Misguided: USPTO Examination Guide Misses the Mark on Booking.com

“The Guide pays lip service to the Supreme Court’s ruling, but instead of instructing examiners to focus exclusively on evidence showing the significance of the mark to consumers, slips back towards the USPTO’s discredited per se rule by emphasizing the importance of factors other than consumer perception.”

Editor’s Note: The authors served as co-counsel to Booking.com B.V. in its successful appeal before the Supreme Court.

https://depositphotos.com/92371636/stock-photo-definition-of-the-word-misguided.htmlLast week, the U.S. Patent and Trademark Office (USPTO) released its long-awaited Examination Guide on so-called generic.coms – domain names comprised of generic elements along with a generic top-level domain (such “gTLDs” include .com, .net, .org, .biz and .info). The Guide (No. 3-20, entitled “Generic Terms after USPTO v. Booking.com”) provides needed guidance to trademark examiners on how to apply the U.S. Supreme Court’s decision in USPTO v. Booking.com B. V., 140 S. Ct. 2298 (2020); it also provides guidance to trademark applicants on the standards they can expect the USPTO to apply in considering whether their domain names are registrable as trademarks. Unfortunately, instead of faithfully applying the Supreme Court’s lesson about the importance of consumer perception in assessing whether a term functions as a trademark, the USPTO has relied on factors close to its discredited per se rule that will make it very difficult to register such marks.

The Booking.com Decision

In Booking.com, the Court rejected the USPTO’s per se rule that the addition of a gTLD to a generic term does nothing to change the generic nature of the proposed mark. In support of that proposed per se rule, the USPTO relied on an 1888 case in which the Court had held more than a century ago that the addition of “Company” to an otherwise generic term (there, Goodyear, which at the time was generic for a process for making rubber) does not transform the combined name into a trademark. Goodyear’s India Rubber Glove Mfg. Co. v. Goodyear Rubber Co., 128 U.S. 598 (1888).

The Goodyear decision was bad law for many reasons. It was written at a time when U.S. trademark law did not distinguish between generic and descriptive marks and did not allow either to be registered (today, of course, descriptive marks are registrable upon a showing of secondary meaning). Moreover, by the time the Lanham Act was adopted nearly 60 years later, the law had evolved to make clear that trademarks should be assessed based on their primary significance to consumers. For that reason, the Court held (in an 8-1 decision by Justice Ginsburg, sadly, her last decision for the Court) in assessing whether a domain name (or any alleged mark for that matter) functions as a trademark or is a generic reference, the touchstone is the meaning of the term to consumers. The USPTO’s proposed per se rule – which was based on the dictionary meaning of the constituent parts of the mark, rather than the significance of the mark as a whole to consumers – was fundamentally inconsistent with the Lanham Act’s primary significance test. See 15 U.S.C. § 1064(3) (“The primary significance of the registered mark to the relevant public . . . shall be the test for determining whether the registered mark has become the generic name of goods or services.”).


The Rush of New Trademark Applications

In the four months since the Supreme Court’s June 30, 2020 decision, applicants have filed well over one thousand applications seeking to register domain names as trademarks. These applications include such domain names as law.com, wine.com, porn.com, cruises.com, business.com, sneakers.com, locals.org, wedogood.org, familydoctor.org, howtoadopt.org, and liquidity.net.

Only a few of these applications have been examined – in part because examiners were waiting for guidance on how to apply Booking.com given the USPTO’s prior policy of treating such domain names as generic. The vast majority of those, including beauty.com, sleep.com, recruiter.com, patentassociate.com and webinar.net, have received initial refusals based, inter alia, on a finding that the domain name appeared to be merely descriptive. In each case, the examiners cited the same pre-Booking.com case law that the USPTO has relied on in the past. For example, in refusing to register sleep.com, the examiner stated:

The applied-for mark consists of the wording “SLEEP”, combined with the non-source-identifying generic top-level domain (gTLD) “.COM.”  The wording “sleep” is merely descriptive of a feature of the services as specified, namely “Providing a website featuring information in the field of sleep; providing a website featuring information in the field of sleep products, namely, mattresses, mattress foundations, pillows, and adjustable bases for mattresses.”  This shows such wording is merely descriptive of applicant’s services because it directly conveys the area or field of information provided.

Further, the attached evidence from https://www.ahdictionary.com/word/search.html?q=COM shows that the gTLD “.COM” is defined as “commercial organization (in Internet addresses).”  This evidence shows that the gTLD is merely descriptive of applicant’s services because it directly conveys that the services are offered on the Internet in association with the mark.

Additionally, a gTLD generally serves no source-indicating function because it will usually be perceived only as part of an Internet address; thus, adding it to an otherwise unregistrable mark typically does not render the mark registrable. TMEP §1209.03(m); see In re Oppedahl & Larson LLP, 373 F.3d 1171, 1177, 71 USPQ2d 1370, 1374 (Fed. Cir. 2004); In re Microsoft Corp., 68 USPQ2d 1195, 1202-03 (TTAB 2003); TMEP §1215.04; see also In re Steelbuilding.com, 415 F.3d 1293, 1297, 75 USPQ2d 1420, 1422 (Fed. Cir. 2005) (“Only in rare instances will the addition of a [g]TLD indicator to a descriptive term operate to create a distinctive mark.”).

In the present case, the non-source-identifying gTLD, when combined with the descriptive wording in the mark, adds no source-identifying significance but retains its significance as indicating an Internet address only. Accordingly, the combination of the merely descriptive wording “sleep” coupled with the non-source-identifying gTLD “.COM” is merely descriptive of applicant’s services because this combination immediately informs prospective purchasers that applicant’s services are offered about “sleep” and are from an Internet business.

Some, like webinar.net, have also received refusals based on the potential for the domain name to be generic:

In addition to being merely descriptive, the applied-for mark appears to be generic in connection with the identified goods and/or services. “A generic mark, being the ‘ultimate in descriptiveness,’ cannot acquire distinctiveness” and thus is not entitled to registration on either the Principal or Supplemental Register under any circumstances. In re La. Fish Fry Prods., Ltd., 797 F.3d 1332, 1336, 116 USPQ2d 1262, 1264 (Fed. Cir. 2015) (quoting H. Marvin Ginn Corp. v. Int’l Ass’n of Fire Chiefs, Inc., 782 F.2d 987, 989, 228 USPQ 528, 530 (Fed. Cir. 1986)); see TMEP §§1209.01(c) et seq., 1209.02(a). Therefore, the trademark examining attorney cannot recommend that applicant amend the application to proceed under Trademark Act Section 2(f) or on the Supplemental Register as possible response options to this refusal. See TMEP §1209.01(c).

These rationales continue to focus, inappropriately, on the meaning of the elements of the domain name – the term “sleep” or “webinar” combined with the gTLD – without considering, as Booking.com requires, how consumers understand the term when considered as a whole.

The New Examination Guide

With its new Examination Guide, the USPTO has now provided trademark examiners with the Office’s interpretation of Booking.com, and how it should be applied to new applications. The Guide pays lip service to the Supreme Court’s ruling that “[w]hether any given ‘generic.com’ term is generic . . . depends on whether consumers in fact perceive that term as the name of a class or, instead, as a term capable of distinguishing among members of the class.” But instead of instructing examiners to focus exclusively on evidence showing the significance of the mark to consumers, the Guide slips back towards the USPTO’s discredited per se rule by emphasizing the importance of the factors other than consumer perception.

The Guide begins by noting that the Court in Booking.com rejected the USPTO’s “proposed per se rule that generic.com terms are automatically generic” and also “specifically declined to adopt a rule that these terms are automatically non-generic.” Instead, the Guide explains, trademark examiners must “assess on a case-by-case basis whether, based on the evidence of record, consumers would perceive a generic.com term as the name of a class of goods and/or services or, instead, as at least capable of serving as a source indicator.”

That statement of the principle is correct, but the Guide goes off the rails when it outlines the following four “examples of the types of evidence that may support the conclusion that consumers would perceive the generic.com term, as a whole, as the name of the class of goods and/or services” (i.e., as being generic):

  • a combination of dictionary excerpts defining the component elements of the generic.com term;
  • significant evidence of generic usage of those elements or the combined term by consumers or competitors in the relevant marketplace;
  • evidence of the “generic.com” term used by third parties as part of their domain names (e.g., “[adjective]generic.com”) in connection with the same or similar goods and/or services; or
  • evidence of the applicant’s own use of the generic.com term.

Three of these four examples are at odds with the Booking.com decision.

The Guide is Misguided

First, the Guide’s reliance on using a combination of dictionary excerpts to assess the meaning of the mark is incompatible with a core principle the Court articulated in Booking.com: the mark must be considered as a whole. As Justice Ginsburg explained, “for a compound term, the distinctiveness inquiry trains on the term’s meaning as a whole, not its parts in isolation.”  Examiners should not be encouraged to start the analysis by assessing the dictionary meaning of the constituent parts of the trademark, since that analysis will inevitably lead to a finding that the mark as a whole is generic. That is exactly the same error of law that the USPTO made in Frito-Lay N. Am., Inc. v. Princeton Vanguard LLC, 109 U.S.P.Q. 2d 1949 (TTAB 2014), where the TTAB ruled that the mark PRETZEL CRISPS must be generic because each of its constituent parts – “pretzel” and “crisps” – are generic terms. As the Federal Circuit explained when it reversed that decision, the TTAB was required to consider genericness of the “PRETZEL CRISPS” mark in its entirety, rather than assessing whether each component term was generic. Princeton Vanguard, LLC v. Frito-Lay N. Am., Inc., 786 F.3d 960, 966-67 (Fed. Cir. 2015). This principal has been ignored by the Guide.

Moreover, and in general, the Guide’s emphasis on dictionary terms lends disproportionate importance to dictionary evidence in general. As the Fourth Circuit in Booking.com noted, it is more appropriate to place greater weight on more valuable consumer perception evidence, like properly designed consumer surveys, than evidence such as dictionary definitions when assessing the primary significance of a term to the public. Booking.com B.V. v. USPTO, 915 F.3d 171, 185 (4th Cir. 2019), aff’d, 140 S. Ct. 2298 (2020).

The Guide then compounds this error by instructing examiners to consider whether there is “significant evidence of generic usage of those elements or the combined term by consumers or competitors in the relevant marketplace.”Applying that test to the BOOKING.COM mark would inevitably have led to a finding that BOOKING.COM is generic, since the “booking” element is used by consumers and competitors in the marketing for making online reservations. This example from the Guide will have the effect of resurrecting through the back door the very per se rule that the Supreme Court rejected.

The Guide’s third example is equally flawed. By suggesting that examiners should consider “evidence of the ‘generic.com’ term used by third parties as part of their domain names in connection with the same or similar goods and/or services,” the USPTO is again endorsing an approach that disregards consumers’ perception of the mark as a whole. Once again, a review of the evidence before the Supreme Court is instructive. In Booking.com, the parties noted that other domain names containing “booking” were registered as trademarks included Ebooking.com (Reg. No. 3,888,087), Fish & Hunt Booking.com (Reg. No. 4,988,037), and Voicebooking.com (Reg. No. 5,352,745). The fact that each of these marks included as a component the term “booking.com” is not evidence that BOOKING.COM is not a trademark; to the contrary, it is evidence that consumers can distinguish different marks containing “booking.com” and that the registration of BOOKING.COM as a trademark will not be anti-competitive, as Justice Breyer worried in dissent (a point we addressed in more detail here). This evidence does nothing to inform the question of whether consumers perceive BOOKING.COM to be a brand name that signifies services from a single source, or whether they perceive it to be the genus of the category of travel reservation websites.

What Evidence Should Applicants Submit?

Although the Guide appears to perpetuate the USPTO’s hostility to registering domain names as trademarks, the Guide does properly highlight some of the evidence that may be relevant to the assessment of consumers’ perception, both as to whether a term is generic and as to whether a descriptive mark has acquired distinctiveness. That evidence may include:

  • properly designed consumer surveys (see generally our earlier article on the importance of survey evidence);
  • consumer declarations;
  • evidence showing the duration, extent, and nature of the applicant’s use of the proposed mark, including the degree of exclusivity of use, related advertising expenditures, and letters or statements from the trade or public; and
  • any other appropriate evidence tending to show that the proposed mark distinguishes the goods or services to consumers.

Applicants looking to overcome an initial refusal to register should focus on all of this evidence in their responses to office actions. Moreover, the USPTO’s recognition that there may be “other appropriate evidence” of the primary significance of the mark to consumers as well opens the door to other relevant evidence. For example, applicants should consider the availability of “voice of consumer” evidence of the kind outlined by Professor Peter Golder in his amicus brief before the Supreme Court. This evidence may include consumer testimonials, online reviews, discussion boards, social media postings, online search queries and consumer engagement metrics from search engines.

Gathering and submitting such evidence will not be inexpensive. Applicants seeking to register their domain names as trademarks should be prepared to make a compelling showing that the primary significance of the domain name to consumers is as a brand name rather than as a generic or merely descriptive reference to a category of websites. But the USPTO should not make this process even more difficult by erecting unnecessary obstacles based on artificial rules that are inconsistent with the dictates of the Supreme Court.

Failure to Function

Although the “failure to function” doctrine was not at issue in Booking.com, the USPTO took the opportunity of the Guide to reemphasize its instruction that examiners should refuse registration if an applied for mark fails to function as a trademark. This long has been an important bulwark against registration of marks that do not indicate source to consumers (see, e.g., Section 1202.17(c) of the Trademark Manual of Examining Procedure), and has become even more important since Tam and Brunetti (which, predictably, led to similar rushes to registration after the Court struck the Lanham Act’s prohibition of registration of disparaging and scandalous marks, respectively). Indeed, one of the principal ways that the USPTO has been able to hold back a tsunami of registrations of offensive marks has been to deny many of these applications on failure to function grounds. For example, three months after the Court invalidated the Lanham Act’s prohibition on registration of scandalous marks in Iancu v. Brunetti, 588 U.S. __, 139 S. Ct. 2294 (2019), the USPTO refused registration of a swastika symbol on failure to function grounds. Trademark Application Serial No. 87,503,998 (Office Action, Sept. 19, 2019). Most recently, the USPTO again relied on the “failure to function” doctrine to refuse registration of the .SUCKS gTLD as a trademark. In re Vox Populi Registry Ltd., ___ U.S.P.Q.2d ___ (TTAB Oct. 29, 2020).

This is one part of the Guide that we believe is both important and appropriate, since many domain names will simply be that – domain names that identify the location on the internet of a website rather than a brand name. Both the Guide and the Court’s decision in Booking.com make clear that domain names are not automatically eligible for registration; they only may be registered if consumers perceive the domain names as brands indicating source, not simply URLs indicating address.

Image Source: Deposit Photos
Author: Devon
Image ID: 92371636


Warning & Disclaimer: The pages, articles and comments on IPWatchdog.com do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of IPWatchdog.com. Read more.

Join the Discussion

2 comments so far.

  • [Avatar for Bernat]
    November 13, 2020 04:22 am

    Thanks for expressing this statement and rendering it public

  • [Avatar for Anon]
    November 3, 2020 06:19 pm


    I think that the “Failure to Function” view actually highlights a reason why every web address COULD be pursued for trademark protection: it is a matter of pure function that a web address IS a sole source indicator because that is how web addresses work. Type in a web address and you ONLY go to one source.

    No chance of confusion (on a specific identity basis).

    Anyone that could not compose a survey that would provide any necessary level of support for trademark purposes should not be in the survey business.