refer to the broad genus as a whole.” Id. at 605. So, for instance, “the term ‘pizzeria’ would be generic for restaurant services”—when such broad services are claimed for registration—even though the public understands the term “to refer to a particular sub-group or type of restaurant rather than to all restaurants.” Royal Crown Co. v. The Coca-Cola Co., 892 F.3d 1358, 1367-68 (Fed. Cir. 2018) (applying Cordua to subcategory of low-calorie drinks subsumed by broader claimed category of sports beverages). So too, then, was “churrascos” as a specialty restaurant name found to be generic when applied to the all-inclusive category of restaurant services—even if the applicant had no plans to open other types of restaurants focused on dishes other than the specialty grilled meat. See Cordua, 825 F.3d at 605.

In short, when Cordua stated that “a term is generic if the relevant public understands it to refer to a key aspect of the genus of goods or services in question,” it was not announcing a novel rule but applying settled trademark principles to the applicant's arguments in that case. Neither of those background rules, though, applies here. No one claims that Snap is trying to register SPECTACLES as a service mark for the broad category of eyewear sales or that its business is concentrated on selling general eyewear. If it were trying to register SPECTACLES for such a wide class of services, then Snap could no more register the mark for eyewear sales than a churrascos restaurant could trademark “churrascos” for restaurant services. And unlike the applicant in Cordua, who sought registration of the contested mark for “all” restaurant services, no one contends that Snap is trying to register SPECTACLES for a similarly capacious product class of “all” eyewear. If it were, everyone agrees that the mark would be generic for that claimed product category—even if Snap intended to use the mark in commerce only for a subcategory of smart glasses.

Thus, nothing in Cordua permits the PTO to substitute an overarching but unclaimed product category as the relevant genus for determining the primary significance of the mark in question. See Snap Inc. v. Vidal, ––– F. Supp. 3d ––––, 2024 WL 1136407, at *4 (C.D. Cal. Mar. 4, 2024); cf. Coach/Braunsdorf Affinity, Inc. v. 12 Interactive, LLC, 110 U.S.P.Q.2d 1458, 2014 WL 1390528, at *11 (T.T.A.B. 2014) (“neither PERKS nor PERKSCARD” was generic for genus of “volume discount buying services” because “while a volume [ ] discount buying service may be a ‘perk,’ not all volume discount buying services are ‘perks’ ”). Nor can Cordua be interpreted, as the PTO would like, to mean that a term can be generic—rather than merely descriptive—when all it does is “refer to” a product attribute.⁸ 823 F.3d at 604 n.5.

That said, even if the PTO's legal interpretation of the Cordua rule were correct, the agency has presented no evidence for the factual predicate needed to make the rule work on its own terms: that the eyewear form of smart glasses is—to the exclusion of the product's computing features—its “key aspect.” When pressed, the PTO conceded that a product's key aspect must have some “primacy” over other material aspects (or else, how could it be “key”?). (ECF 152 at 31-33, 39, 63-68). But then it was at a loss to cite evidence for the supposition that the eyewear form of smart glasses renders their computer functionality obsolete for trademark analysis. Contra, e.g., McZeal v. Amazon Servs., LLC, 2021 WL 5213099, at *2-3 (C.D. Cal. 2021) (“[A]ffixing the term ‘smart’ to a type of product ․ reflects the presence of a microprocessor” and “indicates the product's computing capabilities.”). There is no reason to think that the technological functions of smart glasses are any less essential to the product than its eyewear form. See, e.g., In re Apple Inc., No. 86857587, 2021 WL 487614, at *7 (T.T.A.B. Jan. 12, 2021) (when evaluating whether SMART KEYBOARD is generic for “technologically advanced keyboard,” examiner recognized that the “core features” of product were both the “keyboard and connective technology”).⁹ The PTO offers no convincing argument why the essence of 21st-century “smart” products—smartwatches, smart cars, and (yes) smart glasses—has in no way evolved past the 19th-century attributes those products continue to share. Nor is there an empirical basis to presume that eyewear is invariably the only product category to which smart glasses could belong. To the contrary, the evidence (including the PTO's) revealed that smart glasses are sold retail not only under traditional eyewear categories but also in wearable technology or consumer electronics departments. (EX 40-48, 273-40, 273-42, 273-43, 273-44).

The nature of a smart product, in short, is at least a contestable factual question, so its answer cannot be left to intuition, assumption, or abstraction. See Canfield, 808 F.2d at 299. Unless undisputed, the “key aspect” of a product—even as the PTO construes it under Cordua—is a material fact issue that must be decided in terms of what relevant buyers familiar with the product would perceive to be its integral or paramount feature. See Zobmondo, 602 F.3d at 1113 (In trademark cases, “we are required to consider standards of meaning not our own, but prevalent among prospective purchasers of the article.”) (quoting Bada Co. v. Montgomery Ward & Co., 426 F.2d 8, 11 (9th Cir. 1970)); see also In re Apple, 2021 WL 487614, at *30 (explaining that “determination of whether an alleged mark having the term ‘smart’ is generic depends on whether there is evidence in the record that the term, when combined with ‘keyboard’ as SMART KEYBOARD, identifies the genus of the goods”).

So even accepting the PTO's application of Cordua at face value, the agency has not established the necessary factual predicate that the eyewear form of smart glasses is the exclusive trait by which to judge if relevant consumers think SPECTACLES is a generic category name for the product. Indeed, there is no evidence that either the eyewear aspect or the computing function of smart glasses is the sine qua non of the product. That is why the parties’ stipulation never contemplated that one attribute should override the other for primary significance analysis. Thus, as the parties agreed, the relevant product genus here is “smart glasses”—defined in toto as an emergent product embodying both its essential smart computing elements and its prominent glasses form factor.

B. The PTO Has Not Shown that the Primary Significance of SPECTACLES is as a Product Name for Smart Glasses.

With the relevant product genus of smart glasses properly defined, the next question is whether “consumers in fact perceive” SPECTACLES “as the name” for that product class as such or “as a term capable of distinguishing among members of the class.” Booking.com, 591 U.S. at 560-61. “Evidence informing that inquiry can include not only consumer surveys, but also dictionaries, usage by consumers and competitors, and any other source of evidence bearing on how consumers perceive a term's meaning.” Id. at 561 n.6; see 2 McCarthy § 12:57 (“Evidence of public understanding can be obtained from any competent source, such as, testimony, surveys, dictionaries, trade journals, newspapers, Internet and other publications.”). No single piece of evidence is automatically more persuasive or important than the others; all the evidence must be evaluated, compared, and weighed together considering the trademark issues in dispute, the parties’ respective burdens of production and persuasion, and the applicable legal standards of proof.

1. The PTO's Circumstantial Evidence Proves Only That “Spectacles” Is Descriptive of Smart Glasses

To prove that “spectacles” is a generic name for the product category of smart glasses, the PTO relies, first, on circumstantial usage evidence collected from (a) dictionaries, (b) Snap's own marketing, (c) competitor product listings, (c) and online articles. As alluded to above, however, the PTO built its entire case around the agency's unproven and legally misguided idea that the eyewear form of smart glasses—but none of its computing capabilities—is the key product attribute against which consumer perception of SPECTACLES should exclusively be measured. Once the PTO's circumstantial evidence is detached from that unfounded and unsound presupposition, it proves only that “spectacles” is a descriptive—not generic—term for smart glasses.

Linguistic Evidence. The dictionary definitions cited by the PTO essentially define “spectacles” as “eyeglasses,” a “pair of eyeglasses,” or something “resembling eyeglasses in shape or suggesting them in function.” (A2048-51, A2230-32, A4245; EX 122, 123, 124, 335). Those definitions establish undeniably that “spectacles” is a common synonym for eyewear. Yet “establishing that [SPECTACLES] is generic for something—in this case, the term's own dictionary meaning—does not establish that [it] is generic for the relevant class of goods.” Kudos Inc. v. Kudoboard LLC, 2021 WL 5415258, at *8 (N.D. Cal. Nov. 20, 2021). To say that “spectacles” is generic for eyewear just begs whether the term is generic for smart glasses qua smart glasses. The definition of “smart glasses” offered by the PTO doesn't answer that question, either. (EX 126).

It only reinforces, as explained earlier, that smart glasses have two defining attributes: its outer eyewear design and its inner “computer technology.” Contrary to the PTO's contention, the usage example—“A pair of smart glasses will mean that you can take photos with your spectacles”—doesn't prove that “spectacles” denotes smart glasses as a product; it just illustrates how “spectacles” can connote the smart product's undeniable physical resemblance to conventional glasses. In other words, “spectacles” describes, it doesn't categorically name, smart glasses.

That is why the testimony of the PTO's linguistics expert, Dr. Edward Finegan, is beside the point. Finegan explained that words like smartphone and smartwatch are created by starting with a known category word (what he called “genus”) and extending it with the differentiating properties (what he called “species”) that set the new word apart from others in the same lexical “genus” category. (Finegan, Vol. III at 91-95). Thus, for example, the definition for smartwatch is: “a wristwatch [genus] that includes computerized functionality [species].” (Id. at 95). For smartphone, it is: “a cell phone [genus] that includes additional software functions [species].” (Id.). So too for smart glasses: “a pair of glasses [genus] that contain computer technology [species].” (EX 270-62, 337-3; Finegan, Vol. III at 95-96). But the relevant trademark question is what to make of the differentia in each of these smart product definitions: the “computerized functionality,” the “additional software functions,” the “computer technology.” Finegan doesn't answer that question.¹⁰ And he cannot mean that public understanding of smart products should be inferred only from the meaning of their lexical genus words while ignoring their differentiating species characteristics. If so, that “would endow editors of [dictionaries] with the power to destroy trademarks.” In re Minnetonka, Inc., 212 U.S.P.Q. 772, 1981 WL 40467, at *5 (T.T.A.B. 1981). Besides, unless most technology consumers are also lexicographers, it is hard to imagine any thinking about the Aristotelian schema behind modern definitions of smart products while shopping for smart glasses.

The only pertinent inference to derive from linguistic evidence, then, is that the public understands “spectacles” to signify eyewear generally. That conclusion was cemented by Snap's corpus linguistics expert, Dr. Robert Leonard. (Leonard, Vol. I at 92-94). Based on his analyses of the naturally occurring uses of the term in large data sources (“corpora”), Leonard confirmed that American English readers and speakers regularly use or understand “spectacles” to mean eyewear, not the product name for smart glasses.¹¹ (EX 1-10, 1-11, 1-15, 1-16). That conclusion is persuasive not only because it is consistent with the PTO's own linguistic evidence, but also because it is so unsurprisingly unremarkable. Leonard's testimony on this point may thus be credited even after discounting other opinions of his in view of Dr. Finegan's valid criticisms. (Finegan, Vol. III at 69-71, 77-78).

Snap's Usage Evidence. The PTO highlights Snap's marketing or related public materials, as sampled below, to suggest that the mark-holder itself uses SPECTACLES as a generic product name.

Like the dictionary usage examples for “smart glasses,” however, these excerpts—in context—reveal that Snap's branded use of SPECTACLES is combined with words and images that emphasize the technology features of Snap's product as much as its eyeglasses form. (EX 68, 106-1, 119-70; Chan, Vol. I at 32-33, 59, 61). That is why Snap advertises its branded product, for instance, not only as “the computer you'll wear on your face” but also as “AR-friendly glasses.” (EX 75).

Stripped of that context, these snippets would be meaningless to Snap and misleading to consumers because the social media company is not in the business of selling ordinary eyeglasses. Clipping any mention of augmented reality, for example, would make one of the cited sentences read (unhelpfully to Snap and inaccurately to consumers): “The new Spectacles are our first pair of glasses that bring augmented reality to life.” Thus, it doesn't follow from Snap's own branded uses of SPECTACLES that the term designates only the eyewear component of smart glasses to the exclusion of the product's technology. What it does demonstrate, of course, is that even Snap cannot avoid the descriptive uses of “spectacles” and “eyeglasses” as interchangeable synonyms for each other.

Competitor Usage Evidence. The PTO highlights two product listings to encapsulate alleged competitor use of the word “spectacles” when selling smart glasses. (A2248; EX 104-14). In the first listing, depicted below, the product is described as “eFashion Spectacles Eyewear Glasses DVR Camcorder Camera 720p.” In the second, also depicted below, the product is described as “Relax love Bluetooth Smart Glasses Spectacles Open-ear Glasses Speaker with MIC Waterproof.” Each product is sold for about $30.00. By contrast, the available version of SPECTACLES 3 sells for almost $400. (Chan, Vol. I at 49).

Such haphazard product descriptions do not establish consumer understanding of “spectacles” as a generic name for smart glasses. None even shows that Snap's competitors (if they are indeed competitors) are using the word “spectacles” as a product name—and thus that their shared customer base would understand that term the same way. To the contrary, these listings illustrate how modern online “sellers string together words in their product names to ensure the most hits” by potential consumers “in the age of Amazon, eBay, and Google, in which consumers shop via search bars.” Uncommon, LLC v. Spigen, Inc., 926 F.3d 409, 422 (7th Cir. 2019). Thus, if anything, these product listings show only that while the term “spectacles” isn't used as a common retail term to name smart glasses as a product type, it is naturally exploited as a word to describe one of the product's main attributes. See id. (finding similar listing for cellphone case to be evidence of descriptive not generic use).

More probative for genericism is how competitors display or group products for retail sales to draw and direct consumers to their products.¹² On this issue, both sides’ marketing experts provided helpful testimony consistent with the other. For Snap, Dr. Peter Golder studied how the top 10 U.S. online retailers (based on consumer electronic sales in 2021) classified smart glasses for sale on their websites. (EX 40-47). Golder reasoned that retailers would categorize products based on terms consumers would need to find products they want most easily. (Golder, Vol. II at 45-46). He found that while about half of the retailers displayed no explicit product categories, the others used “smart glasses” or “wearable technology” as product category menus on their websites. For example, Amazon lists “smart glasses” under the “Wearable Technology” category of its Electronics Department. (EX 40-48). Thus, finding no retailers who use “spectacles” as a category name, Golder concluded that most consumers would not use or understand “spectacles” as a product category name, either. (Golder, Vol. II at 46-47).

To be sure, the PTO's rebuttal expert, Dr. Ashish Sood, discovered gaps in Golder's retailer searches. As Sood's work revealed, smart glasses (mostly Ray-Ban STORIES) are also classified under retail categories like “Sunglasses & Eyewear,” “Eyeglasses,” or “Prescription Eyewear.” (EX 273-39, 273-40, 273-41, 273-45). But even then, in nearly every online product menu that Sood found, sellers used “Smart Glasses” as the most common product category name—none used “Spectacles.” (EX 273-40, 273-42, 273-43, 273-44). Of course, a product name need not be the most common word used by consumers or competitors to still be generic. See Premier Nutrition, Inc. v. Organic Food Bar, Inc., 2008 WL 1913163, at *11 (C.D. Cal. Mar. 27, 2008). But that doesn't make it irrelevant either that competitors rarely (or ever) use a term to name a product in the market. See E.T. Browne Drug Co. v. Cococare Prods., Inc., 538 F.3d 185, 198 (3d Cir. 2008) (Evidence of sellers displaying words other than contested term as product category names for prospective customers “tends to prove that the term is not generic.”). Thus, because the marketing evidence reveals that smart glasses are sold in many retail categories—none of which, though, is named “Spectacles”—competitor usage evidence overall fails to support the PTO's case for the claimed genericness of SPECTACLES as applied to smart glasses.

Media Usage Evidence. The PTO cites several online articles to claim that media outlets use “spectacles” as a product name for smart glasses. Yet these sources are again predicated on the PTO's hypothesis that smart glasses should be considered functionally no different than ordinary eyewear for primary significance analysis. Without that linchpin premise, the third-party usage evidence only underscores that “spectacles” is an apt description for the eyewear form of smart glasses. Indeed, as shown in the first set of tables below, most use the term only combined with words like “AR,” “smart,” or “high-tech.” (A2383, A2397, A2400, A2409, A2414, A2873; EX 88-1, 90-7, 92-1, 93-1, 97-3, 108-1).

A couple say that smart glasses “look like regular spectacles.” (A2866, A2871).

And the remaining few combine “spectacles” with a description of computer hardware, smartphone connectivity, or similar technological features. (A2868; EX 86-2, 87-1, 94-2, 97-3).

In short, like its evidence of Snap's and competitor usage, this media evidence reveals only the descriptive—not product naming—uses of “spectacles.” In nearly every article, the word “spectacles” is not used alone to capture the significance of the product class; it is invariably used with an appended technology reference like “AR,” “high-tech,” and of course “smart.” Even the articles that seemingly use the word by itself cannot intelligibly tie into smart glasses unless it is combined with an explanation of built-in or associated computer features. And for the remaining articles, if using “spectacles” to convey how the products “look like” regular glasses isn't a descriptive use but a product designation, it is hard to imagine what space is left between generic names and descriptive marks.

Were there any need to consider Snap's countervailing media evidence, it would only highlight—at a larger scale—what the PTO's cherry-picked media evidence likewise did not reveal: any meaningful use of “spectacles” as a product name. After searching a large database of articles from more than 30,000 sources dating back 40 years called Factiva, Dr. Golder found no significant media usage of “spectacles” as common name for eye-wearable technology. (EX 40-8, 40-13, 40-14, 40-15, 40-46; Golder, Vol. II at 20-21, 27). Similarly, after analyzing hundreds of tweets from a database called Pulsar TRAC using the same methodology, Golder found no significant social media usage of “spectacles” as a common name for smart glasses. (EX 40-8, 40-30, 40-31, 40-32, 40-33, 40-34, 40-46; Golder, Vol. II at 29-33). Even after the weight of Golder's findings are adjusted to account for Dr. Sood's valid criticisms of his historical research methods (Sood, Vol. II at 240-41, 247-48; Sood, Vol. III at 10), the upshot of Golder's findings confirms the absence of evidence—which is the PTO's burden to produce—proving that SPECTACLES is perceived by consumers as a product name rather than a descriptive term for smart glasses.

2. The PTO's Survey Evidence Cannot Reliably Show that Most Consumers Perceive SPECTACLES as a Product Name

The PTO also relies on direct consumer survey evidence to prove that SPECTACLES is generic for smart glasses. In genericism disputes, Teflon and Thermos surveys (each named after the marks at issue in the cases that first used the respective survey formats) are the most recognized and accepted surveys. See 2 McCarthy § 12:14. In a Teflon survey, participants are given a “minicourse” in the difference between a brand and generic name, typically with presumptively obvious examples. Participants are then given a “mini test” where they must classify at least one pair of terms as brand or generic to test their understanding. Those who correctly classify those terms are then presented with the “main test” where they must classify even more terms (including the mark at issue and other “control” or “benchmark” terms) as brand or generic. (Anderson, Vol. II at 137-38). By contrast, a Thermos survey uses open-ended questions asking respondents to list words they would use in imagined purchasing situations to discuss a class of products or to obtain products in a store. The questions might also ask if participants know brand names of relevant products and, if so, to list those. (EX 147-5, 147-26; Sowers, Vol. I at 169).

The PTO retained Dr. Justin Anderson to conduct a Teflon survey (Anderson, Vol. II at 150-51; EX 16), while Snap retained Brian Sowers to conduct a Thermos survey. (Sowers, Vol. I at 163-65). Anderson conducted a second Teflon survey after his first was criticized by Sowers. Both survey types—if properly designed and interpreted—can illuminate how the relevant buying public uses or understands a contested mark as applied to the relevant goods or services. See Booking.com, 591 U.S. at 561 n.6; Princeton Vanguard, LLC v. Frito-Lay N. Am., Inc., 786 F.3d 960, 969 (Fed. Cir. 2015). At the same time, “both types of surveys have their positives and negatives.” Heymann, supra, at 976. For instance, a Thermos survey might “be open to misinterpretation” because it is based in consumers’ subjective experiences rather than focused on their understanding of words. Id. at 976-77. Similarly, though, a Teflon survey may fail to “convey to consumers that words serve as trademarks not in the abstract but in context,” meaning “a Teflon survey might not capture consumer understanding in the marketplace.” Id. at 977.

Contrary to the PTO's insistence, then, the Teflon survey method enjoys no presumption of validity.¹³ (ECF 144 at 14, 31-32; ECF 152 at 8). No survey is the “be-all and end-all” proof of consumer perception. Booking.com, 591 U.S. at 564 (Sotomayor, J., concurring); see Simon Prop. Grp. L.P. v. mySimon, Inc., 104 F. Supp. 2d 1033, 1038-39 (S.D. Ind. 2000) (“No survey model is suitable for every case” or “beyond criticism, especially in the context of litigation.”). “Flaws in a specific survey design, or weaknesses inherent in consumer surveys generally, may limit the probative value of surveys”—especially when the central dispute is about “determining whether a particular mark is descriptive or generic.” Booking.com, 591 U.S. at 564 (Sotomayor, J., concurring). So too here.

According to Dr. Anderson, the topline results of his two Teflon surveys showed that no less than 73 to 82 percent of survey participants classified SPECTACLES as “a generic name that refers to a type of product that may be made by more than one company.” (EX 271-266, 272-46). The aggregated results of his surveys are shown in the two tables below.

(First Teflon Survey, EX 271-16).

(Second Teflon Survey, EX 272-46).

Anderson interprets these combined survey results to mean that relevant consumers think that the “primary significance” of SPECTACLES is not as a brand name but as a generic name for smart glasses. (EX 271-11; Anderson Vol. II at 223-24). But while his opinion that consumers do not primarily recognize SPECTACLES as a brand name could be reliable, it does not reliably follow from his survey results that those consumers therefore perceive SPECTACLES primarily as a generic product name instead. To the contrary, there are many reasons to think that the raw topline percentages from his Teflon surveys disguise the equally possible (if not more likely) conclusion that a statistically significant number of survey respondents—if they weren't guessing or responding to the surveys’ demand characteristics—understood SPECTACLES as a descriptive term for smart glasses rather than a generic product name.

First, the Teflon surveys appear to have been designed by taking as given the PTO's unproven and legally flawed supposition (drawn from Cordua’s key-aspect rule discussed earlier) that any terms consumers may associate with eyewear including smart glasses are only unprotectible generic product names rather than potentially registrable descriptive marks. See Amazon Techs., Inc. v. Daigle, No. 92054425, 2013 WL 3191225, at *11 (T.T.A.B. May 28, 2013) (Teflon survey instructions and questions “inspired by the concept of ‘key aspect’ or ‘central focus’ ” language from Federal Circuit cases “may open too wide a door to potential misunderstanding by the subjects”). Anderson thus evidently saw no need to redesign the default Teflon survey format to test (if possible) whether and to what extent respondents could distinguish between inherently nondistinctive terms—those that, on one hand, may name a product from those, on the other, that may describe the same product. See Eagle Snacks, Inc. v. Nabisco Brands, Inc., 625 F. Supp. 571, 582 (D.N.J. 1985) (when “confusing” survey is not designed to distinguish between “common or descriptive name[s],” survey results cannot exclude possibility that consumers understood contested mark as merely descriptive rather than generic). Thus, the percentages of consumers who classified SPECTACLES as a generic name strictly according to Anderson's survey design cannot differentiate those who, while they may not have recognized the mark as a brand designation, still may have thought the mark was only descriptive of smart glasses rather than a generic name for the product.

Second, consistent with most default Teflon survey formats, both of Anderson's surveys overweighted fanciful and arbitrary (or at best suggestive) marks for the brand control names and common vernacular words for the generic control names. See id. (discounting Teflon survey because test terms were “on extreme ends of the range of protectable trademarks” signaling that “to be a ‘brand’ name, a mark had to be of the strength of” arbitrary or fanciful marks). As a result, participants may have “gain[ed] the false impression that all trademarks must be fanciful names.” Quentin J. Ullrich, Corpora in the Courts: Using Textual Data to Gauge Genericness and Trademark Validity, 108 Trademark Rep. 989, 1001 (2018). That misimpression can then create response bias by unduly “influenc[ing] how [a] respondent categorizes the mark in question if it is not also a fanciful mark.” Id.

Consider the stark contrast in the corresponding pairs of control brand and generic names for the identical minicourses in both Teflon surveys:

(EX 271-43, 271-244, 271-260). Now consider the disparities between the corresponding pairs of control names, respectively, in each of the mini tests:

(EX 271-247, 271-263, 271-264, 271-265, 272-35, 272-41, 272-104, 272-105, 272-106). Finally, consider the choices of control terms in the critical main tests of each survey:

(First Survey, EX 271-29, 271-30).

(Second Survey, EX 272-35).

By the time respondents have gone through these minicourses and mini tests, it takes little survey expertise to reasonably predict how the respondents would classify the target term “spectacles” in the main tests given the chasm between inherently distinctive brand names (in the left column) and inherently nondistinctive generic words (in the right column). When “examples of brand names given in the instructions, the gate-keeper questions, and the survey [are] unnecessarily weighted toward fanciful marks, as opposed to marks composed of ordinary words,” respondents will likely be primed to “believe that any ordinary word or any phrase composed of ordinary words is a common name, rather than a brand name.” Amazon Techs., 2013 WL 3191225, at *11 (finding Teflon survey “to be of limited probative value” for these reasons).

To be sure, the control brand names in the second survey (changed in response to Sowers’ criticisms) are marginally better selections. Even as modified, though, VISION PRO (by Apple) and ECHO FRAMES (by Amazon) are popular brand names if not for eye-wearable technology itself, then as names inextricably linked to two of the most well-known companies, Apple and Amazon. Even BLADE (by Vuzix) is arbitrary, or at best suggestive, for smart glasses. The conceptual distance between those brand control terms and the generic control terms—CAMERA GLASSES, WIFI GLASSES, and LENSES—is too far to mitigate response bias. At the same time, the conceptual closeness between SPECTACLES and the inherently nondistinctive generic names is impossible to miss: two have “GLASSES” in the names while the third, LENSES, is the name of an object probably most associated with eyewear.

Third, the ambiguities and inconsistencies in the mini test control names for both surveys likely biased the representativeness of each of the relevant consumer samples.¹⁴ Those flaws could have either selected only for those more likely to pick SPECTACLES as a common word or excluded those who may have understood SPECTACLES as neither generic nor brand names but a mere descriptive term. “One of the more significant issues with the Teflon survey is the potential effect of demand characteristics. Respondents may answer based on what they think a word is supposed to mean, or what they perceive to be the interviewer's intended meaning of a word, rather than based on how they primarily use or understand the word.” Ullrich, supra, at 999. Thus, when survey participants are not educated about borderline nondistinctive terms that could be either generic or descriptive but then compelled to make a choice between only brand and generic names, respondents can be prone to pick “based on their hunches about what is likely to look like a generic term” rather than on how they would “understand[ ] the term in the context of a purchasing experience.” Heymann, supra, at 980.

Recall that in the minicourse for each Teflon survey, respondents were taught that SMARTPHONE is a product name for Apple's branded IPHONE, that CAMERA is a product name for GoPro's HERO-branded cameras, and that SMART GLASSES is a product name for Lenovo's THINKREALITY-branded smart glasses. But in the first mini test, respondents had to sort—when “thinking of names relating to smart glasses”—not only ANZU as a brand name for the Razer ANZU smart glasses but also DISPLAY as a so-called “generic name” for smart glasses. Many respondents could have wondered what to make of DISPLAY since it cannot name smart glasses in the way that the control generic terms (one of which was SMART GLASSES) named their corresponding branded products in the minicourse. On the other hand, DISPLAY at best described an aspect of smart glasses. But those confused by that ambiguity weren't taught how to classify descriptive names. Nor were they given any meaningful third choice between generic and brand (besides “I don't know”). Thus, this flaw either excluded (to Snap's detriment) any respondents who said “I don't know” or picked brand over generic, or it selected (in the PTO's favor) only those who erred on the side of choosing generic for DISPLAY because that was closest to what they may have (accurately) perceived to be an apt description of smart glasses.

The second survey, while ostensibly redesigned to address this problem, tends to confirm that Anderson's pairing of mini test control words still perplexed respondents after the minicourse. This time, respondents had to sort between Ray-Ban's STORIES product as a brand name and EYEWEAR as a generic name—again, when “thinking of names relating to smart glasses.” Nearly 50 percent of respondents got that classification wrong. (EX 272-43, 272-45). So even more than the first mini test, this second one likely excluded many who couldn't confidently say that EYEWEAR was a generic name for smart glasses given that it also naturally described the product. Alternatively, it selected more for those who opted to say that EYEWEAR was a generic name because it still “relate[d] to smart glasses” by describing them—which the question implied (incorrectly) would be enough to make a word generic. And, of course, a third possibility is that respondents guessed, meaning that the survey sample was not reliably representative of those who understood the difference between generic and brand names—even on the survey's own terms. See Shari S. Diamond, Reference Guide on Survey Research, in Reference Manual on Scientific Evidence, 359, 388 (Fed. Jud. Ctr. 3d ed. 2011) (“When unclear questions are included in a survey, they may threaten the validity of the survey by systematically distorting responses if respondents are misled in a particular direction, or by inflating random error if respondents guess because they do not understand the question.”).

Fourth and not least, the Teflon surveys injected intolerable uncertainty into the results because respondents were taught and asked to classify terms by how much they related or referred to smart glasses. (EX 272). Making the “distinction between descriptive terms and generic names,” is even a “challenge for the courts” because the “generic-descriptive line is too often smudged.” 2 McCarthy § 12:20. That difficult task is made harder when it is sometimes suggested—as in Cordua—that a word can be generic if it just “refers” to a product type. So asking anonymous participants untrained in trademark law (after just two computer screens’ worth of a trademark “minicourse”) if a disputed term relates or refers to a product is “a misleading and incorrect way of asking the crucial question because it smears the critical line between a generic name and a descriptive word.” 2 McCarthy § 12:10. Such “instructions may induce subjects to consider words that describe or suggest any potential use of a [term] to be a ‘common name’.” Amazon Techs., 2013 WL 3191225, at *11.

Consider Anderson's selection of all control terms from both surveys side by side to get a birds-eye view of the surveys’ design structure ¹⁵ :

If respondents were asked to “think[ ] of names relating to smart glasses” and told that it is enough if the word “refers to” a product, is there much chance that most wouldn't classify SPECTACLES alongside the generic control names in the right-hand column? Each is a common word or phrase that identifies if not ordinary eyewear terms (“glasses,” “eyewear,” and “lenses”), then obvious features or functions of smart glasses (“camera,” “display,” “microphone,” “photograph,” and “screen”). On the other hand, there are only two control brand names even remotely related to eyewear (using the terms “vision” and “frames”), but even then, those were made to appear distinctive as recognizable two-word brand names associated with well-known companies (Apple and Amazon).

Those are only four of the most significant design flaws in the Teflon surveys. But they are enough to demonstrate why the surveys’ absolute numbers of respondents who classified SPECTACLES as a generic name are not reliable enough to find empirically that relevant consumers perceive the mark as a product name for smart glasses. Those seemingly high figures obscure the respondents who sorted SPECTACLES into the “generic” column because they thought it named smart glasses from those who did so because in their minds it could just as well have described the product.¹⁶ By following the default Teflon format (aimed at the distinction between inherently distinctive and nondistinctive terms), Anderson failed to reliably reveal consumer understanding of the nuanced difference between generic names and descriptive marks—both inherently nondistinctive sets of words with no source-identifying trademark significance on their face. Thus, this consumer survey evidence cannot carry the PTO's burden to prove that SPECTACLES is a generic product name for smart glasses.¹⁷

IV.

If SPECTACLES is not generic, Snap contends that it is inherently distinctive as a suggestive mark and thus automatically registrable with no proof of secondary meaning needed. Snap also argues that, even if only descriptive, SPECTACLES is still immediately registrable because it has acquired secondary meaning in the smart glasses market. For its part, the PTO not only contests that SPECTACLES is suggestive but also contends that the mark is highly descriptive of smart glasses. As a result, the agency argues that Snap must meet a heightened burden of production to prove secondary meaning but has failed to do so.

Like the genericness determination, whether SPECTACLES is suggestive or descriptive is a question of fact. See Zobmondo, 602 F.3d at 1113. It is now Snap, though, that bears the burden of proving either the inherent or acquired distinctiveness of SPECTACLES by a preponderance of the evidence. See In re La. Fish Fry Prods., Ltd., 797 F.3d 1332, 1335 (Fed. Cir. 2015); In re Gj & Am, LLC, No. 86858003, 2021 WL 2374670, at *17 (T.T.A.B. June 8, 2021). Evidence relevant to both issues, as with genericness, can come from any competent source including dictionaries, consumer surveys, media publications, and usages of the disputed term by both the applicant and its competitors. See In re Northland Aluminum Prods., Inc., 777 F.2d 1556, 1559 (Fed. Cir. 1985). Whatever the source of evidence, the overriding question remains how relevant consumers perceive the disputed mark when applied to the pertinent goods, not whether anyone—including the court—can conclude in the abstract that the mark is suggestive or descriptive. See Entrepreneur Media, 279 F.3d at 1142; In re Nett Designs, 236 F.3d 1339, 1341 (Fed. Cir. 2001).

A. SPECTACLES Is Highly Descriptive Rather than Inherently Suggestive of Smart Glasses.

Suggestive marks share much in common with descriptive marks: both ultimately connote some distinctive attribute of the goods or services for which they are used. See Kendall-Jackson, 150 F.3d at 1047 n.8. The difference lies in “how quickly and easily consumers grasp the nature of the product from the information conveyed.” Cross Com. Media, Inc. v. Collective, Inc., 841 F.3d 155, 162 (2d Cir. 2016); see 1 McCarthy § 11:64 (Suggestive marks, like descriptive marks, “shed some light upon the characteristics of the goods, but so applied they involve an element of incongruity” absent from descriptive marks). A term is suggestive if “imagination or a mental leap is required in order to reach a conclusion as to the nature of the product being referenced.” Zobmondo, 602 F.3d at 1115 (quoting Rudolph Int'l, Inc. v. Realys, Inc., 482 F.3d 1195, 1198 (9th Cir. 2007)). “Descriptive marks,” on the other hand, “define a particular characteristic of the product in a way that does not require any exercise of the imagination.” Quicksilver, Inc. v. Kymsta Corp., 466 F.3d 749, 760 (9th Cir. 2006) (quoting Yellow Cab Co. of Sacramento v. Yellow Cab of Elk Grove, Inc., 419 F.3d 925, 927 (9th Cir. 2005)).

While descriptive and suggestive marks can sometimes be hard to tell apart, see 1 McCarthy § 11:66, the mark SPECTACLES applied to smart glasses is not one of those close calls. Indeed, while capable of serving as a source identifier, SPECTACLES is much closer to the generic-descriptive line than it is to the descriptive-suggestive line—making the mark not just merely descriptive, but highly descriptive. See Filipino Yellow Pages, 198 F.3d at 1151. As already detailed, the evidence overwhelmingly proves that “an entirely unimaginative, literal-minded person would understand the significance of” SPECTACLES as a spot-on description of the prominent eyewear form of smart glasses. Entrepreneur Media, 279 F.3d at 1142. The linguistic and third-party usage evidence also shows that no mental leap is needed by consumers to understand that “spectacles”—as a common synonym for eyeglasses—signals a direct connection to the eyewear design of smart glasses. Competitor product listings likewise use “spectacles” to directly communicate information to consumers about the most visibly prominent feature of smart glasses—its eyewear form factor. See Japan Telecom, Inc. v. Japan Telecom Am. Inc., 287 F.3d 866, 872 (9th Cir. 2002). And not least, Snap's own advertising conveys nothing subtle or incongruous about the eyewear form of smart glasses when its product is identified or described by the company as “Spectacles.”

Indeed, as exemplified in these advertisements commissioned by the company itself, SPECTACLES is unmistakably highly descriptive of the essential eyewear form of smart glasses:

(EX 75). Taken at face value, these ads show that Snap does not want consumers pausing or wondering how to don its wearable technology—they want consumers to know instantly (in a snap, one might say) that SPECTACLES is worn on the face, over the eyes, like ordinary eyewear. (Chan, Vol I. at 59, 61; EX 106-1).

That is no accident. Ryan Chan, the only corporate representative who testified for Snap, made clear that the company wants buyers of its wearable technology to instantly recognize and appreciate—not deduce through imagination or reasoning—the product's visible eyewear form factor. (Chan, Vol. I at 32-33, 59, 61). When Snap was developing its wearable technology, Chan explained that the “most comfortable form factor” wanted by Snap's customers “was something that looked like glasses.” (Id. at 32.). He thus expected that the “glasses form factor” is how wearers would “perceive” the “Spectacles” product. (Id. at 32-33). And that is why when Snap advertised its branded “Spectacles” product, the company identified it as “a pair of glasses” to convey “what the product looks and feels like.” (Id. at 61). So if all the evidence undeniably proves that consumers understand “spectacles” to denote eyewear, Snap cannot seriously maintain that SPECTACLES as a mark is anything if not highly descriptive of its branded smart glasses product.

Importantly, Snap has presented no countervailing evidence that the “commercial impression” of the SPECTACLES mark when “viewed through the eyes of a consumer” is any different from that reflected in its advertisements or corroborated by Chan's testimony. DuoProSS Meditech Corp. v. Inviro Med. Devices, Ltd., 695 F.3d 1247, 1253 (Fed. Cir. 2012). It points to no evidence that buyers of smart glasses need any “imagination, thought, and perception to understand the mark's significance.” Id. Instead, Snap invites the court to discern some suggestive meaning for SPECTACLES just by the term's arcane etymology or alternative dictionary meanings. To be sure, a dictionary is a “suitable starting place for attempting to draw the line between a suggestive and descriptive mark.” Fortune Dynamic, Inc. v. Victoria's Secret Stores Brand Mgmt., 618 F.3d 1025, 1033 (9th Cir. 2010) (cleaned up). But there is no trademark significance to be found at the end of Snap's excursion through dictionary evidence.

First, contrary to Snap's claim that SPECTACLES is suggestive because it is an old-fashioned rather than modern dictionary term for eyewear, mere unfamiliarity with a word that “requires a hearer to think about its meaning does not show that it is suggestive.” Forum Corp. of N.A. v. Forum, Ltd., 903 F.2d 434, 443 (7th Cir. 1990). A “term may be descriptive even if it is an unusual and uncommon term or a play on words, if it still describes the goods or a quality of the goods.” 1 Gilson § 2.03[2][c]. And while the etymology of “spectacles” may render the word an “imaginative” or “creative” business choice from the marketer's perspective, that does not make it a legally suggestive term from the consumer's perspective. Eagle Snacks, 625 F. Supp. 571 at 582. At most, these linguistic idiosyncrasies of the term “spectacles” only help move it out of the generic category and into the descriptive one.

Second, Snap contends that SPECTACLES is a double entendre because the word “spectacles” is also defined as “an event or behavior”—that is, “something that can be seen or viewed, especially something of a remarkable or impressive nature,” or “something exhibited to view as unusual, notable, or entertaining” including one's own “public display.” (EX 1-6, 1-7, 1-8). But for “trademark purposes, a double entendre is an expression that has a double connotation or significance as applied to the goods or services” at issue. TMEP § 1213.05(c). It is not enough, in other words, for a proposed mark to just have two meanings in the dictionary. Instead, the suggestiveness from a mark's alleged double meanings “must be associations that the public would make fairly readily, and must be readily apparent from the mark itself.” Id. Put another way, the dual meanings must make “the merely descriptive significance of the term ․ lost in the mark as a whole.” In re Risesmart, Inc., 104 U.S.P.Q.2d 1931, 2012 WL 6137597, at *3 (T.T.A.B. 2012) (quoting In re Kraft, Inc., 218 U.S.P.Q., 1983 WL 51972, at *2 (T.T.A.B. 1983)). But SPECTACLES has the opposite effect: whatever conceivably suggestive connotations it may have when applied to smart glasses, those meanings are eclipsed by its highly descriptive primary meaning that consumers have no mental choice but to associate with the eyewear design intrinsic to all smart glasses.¹⁸

Moreover, just because a mark may have more than one dictionary meaning does not make it a legally suggestive double entendre. There must be some evidence confirming that consumers would perceive the indirect suggestive meaning in the term over its more direct descriptive meaning. See, e.g., In re Yarnell Ice Cream, LLC, No. 86824279, 2019 WL 3183842, at *9 (T.T.A.B. July 9, 2019) (finding “scoop” merely descriptive of ice cream because applicant had no evidence the public would understand “scoop” by its alternative meaning—a “news scoop”). Yet Snap has presented no evidence that consumers would readily perceive source-identifying significance in SPECTACLES—the very thing needed to make a suggestive term inherently distinctive—just because of its multiple dictionary meanings. See Qualitex Co. v. Jacobson Prods. Co., 514 U.S. 159, 162-63 (1995) (suggestive marks are inherently distinctive because they “almost automatically tell a consumer that they refer to a brand,” while descriptive terms require a secondary meaning to serve the same purpose); 1 McCarthy § 11:71 (whether consumers are “likely to regard the mark really as a symbol of origin” is important measure of suggestiveness since “descriptive marks cannot pinpoint one source”).

Third and last, Snap curiously argues that because “spectacles” standing alone could mean just eyewear, SPECTACLES is somehow suggestive since consumers might not know—without context—if the term is referring to eyewear or to smart glasses. (ECF 152 at 152-54). It is hard to see, though, how the evidence establishing why SPECTACLES is not generic for smart glasses can somehow unexplainably become evidence proving that the mark is suggestive. It is one thing for SPECTACLES to have been spared the fate of a generic product name for smart glasses because “spectacles” denotes eyewear generally rather than smart glasses specifically. But that doesn't mean the mark is suddenly shed of its highly descriptive “commercial impression” and imbued with inherent suggestiveness in “the eyes of a consumer.” DuoProSS Meditech, 695 F.3d at 1253. The PTO's evidence only failed to prove that SPECTACLES is a generic name for smart glasses precisely because of how well it managed to show that the mark is highly descriptive of the form and function of smart glasses instead. See id. at 1251-52. So just because the evidence may have nudged SPECTACLES over the “fine line between a highly descriptive designation and a generic name” cannot mean that the term catapulted itself onto the inherently suggestive side of the distinctiveness spectrum. 2 McCarthy § 12:20.

In the end, the significance of a mark to consumers must be evaluated as if it were “seen on the goods or services” at issue. Lahoti v. VeriCheck, Inc., 586 F.3d 1190, 1201 (9th Cir. 2009). Suggestiveness is not discerned by “asking whether one can guess, from the mark itself, considered in a vacuum, what the goods or services are.” 1 McCarthy § 11:16. Contrary to the premise of Snap's argument, then, the suggestiveness of a mark does not turn on how “relatively difficult” it is for “consumers to divine the nature of the particular product at issue from the mark alone.” Cross Com. Media, 841 F.3d at 163. The test for suggestiveness “does not ask what information about the product could be derived from a mark, but rather whether a mental leap is required to understand the mark's relationship to the product.” Zobmondo, 602 F.3d at 1116 (cleaned up).¹⁹ For the many reasons now detailed, Snap has not proven that consumers need to employ any imagination to instantaneously associate SPECTACLES with one of the most essential aspects of smart glasses—its eyewear design.²⁰

B. SPECTACLES Has Not Yet Acquired Secondary Meaning as a Highly Descriptive Mark

Snap contends that even if SPECTACLES is descriptive of smart glasses, it has attained secondary meaning. But because SPECTACLES is highly descriptive of smart glasses, Snap's burden of production to establish secondary meaning is higher: the more descriptive the term, the better the produced evidence must be—qualitatively or quantitatively—to meet an applicant's overall burden of persuasion even under a preponderance of the evidence standard. See In re Steelbuilding.com, 415 F.3d 1293, 1300 (Fed. Cir. 2005); Filipino Yellow Pages, 198 F.3d at 1151; La. Fish Fry, 797 F.3d at 1336-37; In re Gj & Am, 2021 WL 2374670, at *17. As an issue of fact, secondary meaning is “the mental association by a substantial segment of consumers and potential consumers between the alleged mark and a single source of the product.” Levi Strauss & Co. v. Blue Bell, Inc., 778 F.2d 1352, 1354 (9th Cir. 1985) (en banc). Secondary meaning “is directed towards the consumer's attitude about the mark in question: does it denote to him a single thing coming from a single source?” Carter-Wallace, Inc. v. Procter & Gamble Co., 434 F.2d 794, 802 (9th Cir. 1970) (cleaned up). In other words, “an applicant must show that in the minds of the public, the primary significance of a product feature or term is to identify the source of the product rather than the product itself.” La. Fish Fry, 797 F.3d at 1336 (quoting Coach Servs., Inc. v. Triumph Learning LLC, 668 F.3d 1356, 1379 (Fed. Cir. 2012)).

That showing can be established with direct or circumstantial evidence. Direct evidence such as consumer surveys or consumer testimony naturally provides the strongest evidence of secondary meaning. See Levi Strauss, 778 F.2d at 1358; Vision Sports, Inc. v. Melville Corp., 888 F.2d 609, 615 (9th Cir. 1989). But circumstantial evidence can be equally probative too. See Filipino Yellow Pages, 198 F.3d at 1151. Such evidence is relevant if it addresses “(1) whether actual purchasers of the product bearing the claimed trademark associate the trademark with the producer, (2) the degree and manner of advertising under the claimed trademark, (3) the length and manner of use of the claimed trademark, and (4) whether use of the claimed trademark has been exclusive.” Levi Strauss, 778 F.2d at 1358 (cleaned up). Evidence of widespread copying of the contested mark can also be a useful proxy for secondary meaning. See Vision Sports, 888 F.2d at 615. In all, the court's consideration of these factors reveals that Snap has produced inadequate evidence under its elevated burden of production to establish that SPECTACLES has more likely than not attained secondary meaning in the market for smart glasses.

On the one hand, Snap's seven years of continuous use of the SPECTACLES mark is prima facie evidence of secondary meaning. See 15 U.S.C. § 1052(f). And none of the PTO's competitor product listings is enough to prove that Snap's use of the mark hasn't been exclusive enough. (EX 104-1, 111-5). After all, “absolutely exclusive use on the part of the applicant” is not required. Alphaville Design, Inc. v. Knoll, Inc., 627 F. Supp. 2d 1121, 1131 (N.D. Cal. 2009). The limited evidence of Snap's advertising expenditures and gross sales revenues is superficially probative of secondary meaning, too. The PTO does not contest Snap's claim that it has spent tens of millions of dollars on advertising between 2016 and 2022, including with billboards, vending machines, and both print and digital advertising. (Chan, Vol. I at 24-25, 41-42; A2582-87, A346-48; EX 113-7). Nor does the PTO deny Snap's claim that it has sold (as of trial) an estimated 300,000 units of its products, not even including SPECTACLES 4, generating more than $37 million in gross revenues. (EX 113-6, 113-7). Finally, the PTO has no basis to challenge Snap's claim that the launch of SPECTACLES 1 generated more than 10,000 media articles, resulting in over 1.6 billion impressions (the number of times a piece of content is viewed online). (Chan, Vol. I at 25).

But this circumstantial evidence combined still fails to meet Snap's burden given the highly descriptive nature of the SPECTACLES mark. See La. Fish Fry, 797 F.3d at 1337 (continuous and exclusive use evidence held insufficient given highly descriptive nature of mark); In re Boston Beer Co. L.P., 198 F.3d 1370, 1371, 1373 (Fed. Cir. 1999) ($85 million in annual sales and $10 million in advertising costs held insufficient given highly descriptive nature of mark). For starters, while Snap's estimated advertising expenditures are high in absolute terms, it has presented no evidence to establish any baseline to understand those costs in relative terms. The mere “large expenditure of money does not in itself create legally protectable rights.” Carter-Wallace, 434 F.2d at 800 (quoting Smith v. Chanel, Inc., 402 F.2d 562, 568 (9th Cir. 1968)). Similarly, Snap's sales and revenue figures provide no convincing proxy for secondary meaning because those absolute numbers may only point to product popularity or rapid market gains. See In re Bongrain Int'l Corp., 894 F.2d 1316, 1318 (Fed. Cir. 1990); Cont'l Lab'y Prods., Inc. v. Medax Int'l, Inc., 114 F. Supp. 2d 992, 1003 (S.D. Cal. 2000). “Raw sales figures need to be put into context to have any meaning.” 2 McCarthy § 15:49.

What contextual evidence there is in the record undermines rather than buttresses Snap's claim of secondary meaning.²¹ In the company's own Thermos survey, for instance, about 42 percent of the qualified 273 participants said they were familiar with brand names for smart glasses. (EX 147-26, 147-27). Yet of those 114 respondents, only one listed Snap's SPECTACLES product as a branded smart glasses product.

(EX 147-28). Most of the other respondents listed branded products from Google, Ray-Ban, or Bose instead. While there may be no precise cutoff for when source-identification is enough to show secondary meaning, it is safe to say that less than 1 percent probably isn't enough.²² See generally 5 McCarthy § 32:190; see, e.g., Warner Bros. Ent. v. Global Asylum, Inc., 2012 WL 6951315, at *4 (C.D. Cal. 2012) (relying on 48 percent source-identification).

What's more, a 2019 brand awareness study that Snap introduced to demonstrate secondary meaning mostly undercuts it. According to Snap, the launch of SPECTACLES 3 in 2019 generated around 1,600 favorable media articles receiving about 1.3 billion impressions online. (Chan, Vol. I at 38-39; EX 75-6). To keep boosting brand awareness around SPECTACLES 3, Snap carried out its own advertising campaign in four countries including the United States (covering Los Angeles, New York, Chicago, and Miami). (EX 75-6). It then conducted a “Brand Lift Study” designed to assess the effect of the SPECTACLES 3 advertising blitz on those exposed to the advertisements. (Chan, Vol. I at 38-39). In the United States market survey, the participants were divided between 472 in the “control” group and 656 in the “exposed” group. (EX 75-16, 75-17).

As pertinent here, one of the Brand Lift survey questions was aimed at measuring “Unaided Brand Awareness.” It did so by asking survey participants an open-ended question: “When thinking of wearable cameras, which come to mind?” (EX 75-18). According to the survey results (shown in the summary table below), 0 percent of the 472 respondents in the control group and only 1 percent of the 656 respondents in the exposed group mentioned Snap's branded SPECTACLES product first in their “unaided” answers. (EX 75-23). Among those who mentioned Snap's product at all (again unaided with no prompting), only 1 percent of the control group listed the SPECTACLES product and only 1 percent of the exposed group also listed the SPECTACLES product. (Id.).

The Brand Lift Study also analyzed brand awareness among U.S. Snapchat users and non-users who were surveyed. Evidently, 58 percent of the “reached US audience” from the survey respondents were daily Snapchat users. (EX 75-41). As shown in the next summary table, among that cohort of daily Snapchat users, only 1 percent of the control group and 1 percent of the exposed group made any unaided mention of Snap's branded SPECTACLES product when asked to list wearable cameras. Among monthly Snapchat users, 0 percent of the control group and 3 percent of the exposed group made any unaided mention of Snap's branded product to the same question. (Id.). Results were similar among non-users, at 0 percent in the control group and 2 percent in the exposed group mentioning SPECTACLES as a branded product for wearable cameras. (Id.).

While not conclusive, this evidence is the only available probative context for Snap's advertising expenditures, gross sales revenues, and even favorable media coverage. While advertising input is relevant to be sure, “it is the effect of such advertising that is important, not its extent.” Co-Rect Prods., Inc. v. Marvy! Adv. Photography, Inc., 780 F.2d 1324, 1332 (8th Cir. 1985). Similarly, while successful unit sales of products are naturally relevant, that success “is not synonymous with secondary meaning” unless it can be meaningfully distinguished from mere “[p]opularity of a product.” 2 McCarthy § 15:47. Yet Snap's own evidence from its Thermos and brand surveys suggests that the company's extensive (and expensive) brand promotion efforts on top of its product sales haven't yet “effectively create[d] an association with [Snap's] product” in the minds of relevant consumers. Mercado Latino, Inc. v. Indio Prod., Inc., 2018 WL 3490752, at *4 n.7 (C.D. Cal. July 17, 2018). If that disassociation between mark and brand affects even Snap's own daily and monthly Snapchat users, there is little room to find that the company has met its burden to prove secondary meaning. See, e.g., Jackpocket, Inc. v. Lottomatrix NY LLC, 645 F. Supp. 3d 185, 253-54 (S.D.N.Y. 2022) (finding little probative value in plaintiff's marketing and sales efforts where consumer surveys showed “that even [plaintiff's] consumers d[id] not recognize [the product] as the service they have used”), aff'd, 2024 WL 1152520 (2d Cir. Mar. 18, 2024).

Finally, Snap has presented no evidence suggesting that Snap's rivals are poaching the SPECTACLES mark. Snap considers its primary competitors to be Ray-Ban, GoPro, Microsoft, and Magic Leap. (Chan, Vol. I at 54-56). But the takedown notices Snap introduced to help show alleged infringement of its mark were not directed at any of these principal competitors. And those notices merely show that many were evidently sent out but with no indication of what underlying infringing activities precisely the notices were directed at. (EXs 306 to 327). Such one-sided bulk evidence provides no sound basis to determine whether there is significant and deliberate copying of the SPECTACLES mark “intended to deceive consumers about the product's source.” TBL Licensing, LLC v. Vidal, 98 F.4th 500, 518 (4th Cir. 2024). Indeed, if the two or three meager examples of alleged infringement presented by Snap are any clue (EX 302-2, 302-3), it is hard to see enough evidence of widespread copying by even lower-grade competitors, much less genuine competitors of Snap's ilk, to suggest that SPECTACLES has achieved sufficient secondary meaning.

To the contrary, Snap's own evidence is again self-defeating. In Dr. Golder's historical Google Trends research, two hypotheses were tested: that online search popularity for a generic product name should spike whenever any type of that product is launched by a brand, while search popularity for a branded product should be more variable, spiking primarily only when there are significant events related to that specific branded product. (Golder, Vol. II at 48-49; EX 40-38, 40-39). Consistent with Golder's hypothesis about a common category name, the popularity of searches for “smart glasses” increased over time generally but with several spikes correlating with product launches of Google GLASS, Snap SPECTACLES, and Ray-Ban STORIES. (EX 40-37). At the same time, consistent with his brand name hypothesis, searches for “spectacles” spiked significantly with the release of Snap SPECTACLES products—but not when other companies’ competing products were launched. (Id.). Based on these trend lines, Golder found that it is not “spectacles” but rather “smart glasses” that is the most common generic term used by consumers to search for (what else?) smart glasses—thereby further disproving the PTO's genericism theory. (Golder, Vol. II at 51-52). But his findings also suggest that genuine competitors in the smart glasses market are not discernibly aping the term “Spectacles” when launching or promoting their own branded smart glasses to confuse or siphon away Snap's existing or potential customers—thereby undermining Snap's case for secondary meaning in its highly descriptive mark, SPECTACLES, based on any alleged widespread copying of the mark by third parties.

V.

For all the reasons given above, and in accordance with Federal Rule of Civil Procedure 52(a), the court finds and concludes that:

1. The PTO has not met its burden to prove by clear and convincing evidence that SPECTACLES is primarily understood by most consumers of smart glasses as a generic name for that product. To the contrary, the evidence establishes that those consumers overwhelmingly understand “spectacles” as a description of the product's prominent eyewear form rather than a generic name for smart glasses.

2. Snap has not carried its burden to prove by a preponderance of the evidence that SPECTACLES is suggestive and thus eligible for registration even with no proof of secondary meaning. Whatever conceivably suggestive or other meaning beyond eyewear that “spectacles” may carry is eclipsed by its highly descriptive connotation for the essential eyewear form of smart glasses.

3. Snap has not produced enough evidence to prove even by a preponderance standard that SPECTACLES carries enough secondary meaning to be included immediately on the principal register. Still, the mark can acquire such source-identifying significance if for no other reason than the term's arcane etymology and alternative dictionary meanings. SPECTACLES should thus be eligible for supplemental registration.

* * *

Judgment will therefore be entered remanding Snap's registration applications—Serial Nos. 87/177,292 and 87/211,977—to the PTO consistent with the court's findings of fact and conclusions of law, so that the agency can take the necessary actions for amendment or approval of those applications to permit supplemental registration of SPECTACLES both as a word and stylized mark.²³

IT IS SO ORDERED.

FOOTNOTES

1.   As used here, facts from the certified administrative record, admitted into evidence as Trial Exhibits 110 and 111 (ECF 42; ECF 134 at 8), are cited with their original pagination references. Relevant trial testimony, transcripts of which are filed under ECF 138-40, is cited by the last name of the witness, followed by the corresponding transcript volume and the referenced page numbers. Trial exhibits are cited using the parties’ stamped bates pagination, consisting of the prefix EX followed by the exhibit number separated from the cited page numbers with a hyphen.

2.   The photographs also displayed SPECTACLES in its stylized form. (A0012-13; A2205-06).

3.   “This does not mean that the buyer must know the identity of that ‘single source’ in the sense that she knows the corporate name of the producer or seller. In fact, few buyers know, or care about, the corporate identity of the seller of a trademarked product.” 2 McCarthy on Trademarks and Unfair Competition § 15:8 (5th ed. 2017) (hereinafter McCarthy).

4.   The PTO and TTAB applied the clear and convincing standard of proof in administrative proceedings below. (A2043, A2109, A4239, A4306). The Federal Circuit has long required that elevated standard in registration proceedings. E.g., In re Hotels.com, L.P., 573 F.3d 1300, 1302 (Fed. Cir. 2009). In May 2022, however, the PTO changed its policy and now maintains that only a preponderance of the evidence is needed. See PTO Examination Guide No. 1-22, Clarification of Examination Evidentiary Standard for Marks Refused as Generic (May 2022). But because that change was announced only after Snap filed suit here, the PTO has agreed not to dispute in this case—while preserving its right to do so in another—that it must prove genericness by clear and convincing evidence. (ECF 152 at 74-78, 146-47).

5.   For instance, “canoes,” “kayaks,” and “ships” could each be its own relevant product genus depending on the trademark dispute. Fletcher, supra, at 488. But those same goods could also be subgenera under any one of higher-level genera like “boats, vessels, and watercraft.” Id. Thus, the genus language used in trademark law should not be applied literally like the genus-species taxonomy used in scientific fields. See 1 Anne Gilson LaLonde, Gilson on Trademarks § 2.02 (Matthew Bender 2024) (hereinafter Gilson) (“genus/species” terminology can be “awkward,” “decidedly unhelpful,” and “may create confusion” in trademark context). In trademark law, “species” is the designation that identifies the source of the product while a “genus” term is the designation that names the type of product. See Filipino Yellow Pages, Inc. v. Asian J. Publ'ns, Inc., 198 F.3d 1143, 1147 (9th Cir. 1999). Thus, for example, the relevant genus for “a floating craft can be increased by using the term ‘boat,’ which refers to a broad genus, in preference over ‘canoe,’ which refers to a narrow genus.” Tom Brody, Functional Elements Can Ensure Allowance of Genus Claims, 90 J. Pat. & Trademark Off. Soc'y 621, 621 (2008) (discussing genus claims in patent law). In either case, a mark like “Old Town Trapper 15” would be the “species” designation. Id.

6.   For instance, in the classic dispute between Kellogg and Nabisco over the term “shredded wheat,” if “wheat cereals” were the “relevant product class, then Shredded Wheat would be merely a brand.” Canfield, 808 F.2d at 301 n.12. “But once it was decided that cereals containing pillow-shaped forms of wheat shreds was the relevant product class, the term ‘shredded wheat’ was obviously generic.” Id.

7.   Cordua also relied in part on an earlier set of cases that affirmed registration denials for online businesses—LAWYERS.COM, HOTELS.COM, and 1800MATTRESS.COM—even when combined with the “.com” top-level domain. 823 F.3d at 603-04 (discussing In re Reed Elsevier Props. Inc., 482 F.3d 1376 (Fed. Cir. 2007); Hotels.com, 573 F.3d 1300; and In re 1800Mattress.com IP, LLC, 586 F.3d 1359 (Fed. Cir. 2009)). However, it is an open question whether (or to what extent) those GENERIC.COM cases remain viable after the Supreme Court's 2020 decision in Booking.com, 591 U.S. 549.

8.   To the extent that the Federal Circuit's decision in Royal Crown seemingly endorsed that idea, it has been criticized. See 2 McCarthy §§ 12:1, 12:10, 12:20. But much as in Cordua, Royal Crown used the key-aspect language to address a subcategory of goods (zero-calorie drinks) nested under a broader category of claimed goods (sports beverages). 892 F.3d at 1368. Thus, Royal Crown can still be read consonant with trademark law without implying that any reference to a product attribute is generic.

9.   In this case, the TTAB defined “smart” applied to products as signifying hardware “incorporating some kind of digital electronics” or “computer systems.” In re Apple, 2021 WL 487614 at *4. It also defined “Smart Device” as “an electronic gadget that is able to connect, share and interact with its user and other smart devices.” Id.

10.   To his credit, Dr. Finegan conceded that his use of “genus” terminology was conveyed only “as a linguist” to explain “how dictionaries work,” not to provide a “legal opinion” about “the way genus is used in trademark law.” (Finegan, Vol. III at 96).

11.   The corpora included the Corpus of Historical American English (COHA), the Corpus of Contemporary American English (COCA), and Google Ngram. (EX 1-8, 1-9, 1-12; Leonard, Vol. I at 93). COHA contains about 475 million words from the 1820s to the 2010s with texts from television, movies, fiction, magazines, and newspapers. (EX 1-9, 1-10). COCA contains over 1 billion words from around 1990 to 2019 from similar sources but also blogs and web pages. (EX 1-14). Google Ngram is made up of words found in books written between the 1500s and 2019. (EX 1-13).

12.   The point of considering competitor use information, after all, is to draw an inference of consumer understanding. Because “conceptually” consumers do an “aisle search” for “the type of good they want” in the market and a “shelf search” for “the particular brand of that good,” the “genericism inquiry is therefore about assessing terms to determine whether they are related to an aisle search or a shelf search.” Laura A. Heymann, Trademarks in Conversation: Assessing Genericism After Booking.com, 39 Cardozo Arts & Ent. L.J. 955, 958 (2021). “Competitors, for their part, need to be able to use the term that will allow them to be grouped with comparable goods in the aisle search and to use trademarks as one method of winning the competition in the shelf search.” Id.

13.   The PTO's view is especially odd because the agency has previously taken the stance (shared by some courts but not adopted here) that “Teflon surveys are only appropriate to consider in a case where the question is whether a coined or arbitrary mark has become generic,” not “to prove recognition of an otherwise not inherently distinctive mark.” Frito-Lay N. Am., Inc. v. Princeton Vanguard, LLC, 124 U.S.P.Q.2d 1184, 2017 WL 3948367, at *13 (T.T.A.B. 2017).

14.   This is to say nothing of how the number of respondents who passed the mini tests of each survey is unvalidated in the first and intolerably low in the second. According to Anderson, “a total of 652 respondents completed the mini test” in the first survey and “512 (78.5%) passed.” (EX 271-31). But according to his “Termination and Removal Summary,” 502 respondents failed. (EX 271-281). That discrepancy is nowhere reconciled in his report, nor did the PTO address it at trial. Contrary to Anderson's assertion, then, the court cannot validate if the “passing rate” from the first survey “exceeds the minimum standards for a mini test.” (EX 271-31). Anderson himself testified that a high number of unexplained removals from a survey undermines its reliability. (EX 147-18, 171-20, 171-21). If the mini test passage rate is problematic in the first survey, it is near fatal in the second. Of the 412 respondents who took this mini test, only 217 passed—a pass rate of 52.7 percent. (EX 272-43, 272-45). If such “a substantial percentage of otherwise qualified survey respondents [are] excluded as ineligible for failing to correctly classify the qualifying test names,” the “court may give less weight to the survey as a measure of whether the disputed mark is generic.” Jay E. Deborah, Genericness Surveys in Trademark Disputes: Under the Gavel, in Trademark and Deceptive Advertising Surveys: Law, Science, and Design 107, 126 n.86 (Shari S. Diamond & Jerre E. Swann eds., 2d ed. 2022).

15.   It is not lost on the court that the Teflon surveys randomized the order of questions and of course didn't present the control names en masse as shown in the table. (EX 271-278). But the point is that no matter how well the survey questions were each ordered or presented to yield individually valid answers, if the overall structure and underlying design of the surveys focused only on testing consumers’ ability to sort between highly distinctive brand names and intrinsically nondistinctive generic names, then the surveys’ results can't be uncritically accepted at face value as a reliable measure of consumer understanding between purely nondistinctive generic and descriptive terms.

16.   Anderson himself was acutely aware of the danger posed by ambiguous survey questions: he criticized Sowers for using questions in his Thermos survey that could lead respondents to give product “descriptions” rather than product names. (Anderson, Vol. II at 169-70; EX 171-21, 147-48, 147-49, 147-50, 147-51). In Anderson's (correct) view, questions phrased to elicit descriptive terms—rather than words that name the type of product—cannot yield reliable evidence of whether consumers understand a term to be the common category name. (Anderson, Vol. II at 170-71). So too for Anderson's Teflon surveys. What's bad for the goose is bad for the gander.

17.   There is no need, as a result, to weigh the Teflon surveys against Snap's Thermos survey or Golder's “voice of the consumer” testimony derived from his historical research of Amazon product reviews and Google Search data. If the court were to consider that evidence, though, it would prevent the PTO from carrying its burden even by a preponderance of the evidence. For instance, while the court generally agrees with Dr. Anderson that most of the main Thermos survey questions were unhelpful (EX 171-17, 171-18, 171-19; Anderson, Vol. II at 168-69, 174), at least one question still asked a relevant question: what respondents would “tell the salesperson [they] wanted” or what they would “type into a search bar online” if they “wanted to purchase” smart glasses. Only one out of 273 qualified respondents said “spectacles” as a product name for that question. (EX 147-21). Similarly, Golder's historical research of how consumers review or search for smart glasses, even with its many flaws (Sood, Vol. II at 240-41, 247-48; Sood, Vol. III at 10), still underscored that “spectacles” is not used by relevant consumers to name the product; the term most frequently used instead was, unsurprisingly, “smart glasses.” (Golder, Vol. II at 15-20, 24, 28, 40-42, 45-53).

18.   Besides, the extra meanings Snap identifies—people making “spectacles” of themselves or capturing public “spectacles” while wearing smart glasses—still just describe desirable purposes of the SPECTACLES product. See, e.g., Gruner + Jahr USA Pub. v. Meredith Corp., 991 F.2d 1072, 1076 (2d Cir. 1993) (descriptive mark “may point to a product's intended purpose, its function or intended use, ․ or its merit”). What's more, according to Snap's own brand research, only 12 to 17 percent of surveyed consumers previously targeted by the company's Spectacles 3 advertising blitz said that they recognized “Capture your world in 3D” or “Share your perspective” as brand messages communicated by Snap's ads. (EX 75-18, 75-23). So it may be that the double meanings Snap contends makes “spectacles” suggestive is still too obtuse or hidden to register in its own target consumers’ minds, even with imagination.

19.   Snap's literal reading of Cross Commerce, which it relies on heavily, is hard to reconcile with the accepted test for distinctiveness: “whether ‘someone who knows what the goods and services are will understand the mark to convey information about them.’ ” 1 McCarthy § 11:21 (quoting In re Tower Tech, Inc., 64 U.S.P.Q.2d 1314, 2002 WL 992268, at *3 (T.T.A.B. 2002)). It is unclear whether Cross Commerce is endorsing some discredited “blank slate” approach to descriptiveness, which asks whether a mark conveys information “to one who has never seen the product and does not know what it is.” Id. Of course, if Cross Commerce is contrary to Zobmondo, the court is bound to follow the Ninth Circuit's decision. In any event, Cross Commerce is no help to Snap on its own terms because the dispute there was about whether the term “collective” applied to data analytics software was descriptive or suggestive. See 841 F.3d at 163. It is not hard to distinguish that contested term applied to something as divergent as software from a mark like SPECTACLES when applied to a product that calls to mind eyewear.

20.   Nor has Snap shown that SPECTACLES is suggestive under the alternative to the “imagination” test—the “competitor need” test. Under that test, if a mark conveys an association with a product or service “so direct and clear that competing sellers would be likely to need to use the term in describing or advertising their goods or services,” then the term is not suggestive but descriptive. Rodeo Collection, Ltd. v. W. Seventh, 812 F.2d 1215, 1218 (9th Cir. 1987) (cleaned up). As discussed, there is abundant evidence of competitors and third parties using “spectacles” (as well as other similar terms for eyewear) to describe smart glasses. (A2381-2407; A2866-867; EX 87, 88, 90-7, 91-1, 92, 93-1, 94-1, 97-1, 104-1, 108, 111-51, 111-52, 111-53, 111-54, 111-55, 111-56). That evidence only further undermines Snap's claim to the suggestiveness of SPECTACLES. See, e.g., SportFuel, Inc. v. PepsiCo, Inc., 932 F.3d 589, 599-600 (7th Cir. 2019) (finding “sport fuel” descriptive for athletic nutritional products because most producers generally used that term to also describe their products); TMEP § 212.01 (If “third parties in applicant's field or closely related fields use the same or substantially the same wording as the mark, or very similar wording as the mark,” that usage “tends to indicate the mark is at least highly descriptive.”).

21.   Snap's retained marketing expert testified that he was unaware of any Snap secondary-meaning consumer survey “that was done properly to track consumer perception of the term ‘spectacles’.” (Golder, Vol. II at 99).

22.   Similarly, the PTO's Teflon surveys, even with their flaws as to genericness, showed at most 23.8 percent source identification of SPECTACLES as a brand name among respondents. (EX 272-16).

23.   To that end, the parties are asked to lodge a stipulated form of final judgment consistent with this order to be entered no later than 14 days from this order.

STEVE KIM, United States Magistrate Judge