Plaintiff Booking.com B.V. ("Booking.com" or "plaintiff") filed this civil action challenging the denial by the Trademark Trial and Appeal Board ("TTAB") of the United States Patent and Trademark Office ("USPTO") of four trademark applications involving the mark "BOOKING.COM" for services in Classes 39 and 43. One of the applications was for the word mark and three were for stylized versions of the mark. For each of the applications, the TTAB found plaintiff's marks ineligible for registration as trademarks because it concluded that BOOKING.COM is generic for the services identified in the applications or, alternatively, that it is merely descriptive and lacks acquired distinctiveness.
Before the Court are plaintiff and defendants' cross-motions for summary judgment. For the reasons that follow, plaintiff's Motion for Summary Judgment [Dkt. No. 63] will be granted in part and denied in part; defendants' Motion for Summary Judgment [Dkt. No. 60] will be granted in part and denied in part; and the USPTO will be ordered to register the mark BOOKING.COM as to the Class 43 services identified in plaintiff's applications but not as to the Class 39 services.
On December 1, 2011, plaintiff filed a federal trademark application, Serial No. 85485097 ("'097 Application"), based on use, for the mark:
A3166. The services identified in the application (as amended) were:
Class 39: Travel agency services, namely, making reservations for transportation; travel and tour ticket reservation services; travel agency services, namely making reservations for transportation for tourists; provision of travel information; providing consultation related to making reservations for transportation, and travel and tour ticket reservation; all of the foregoing services rendered in-person and via the internet.
Class 43: Making hotel reservations for others in person and via the internet; providing personalized information about hotels and temporary accommodations for travel in-person and via the Internet; providing on-line reviews of hotels; consultation services related to making hotel reservations for others, provision of personalized information about hotels and temporary accommodations for travel, and on-line reviews of hotels.
Moskin Decl. [Dkt. No. 65-5] ¶ 2.
On June 5, 2012, plaintiff filed Application Serial No. 7911498 ("'998 Application"), for recognition in the United States of its International Registration (hereinafter referred to by the name of the authorizing treaty, "the Madrid Protocol") for the mark:
A4. The services identified in the application (as amended) were:
Class 39: Arranging of tours and arranging of tours online; reservation and sale of travel tickets and online reservation and sale of travel tickets; information, advice and consultancy regarding the arranging of tours and the reservation and sale of travel tickets; provision of information relating to travel and travel destinations; travel and tour agency services, namely, travel and tour ticket reservation services; travel agency services; tourist agency services; providing online travel and tourism services, namely, providing online travel and tour ticket reservation services, online travel agency services, online tourist agency services and providing online information relating to travel and travel destinations.
Class 43: Making hotel reservations for others; holiday accommodation reservation services and resort reservation services, namely providing online hotel and resort hotel room reservation services; providing information about hotels, holiday accommodations and resort accommodations, whether or not based on the valuation of customers; providing information, advice and consultancy relating to making hotel reservations and temporary accommodation reservations; providing online information, advice and consultancy relating to making hotel reservations and temporary accommodation reservations.
Moskin Decl. [Dkt. No. 65-5] ¶ 3.
On November 7, 2012, plaintiff filed two federal trademark applications, Serial No. 79122365 ("'365 Application") and Serial No. 79122366 ("'366 Application"), under the Madrid Protocol for the following marks:
A2153, A1138. The services identified in the two applications (as amended) were limited to a subset of services in Class 43:
Hotel reservation services for others; holiday accommodation reservation services and resort reservation services, namely, providing hotel room reservation services and resort hotel reservation services and providing online hotel and resort hotel room reservation services; providing information about hotels, hotel accommodations and resort accommodations, whether or not based on the valuation of customers; information, advice and consultancy relating to the aforesaid services; aforesaid services also provided electronically.
Moskin Decl. [Dkt. No. 65-5] ¶ 4.
During review by the USPTO, all four applications followed the same procedural history. The examiner initially rejected each application on the ground that BOOKING.COM is merely descriptive of plaintiff's services and therefore unregisterable. A1074, A2089, A3765. After plaintiff objected that the mark BOOKING.COM had acquired distinctiveness, the examiner issued a new refusal, this time on the basis that the word mark is generic as applied to the relevant services and, in the alternative, that the mark is merely descriptive and that plaintiff had failed to establish acquired distinctiveness. A1074, A2089-90, A3766. For each application, plaintiff sought reconsideration of the new refusal and in each instance reconsideration was denied. A1075, A2090, A3766.
Plaintiff filed a Notice of Appeal for each application and requested consolidated briefing before the TTAB, which was granted. A3766. The evidence submitted to the TTAB included dictionary definitions of the words "booking" and ".com;" print-outs of plaintiff's webpages; examples from news articles and travel websites of terms such as "online booking services" and "booking sites," used to refer to hotel reservation and travel agency services; examples of eight third-party domain names that include "booking.com;" a 2012 JD Power & Associates press release and survey results, indicating that Booking.com ranked highest in overall customer satisfaction; and a declaration from plaintiff's director listing awards won by plaintiff and figures regarding plaintiff's sales success, advertising campaigns, followers on social media, and unsolicited news articles. See Def. Mem. at 6; A1089-92.
Following the hearing, the TTAB affirmed the four refusals of registration in three separate opinions. See A1073-111 (denying the appeal for the '998 Application), A2088-126 (denying the appeals for the '365 and '366 Applications), A3764-801 (denying the appeal for the '097 Application). Although there are minor differences among the three opinions, all share the same central conclusions that "booking" refers to "a reservation or arrangement to buy a travel ticket or stay in a hotel room" or "the act of reserving such travel or accommodation;" that ".com" indicates a commercial website, which does not negate the generic character of the term "booking;" and that the combined term BOOKING.COM would be understood by consumers "primarily to refer to an online reservation service for travel, tours, and lodging," which is consistent with the services proposed in the applications, making the mark generic for the services offered. See, e.g., A1092, A1096, A1107. In the alternative, the TTAB concluded that BOOKING.COM is descriptive of plaintiff's services and that plaintiff "failed to demonstrate that the term has acquired distinctiveness." See, e.g., A1111.
On April 15, 2016, plaintiff filed this civil action under 15 U.S.C. § 1071(b) against Michelle Lee, who was then the USPTO Director ("the USPTO Director"),1 and the USPTO (collectively "defendants"), challenging the USPTO's denial of registration of the four applications. The parties have filed the administrative record from the USPTO proceedings and both sides have produced new evidence on the questions of genericness and descriptiveness. Although the body of evidence before this Court is similar to what was before the TTAB, of significance, plaintiff has now submitted a "Teflon survey," which, as will be discussed below, is the most widely used survey format for measuring consumer opinion in a genericness challenge, and defendants have provided a report by a rebuttal expert. By way of relief, plaintiff asks the Court to reverse the decisions of the TTAB and order the USPTO Director to publish each application in the Principal Register. Compl., [Dkt. No. 1] at 17.
A. Standards of Review
A trademark applicant "dissatisfied with the decision" of the USPTO has two remedies under the Lanham Act: either "appeal to the United States Court of Appeals for the Federal Circuit," see 15 U.S.C. § 1071(a), or file a civil action against the USPTO Director in federal district court, see 15 U.S.C. § 1071(b). Under § 1071(a), an appeal to the Federal Circuit is taken "on the record" before the USPTO, id. § 1071(a)(4), and the USPTO's factual findings will be upheld if they are supported by "substantial evidence," see, e.g., Recot, Inc. v. Becton, 214 F.3d 1322, 1327 (Fed. Cir. 2000). In contrast, in a civil action under § 1071(b), "the district court reviews the record de novo and acts as the finder of fact." Swatch AG v. Beehive Wholesale, LLC, 739 F.3d 150, 155 (4th Cir. 2014) (citing Durox Co. v. Duron Paint Mfg. Co., 320 F.2d 882, 883-84 (4th Cir. 1963)).2 Placement of a mark on the generic-descriptive-suggestive-fanciful continuum is a question of fact. In re Dial-a-Mattress Operating Corp., 240 F.3d 1341, 1344 (Fed. Cir. 2001).
Upon the motion of a party, the district court must admit the USPTO record and give it the "same effect as if originally taken and produced in the suit." § 1071(b)(3). "[T]he district court may, in its discretion, `consider the proceedings before and findings of the [USPTO] in deciding what weight to afford an applicant's newly-admitted evidence.'" Kappos v. Hyatt, 566 U.S. 431, 132 S.Ct. 1690, 1700, 182 L.Ed.2d 704 (2012) (quoting Hyatt v. Kappos, 625 F.3d 1320, 1335 (Fed. Cir. 2010)). The district court also "has authority independent of the [USPTO] to grant or cancel registrations." Swatch AG, 739 F.3d at 155 (citing § 1071(b)(1)).
Summary judgment is appropriate where the record demonstrates that "there is no genuine dispute as to any material fact and that the moving party is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(a). Although the Court must view the record "in the light most favorable to the non-moving party," Dulaney v. Packaging Corp. of Am., 673 F.3d 323, 324 (4th Cir. 2012), "[t]he mere existence of a scintilla of evidence in support of the [non-movant's] position will be insufficient" to overcome a motion for summary judgment. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 252, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986); see also Am. Arms Int'l v. Herbert, 563 F.3d 78, 82 (4th Cir. 2009). Rather, a genuine issue of material fact exists only "if the evidence is such that a reasonable jury could return a verdict for the nonmoving party." Anderson, 477 U.S. at 248, 106 S.Ct. 2505. Summary judgment does not become disfavored simply because there is an "important, difficult or complicated question of law." Lewis v. Coleman, 257 F.Supp. 38, 40 (S.D. W. Va. 1966); Bradacs v. Haley, 58 F.Supp.3d 514, 521 (D.S.C. 2014).
The parties have "expressly agree[d] that if the Court determines after reviewing the briefs and evidence on summary judgment that any material issue of fact exists, the Court is authorized to resolve any such factual dispute." [Dkt. No. 26] ¶ 4(B). Accordingly, the Court will make factual determinations as well as weighting decisions that are not normally appropriate on a motion for summary judgment.
Although plaintiff filed four trademark applications, neither plaintiff nor defendants contend that the stylized elements described in those applications affect the protectability of the mark. Instead, the parties focus on the word mark BOOKING.COM and on where along the generic-descriptive-suggestive-fanciful continuum the mark is situated. Def. Mem. at 1-2; Pl. Mem. at 10-12. Therefore, rather than addressing each application individually, the Court will disregard the stylized elements and focus on the appropriate categorization of the word mark BOOKING.COM; however, because a multi-class application is regarded as a series of separate applications, the Court must independently assess the protectability of the mark for the two classes of services plaintiff claims in its applications, Classes 39 and 43. See 3 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 19:56.50 (4th ed.) (hereinafter McCarthy on Trademarks).
1. The Framework of the Lanham Act
The Lanham Act provides nationwide protection of trademarks. A trademark is "any word, name, symbol, or device, or any combination thereof" used "to identify and distinguish ... goods [or services], including a unique product [or service], from those manufactured or sold by others and to indicate the source of the goods [or services], even if that source is unknown." 15 U.S.C. § 1127. The Act has two purposes. The first is to prevent consumer confusion regarding the source of goods and services and to reduce consumers' information costs by "quickly and easily assur[ing] a potential customer that this item — the item with this mark — is made by the same producer as other similarly marked items [or services] that he or she liked (or disliked) in the past." Qualitex Co. v. Jacobson Prod. Co., 514 U.S. 159, 163-64, 115 S.Ct. 1300, 131 L.Ed.2d 248 (1995) (emphasis in original). Second, the Act incentivizes brand investment by assuring the "producer that it (and not an imitating competitor) will reap the financial, reputation-related rewards associated with a desirable product [or service]," id. at 164, 115 S.Ct. 1300, thereby "secur[ing] to the owner of the mark the goodwill of his business," Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189, 198, 105 S.Ct. 658, 83 L.Ed.2d 582 (1985). By allowing the producer to reap the benefits of consumer goodwill, trademark law "encourages the production of quality products [and services] and simultaneously discourages those who hope to sell inferior products [or services] by capitalizing on a consumer's inability quickly to evaluate the quality of an item [or service] offered for sale." Qualitex, 514 U.S. at 164, 115 S.Ct. 1300 (internal quotation marks omitted). Thus, the Lanham Act reflects Congress's conclusion that "[n]ational protection of trademarks is desirable ... because [it] foster[s] competition and the maintenance of quality by securing to the producer the benefits of good reputation." Park 'N Fly, 469 U.S. at 198, 105 S.Ct. 658.
In keeping with these twin purposes, the Lanham Act identifies four categories of marks. "Arrayed in an ascending order which roughly reflects their eligibility [for] trademark status and the degree of protection accorded, these classes are (1) generic, (2) descriptive, (3) suggestive, and (4) arbitrary or fanciful." Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4, 9 (2d Cir. 1976). "A generic mark refers to the genus or class of which a particular product [or service] is a member and can never be protected." Ashley Furniture Indus., Inc. v. SanGiacomo N.A., 187 F.3d 363, 369 (4th Cir. 1999). Examples include Light Beer for ale-type beverages and Thermos for vacuum-insulated bottles. Sara Lee Corp. v. Kayser-Roth Corp., 81 F.3d 455, 464 (4th Cir. 1996). A descriptive mark "describes a function, use, characteristic, size, or intended purpose" of the product or service, such as 5 Minute glue and the Yellow Pages telephone directory. Id. "Marks that are merely descriptive are accorded protection only if they have acquired a secondary meaning [also called `acquired distinctiveness'], that is, if in the minds of the public, the primary significance of a product [or service] feature or term is to identify the source of the product [or service] rather than the product [or service] itself." Id. (internal citations omitted). Although eligible for protection in some instances, descriptive marks are considered weak marks. Shakespeare Co. v. Silstar Corp. of Am., 110 F.3d 234, 239-40 (4th Cir. 1997). Suggestive marks, such as Coppertone for sunscreen and Orange Crush for orange flavored soda, "connote, without describing, some quality, ingredient, or characteristic of the product [or service]." Sara Lee Corp., 81 F.3d at 464. Marks that are "comprised of words in common usage" but "do not suggest or describe any quality, ingredient, or characteristic of the goods [or services] they serve, are said to have been arbitrarily assigned." Id. Examples of arbitrary marks include Tea Rose brand flour and Apple for computers. Id. Lastly, fanciful marks are "in essence, made-up words expressly coined for serving as a trademark," such as Clorox for a bleach product and Kodak for photography-related products. Id. Because the "intrinsic nature" of suggestive, arbitrary, and fanciful marks "serves to identify a particular source of a product [or service]," these categories "are deemed inherently distinctive and are entitled to protection." Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992).
Because a generic mark, which is statutorily defined as "the common descriptive name of an article or substance," Federal Trade Commission Improvements Act of 1980, Pub. L. No. 96-252, § 18, 94 Stat. 374, 391, by definition neither signifies the source of goods or services nor distinguishes the particular product or service from other products or services on the market, it cannot be protected as a trademark nor registered as one. Retail Servs., Inc. v. Freebies Publ'g, 364 F.3d 535, 538 (4th Cir. 2004) (citing Park'N Fly, 469 U.S. at 194, 105 S.Ct. 658). To permit otherwise "would grant the owner of the [generic] mark a monopoly since a competitor could not describe his goods [or services] as what they are." CES Publ'g Corp. v. St. Regis Publ'ns, Inc., 531 F.2d 11, 13 (2d Cir. 1975); see also Renaissance Greeting Cards, Inc. v. Dollar Tree Stores, Inc., 405 F.Supp.2d 680, 688 (E.D. Va. 2005), aff'd, 227 Fed.Appx. 239 (4th Cir. 2007) ("[T]he goals of trademark protection... must be balanced by the concern that trademark protection not become a means of monopolizing language or stifling productive competition."). To the contrary, such marks must remain in the public domain where they are free for all to use. See Am. Online, Inc. v. AT & T Corp., 243 F.3d 812, 821 (4th Cir. 2001) (explaining that trademark law "protects for public use those commonly used words and phrases that the public has adopted, denying to any one competitor a right to corner those words and phrases by expropriating them from the public `linguistic commons'"); 2 McCarthy on Trademarks § 12:2.
"The rub ...is in trying to distinguish generic marks from [protectable marks]." Ashley Furniture Indus., 187 F.3d at 369. According to the test adopted by the Supreme Court in Kellogg Co. v. Nat'l Biscuit Co., a plaintiff seeking to establish a valid trademark as compared to a generic mark "must show that the primary significance of the term in the minds of the consuming public is not the product but the producer." 305 U.S. 111, 118, 59 S.Ct. 109, 83 S.Ct. 73 (1938). A mark is not generic simply because it plays some role in denoting to the public what the product or service is; rather, a mark may serve a "dual function — that of identifying a product [or service] while at the same time indicating its source." S. Rep. No. 98-627, at 5 (1984). Hence, Kellogg focuses on whether "the primary significance of the mark [is] indication of the nature or class of the product or service, rather than an indication of source." Glover v. Ampak, Inc., 74 F.3d 57, 59 (4th Cir. 1996) (emphasis in original); see also Retail Servs., 364 F.3d at 544 (explaining that a generic mark "neither signifies the source of goods nor distinguishes the particular product from other products on the market").
Determining whether a mark is generic involves three steps: "(1) identify[ing] the class of product or service to which use of the mark is relevant; (2) identify[ing] the relevant purchasing public of the class of product or service; and (3) [determining whether] the primary significance of the mark to the relevant public is to identify the class of product or service to which the mark relates." Glover, 74 F.3d at 59. Evidence of public understanding of the primary significance of a mark can come from "purchaser testimony, consumer surveys, listings and dictionaries, trade journals, newspapers, and other publications." Id. The burden of proof rests with the party seeking to establish genericness, in this case the defendants, who must prove that the mark is generic by clear and convincing evidence. In re Cordua Restaurants, Inc., 823 F.3d 594, 600 (Fed. Cir. 2016).
a. Classes of Services
The first step in analyzing the proposed BOOKING.COM mark is to determine the classes of services (sometimes referred to as "genera of services") at issue in each application. Glover, 74 F.3d at 59. The defendants recognize that the services identified in each registration vary, but summarize the classes of services as "online travel agency services, namely the arrangement of transportation and tours," for Class 39 and "online hotel and lodging services" for Class 43. Def. Mem. at 17-18. Plaintiff does not appear to understand that this is a class specific analysis and argues that the USPTO's "inability to adopt a single genus ...requires reversal." See Pl. Reply at 16-17 & n.5. This argument is inconsistent with the longstanding principle that a single application to register multiple classes, i.e., a combined application, is treated "as though it were a group of individual applications" requiring "separate analyses for each class of goods [or services]." Federated Foods, Inc. v. Fort Howard Paper Co., 544 F.2d 1098, 1102 (C.C.P.A. 1976).
Plaintiff further criticizes the defendants' "new proposed genera" for "ignor[ing] most of plaintiff's actual services," including "the information and search (or research) services ... and business oriented services ... used by hotels and other travel services seeking to advertise and list their accommodations for rental." Pl. Opp. at 28. This critique is baseless for several reasons. As an initial matter, plaintiff overlooks that it too has taken a reductivist view of the class of service by defining the relevant class of services or genus as "travel agency services." Pl. Mem. at 13. Further, plaintiff's suggestion that the class of services should include specific services, such as business consulting, stems from plaintiff's argument that a granular approach that emphasizes services that would not be described with the word "booking" can save its mark from genericness. See Pl. Opp. at 28 n.18 ("[T]here is no reason a broad specification of services or goods cannot be held distinctive for some of the services and not so for others."). Defendants, on the other hand, maintain that "registration is properly refused if genericness is found as to any service specified in the application." Def. Reply at 9. On this issue, defendants have the better of the argument. Registration must be refused if a mark "is the generic name of any of the goods or services for which registration is sought." See Cordua Restaurants, 823 F.3d at 605 (quoting 2 McCarthy on Trademarks § 12:57). Therefore, even though plaintiff provides business consulting services, its mark will fail if it is generic as to plaintiff's hotel reservation services. Similarly, even if, as plaintiff argues, the appropriate genus is "travel agency services," to the extent that this encompasses hotel reservation services, plaintiff's mark is not entitled to protection if it is generic for hotel reservation services. See Otokoyama Co. v. Wine of Japan Imp., Inc., 175 F.3d 266, 271 (2d Cir. 1999) ("Generic words for sub-classifications or varieties of a good are ... ineligible for trademark protection.").
In addition, as defendants acknowledge, "[p]laintiff is the master of its application, including the identification of services covered by it" and "[i]t was from the identification of services provided by plaintiff that the TTAB derived its recitation of the services." Def. Opp. at 10; see also A1076 (describing the genus of services as an "accurate summar[y]" of the recitation provided by the plaintiff); A3768 (same). Because "the question of registrability of an applicant's mark must be decided on the basis of the identification of goods [and services] set forth in the application regardless of what the record may reveal as to the particular nature of an applicant's goods," Octocom Sys., Inc. v. Houston Computer Servs., Inc., 918 F.2d 937, 942 (Fed. Cir. 1990), the identification of services in plaintiff's applications "controls this analysis," In re Dayan, 61 Fed.Appx. 695, 696 (Fed. Cir. 2003). Although plaintiff's applications reference a multitude of services, because "registration is properly refused if the word is the generic name of any of the goods or services for which registration is sought," 2 McCarthy on Trademarks § 12:57, the Court need not analyze each service. Instead, it will focus on the broadest services described in plaintiff's applications: "travel and tour ticket reservation services" for Class 39 and "[m]aking hotel reservations for others" for Class 43. Moskin Decl. [Dkt. No. 65-5] ¶¶ 2-4.
b. Relevant Purchasing Public
To define the relevant purchasing public, a court must look to the class of goods and services for which the trademark application was submitted. Cf. Glover, 74 F.3d at 59; Retail Servs., Inc. v. Freebies Publ'g, 247 F.Supp.2d 822, 826 (E.D. Va. 2003), aff'd Retail Servs., 364 F.3d at 535. Here, the applications sought registration for travel, tour, and hotel reservation services, including those offered online, making consumers who use travel, tour, and hotel reservation services offered via the internet or in person the relevant purchasing public.
c. Primary Public Understanding
The next consideration is whether "the primary significance of the term in the minds of the consuming public is not the product but the producer." Kellogg, 305 U.S. at 118, 59 S.Ct. 109. The public's primary understanding of a mark "is derived from it as a whole, not from its elements separated and considered in detail;" therefore, "it should be considered in its entirety." Estate of P.D. Beckwith, Inc. v. Comm'r of Patents, 252 U.S. 538, 545-46, 40 S.Ct. 414, 64 S.Ct. 705 (1920). Although "a mark must be considered as a whole," this "does not preclude courts from considering the meaning of individual words in determining the meaning of the entire mark." Hunt Masters, Inc. v. Landry's Seafood Rest., Inc., 240 F.3d 251, 254 (4th Cir. 2001). Accordingly, this Court will consider the two elements "booking" and ".com" separately before considering them in combination.
Defendants' argument that BOOKING.COM is generic rests primarily on its view of the meaning of "booking," for which its primary sources are various dictionary definitions. "Although not controlling, `dictionary definitions are relevant and sometimes persuasive' on the issue of genericness `based upon the assumption that dictionary definitions usually reflect the public's perception of a word's meaning and its contemporary usage.'" Retail Servs., 364 F.3d at 544-45 (citing 2 McCarthy on Trademarks at § 12:13). The definitions in the record, which are nearly identical to those relied on by the TTAB, include:
1. An engagement, as for a performance. 2. A reservation, as for accommodations at a hotel. American Heritage College Dictionary (3d ed. 1997).
1. a contract, engagement or scheduled performance of a professional entertainer. 2. a reservation. 3. the act of a person who books. Random House Unabridged Dictionary (2d ed. 1993).
1: the act of one that books 2: an engagement or scheduled performance...3: RESERVATION; esp one for transportation, entertainment, or lodging 4: ORDER. Webster's Third New International Dictionary (1993).
1: the act of one that books 2: an engagement or scheduled performance 3: RESERVATION. Merriam-Webster's Collegiate Dictionary (11th ed. 2008).
Plaintiff points out that the primary definition of "booking" in the definitions produced by the defendants refers to a performance, as in a theatrical engagement, and that the word has numerous other meanings. Pl. Mem. at 21. Although this observation is correct, it does not advance plaintiff's cause as "a word may have more than one generic use." Abercrombie & Fitch, 537 F.2d at 11.
Defendants also cite evidence that plaintiff and its competitors routinely use the word "booking" in reference to their services. For example, plaintiff's website uses "booking" as a noun, to describe a reservation, see A345 ("Sign in to manage your bookings."); id. ("Latest booking 10 minutes ago."), and as a verb, meaning to make a reservation, see A349 ("Our goal is to provide business and leisure travelers with the most accessible and cost effective way of discovering and booking the broadest section of accommodations in every corner of the world."), as do its confirmation notices, which refer to the reservation as a "booking" in the subject line, see Def. Ex. A, PTO-00011. Similarly, plaintiff's competitors use "booking" as both a noun and a verb in describing their services. For example, Hotwire, which provides services for making hotel, car, and flight reservations, advertises "easier booking," id. at PTO-00322, Hotels.com claims to be the preferred choice "when it comes to booking the perfect hotel," id. at PTO-000298, and Travelocity and Expedia offer services for "hotel booking," id. at PTO-00326; PTO-00313, "vacation package booking," PTO-00315, and "booking a rental car," PTO-00327. "Booking" is also a common component of descriptors for hotel reservation and travel agency services. A 2016 New York Times article regarding the impact of online reservation services on hotel loyalty programs referred to "Hotels.com, Hotwire.com, Trivago.com, and Travelocity.com" as "booking sites." Id. at PTO-00261-64; see also id. at PTO-00286-92. And, Skift, an information platform for the travel sector, headlined an article discussing flight reservation services with reference to "booking sites." Id. at PTO-00250-54. Finally, defendants identify fifteen third-party websites that include "booking.com" or "bookings.com" as components of their domain names. See id. at PTO-00148-65; A764-68, A772-81, A1085-86.
Plaintiff contends that this evidence is not enough to show that the word "booking" is "ever used as a generic term for travel agency services" and "has no relation whatsoever to plaintiff's business consulting services." Pl. Mem. at 21. This argument parallels plaintiff's argument regarding the proper genus of services, which effectively contends that the term "booking" is too narrow to describe the broad "travel agency services" offered by plaintiff while simultaneously too general to capture plaintiff's consulting services. But, this "heads I win, tails you lose" approach has no legal support. See Nat'l Nonwovens, Inc. v. Consumer Prod. Enters., 397 F.Supp.2d 245, 252 (D. Mass. 2005) (rebuffing plaintiff's "subtle rhetorical move that attempts to abstract [the genus] to a higher level of generality"). As explained above, it is well-established that "[a] registration is properly refused if the word is the generic name of any of the goods or services for which registration is sought." Cordua Rests., 823 F.3d at 605 (citing 2 McCarthy on Trademarks § 12:57). This makes good sense. Otherwise, applicants could elude a finding of genericness by simply tailoring their recitation of the goods and services at issue to be broader or narrower than the linguistic scope of their generic or descriptive mark. Permitting such gamesmanship would defeat one of the central purposes of the Lanham Act, which "is carefully crafted to prevent commercial monopolization of language that otherwise belongs in the public domain." S.F. Arts & Athletics, Inc. v. U.S. Olympic Comm., 483 U.S. 522, 573, 107 S.Ct. 2971, 97 L.Ed.2d 427 (1987) (Brennan, J., dissenting).
The evidence presented by the defendants establishes that, by itself, the word "booking" is generic for the classes of hotel and travel reservation services recited in plaintiff's applications. The dictionary definitions include as a meaning of "booking" a reservation or the act of making a reservation. Even more tellingly, plaintiff and its competitors use the term in this manner. This definition is also consistent with public usage, as indicated by the news sources quoted in the record. In this respect, the word "booking," standing alone, is the common descriptive name for both the act of making a reservation and the reservation itself. This conclusion is equally true for hotel reservations and the wider set of reservations offered by a travel agency service, because hotel, flight, and tour reservations are all referred to as "bookings," just as the act of making these types of reservations is often called "booking." Based on this evidence, the Court finds that the term "booking" is generic for these types of services.
ii. Top-Level Domains
The finding that "booking" is a generic term does not end the analysis because the mark at issue is BOOKING.COM. Therefore the Court must consider whether the term resulting from combining "booking" with ".com" remains generic. According to dictionary definitions, ".com" refers to a "commercial organization (in Internet addresses)," American Heritage College Dictionary (3d ed. 1997), or "[p]art of the internet address of many companies and organizations," Dictionary.com, http://www.dictionary.com/browse/-com (last accessed Apr. 7, 2017). In addition, some dictionaries state that "the phrase dot-com is used to refer generically to almost anything connected to business on the Internet." Id. Plaintiff argues that ".com" should be read as a top-level domain (TLD), in the same family as ".net," ".org," and ".edu." Pl. Mem. at 22. A TLD can be contrasted with a second-level domain (SLD), which is the next level of organization in the domain name hierarchy. For example, in "booking.com," "booking" is the SLD and ".com" is the TLD. According to plaintiff, the combination of "booking" and ".com" signals a domain name,3 which is a unique identifier capable of indicating the source of a product or service. Id. at 21, 23. The defendants, on the other hand, argue that ".com" is merely a term that denotes services offered via the Internet, and point to Federal Circuit cases holding that a TLD has no source identifying significance.
Although Federal Circuit case law on trademark is not controlling in this jurisdiction, it is persuasive authority. Because the parties acknowledge that there is no Fourth Circuit precedent regarding the source identifying significance of a TLD, Def. Mem. at 19 n.13, the reasoning of the Federal Circuit, which has addressed the role of TLDs in at least five cases, is a helpful starting point; however, it is important to appreciate that all of these opinions arose in § 1071(a) proceedings, in which the Federal Circuit reviewed the TTAB's decisions regarding genericness and descriptiveness for substantial evidence, which is a more deferential standard than the de novo review applicable in this civil action brought under § 1071(b).
The Federal Circuit first addressed the legal effect of combining a SLD consisting of a generic word (henceforth "generic SLD") and a TLD in In re Oppedahl & Larson LLP, 373 F.3d 1171 (Fed. Cir. 2004). There, the USPTO found that PATENTS.COM was generic for software that allowed consumers to track the status of U.S. trademark and patent applications. This holding relied on the conclusion that "patents" was generic and the Trademark Manual of Examining Procedure's instruction that "[b]ecause TLDs generally serve no source-indicating function, their addition to an otherwise unregistrable mark typically cannot render it registrable." Id. at 1174-75 (citing 1209.03(m) Domain Names [R-2]). Before the Federal Circuit, the applicant argued that domain name marks were inherently distinctive and therefore all such marks were entitled to registration. Id. at 1176. The Federal Circuit rejected this argument and affirmed the USPTO, reasoning that "[t]elephone numbers and street addresses are also unique, but they do not by themselves convey to the public the source of specific goods or services." Id. at 1176-77. Nevertheless, the Federal Circuit cautioned that "a bright-line rule that the addition of a TLD to an otherwise descriptive term will never under any circumstances affect the registratibility [sic] of a mark" would "be a legal error," concluding that the USPTO's policy was not a bright-line rule. Id. at 1175.
The Federal Circuit's next TLD case, In re Steelbuilding, 415 F.3d 1293 (2005), is the only case in which the Federal Circuit reversed the TTAB's finding that a domain name was generic, although it ultimately sustained the USPTO's denial of registration. In that case, the applicant sought registration of STEELBUILDING.COM for "computerized on-line retail services in the field of pre-engineered metal buildings and roofing systems." Id. at 1296. On appeal, the Federal Circuit determined that the evidence did not support the TTAB's genericness finding because "[t]he applicant's web site permits a customer to first design, then determine an appropriate price, for its own unique design," and that the TTAB "misunderst[ood] the proper genus." Id. at 1298. With respect to the TLD, the court concluded that "[i]n [that] unusual case, the addition of the TLD indicator expanded the meaning of the mark to include goods and services beyond the mere sale of steel buildings" by "expand[ing] the mark to include internet services that include `building' or designing steel structures on the web site and then calculating an appropriate price before ordering the unique structure." Id. at 1299. Although it rejected the genericness finding, the Federal Circuit affirmed the TTAB's alternative conclusion that the mark was descriptive for the online services specified in the application and that the applicant had failed to meet its burden of proving acquired distinctiveness. Id. at 1299-300. In a separate opinion that diverged from Oppedahl & Larson's conclusion that TLDs generally serve no source identifying function, Judge Linn argued that "[i]n the Internet world, domain-name recognition is a form of source identification" and argued that the case should be remanded to the TTAB for a reassessment of the evidence. Id. at 1301 (Linn, J., concurring-in-part and dissenting-in-part).
The Federal Circuit's next case4 involved the mark HOTELS.COM, which the examiner concluded was descriptive for the class of services — "providing information for others about temporary lodging; travel agency services, namely, making reservations and bookings for temporary lodging for others by means of telephone and global computer network" — and had not acquired secondary meaning. In re Hotels.com, L.P., 573 F.3d 1300, 1301 (Fed. Cir. 2009). The TTAB subsequently affirmed the rejection but on the alternative basis that HOTELS.COM is a generic term for hotel information and reservation services and that the addition of ".com" to "hotels" did not convert the generic term "hotels" into a protectable mark. Id. In reaching this conclusion, the TTAB relied on dictionary definitions of "hotel," computer printouts of the applicant's website featuring links to hotels, and the inclusion of the characters "hotel.com" in other domain names. Id. at 1301. The applicant presented rebuttal evidence, including sixty-four declarations from customers, vendors, and competitors, who each stated that "the term HOTELS.COM is not the common, generic name of any product, service, or field of study," as well as a Teflon survey drawn from 277 respondents in which 76% regarded HOTELS.COM as a brand name, both of which the TTAB declined to credit. Id. at 1304-05. On appeal, the Federal Circuit concluded that "on the entirety of the evidence before the TTAB, and with cognizance of the standard and burden of proof borne by the USPTO, the TTAB could reasonably have given controlling weight to the large number of similar usages of `hotels' with a dot-com suffix, as well as the common meaning and dictionary definition of `hotels' and the standard usage of `.com' to show a commercial internet domain" and held that "Board's finding that HOTELS.COM is generic was supported by substantial evidence." Id. at 1305-06.
The USPTO also denied registration to MATTRESS.COM for services identified as "online retail store service in the field of mattresses, beds, and bedding" on the basis of genericness. In re 1800Mattress.com IP, LLC, 586 F.3d 1359, 1361 (Fed. Cir. 2009). The Federal Circuit affirmed, rejecting the applicant's argument that because consumers did not refer to such stores as "mattresses.com's" the term could not be generic and instead holding that "substantial evidence ... support[ed] the Board's conclusion that `[c]onsumers would see MATTRESS.COM and would immediately recognize it as a term that denotes a commercial website rendering retail services featuring mattresses.'" Id. at 1362, 1364. In addition, the court found that "[b]ecause websites operate under the term `mattress.com' to provide mattresses, and they provide them online, the [TTAB] properly concluded that the relevant public understands the mark MATTRESS.COM to be no more than the sum of its constituent parts, viz., an online provider of mattresses." Id. at 1363.
Although the Court recognizes the persuasive force of Federal Circuit cases, a number of factors caution against crediting these precedents here. From a chronological perspective, the Federal Circuit's first TLD case, Oppehahl & Larson, which held that "TLDs generally serve no source-indicating function," was decided in 2004 when the internet was in its infancy and norms regarding domain names were just taking root. See 373 F.3d at 1176. Subsequent opinions have undermined Oppehahl & Larson's reasoning by recognizing that a TLD indicates a domain name and "domain-name recognition is a form of source identification." Steelbuilding, 415 F.3d at 1301 (Linn, J., concurring-in-part and dissenting-in-part). There also appears to be a tension between the Federal Circuit's statement that a per se rule that TLDs cannot be source identifying would be "legal error," Oppedahl & Larson, 373 F.3d at 1177, and the outcomes of these cases, which show that the USPTO's guidance on TLDs functions as a per se rule, see Oppedahl & Larson, 373 F.3d at 1177; Hotels.com, 573 F.3d at 1306; 1800Mattress.com, 586 F.3d at 1363. As discussed above, Steelbuilding is a notable exception, but as Professor McCarthy explains, this case "muddied the waters" and appears to be based on an "erroneous" characterization of STEELBUILDING.COM. 1 McCarthy on Trademarks § 7:17.50.
Beyond the tension within the cases, the Federal Circuit's TLD precedents also demonstrate the difficulty of distinguishing between generic and descriptive marks, an indeterminacy evidenced both by the anomalous holding in Steelbuilding and the multiple cases in which examining attorneys denied registration based on descriptiveness only to be affirmed by TTAB decisions concluding that the mark was actually generic, Oppedahl & Larson, 373 F.3d at 1173; Hotels.com, 573 F.3d at 1301. As discussed below, because "categorizing trademarks is necessarily an imperfect science," Fortune Dynamic, Inc. v. Victoria's Secret Stores Brand Mgmt., Inc., 618 F.3d 1025, 1033 (9th Cir. 2010), it would be imprudent to adopt a sweeping presumption denying trademark protection to a whole category of domain name marks in the absence of robust evidence that public ownership of this language is necessary for consumers and competitors to describe a class of products or services — evidence that does not appear in the Federal Circuit cases. Most importantly, in each of these TLD cases the Federal Circuit reviewed TTAB decisions under the deferential substantial evidence standard, a point that was repeatedly emphasized in the cases. See, e.g., Hotels.com, 573 F.3d at 1301. By contrast, under § 1071(b) this Court is required to conduct a de novo review. For all these reasons, this Court declines to rely on the Federal Circuit's precedents regarding TLDs and will treat this question as an issue of first impression. And, for the reasons developed below, the Court concludes that, when combined with an SLD, a TLD generally has source identifying significance and the combination of a generic SLD and a TLD is generally a descriptive mark that is protectable upon a showing of acquired distinctiveness.
To illustrate this conclusion, it is helpful to consider the Federal Circuit's reasoning in a case involving telephone numbers as marks. In 2001, before the Federal Circuit first confronted the issue of TLDs, it held that the mark 1-888-M-A-T-R-E-S-S was protectable as a descriptive mark. In re Dial-a-Mattress, 240 F.3d 1341, 1346 (Fed. Cir. 2001). In that case, the applicant applied to register 1-888-M-A-T-R-E-S-S was a service mark for "telephone shop-at-home retail services in the field of mattresses." Id. at 1343. The examining attorney rejected the mark as generic for the relevant services or, in the alternative, as a descriptive mark with insufficient evidence of acquired distinctiveness. Id. at 1344. The TTAB affirmed both rationales. Id. On appeal to the Federal Circuit, Dial-a-Mattress conceded that the area code in the mark was devoid of source identifying significance by itself and that the word "mattress," no matter how creatively spelled, was generic for retail services in the field of mattresses; however, it argued that, considered in its entirety, the mark was not generic. Id. at 1345. The Federal Circuit agreed, holding that although area codes have no source identifying significance by themselves and the term "mattress" was generic, the combination of an area code and a generic term (1-888-MATRESS) was source identifying. Id. at 1346. Specifically, it was descriptive, as it indicated that "a service relating to mattresses [was] available by calling the telephone number." Id. Yet, even though the telephone mnemonic was source identifying, the Federal Circuit explained that the applicant still needed to establish "acquired secondary meaning" (also termed "acquired distinctiveness") in order to register the descriptive mark. Id. at 1347.5
The reasoning in Dial-a-Mattress maps seamlessly onto TLDs. Although a TLD, like an area code, has no source identifying significance by itself, in combination with a SLD, it indicates a domain name, which, like a telephone number, is unique. Moreover, like the mnemonic phone number 1-888-M-A-T-R-E-S-S, the combination of a TLD and a generic SLD creates a descriptive mark by indicating that services relating to the generic SLD are available by accessing the domain name. Finally, whether such a mark is entitled to trademark protection depends on whether the applicant can demonstrate that it has acquired distinctiveness. In short, TLDs generally do have source identifying value when used in conjunction with an SLD and a mark comprised of a generic SLD and a TLD is generally a descriptive mark entitled to trademark protection if the mark holder can establish acquired distinctiveness.
Defendants resist this conclusion. Beyond invoking the Federal Circuit cases, defendants' argument that a TLD does not have identifying significance relies principally on the Supreme Court's 1888 decision in Goodyear's Rubber Mfg. Co. v. Goodyear Rubber Co., 128 U.S. 598, 602, 9 S.Ct. 166, 32 S.Ct. 535 (1888), which held that adding terms such as "Corp.," "Inc.," and "Co." to a generic term does not add any trademark significance to an otherwise unregistrable mark. Def. Mem. at 13, 21. By analogy, defendants argue that "[a]dding `.com' to a generic term does not create a composite that is capable of identifying source, just as Plaintiff would not have created a protectable mark by adopting the designation `Booking Company.'" Id. at 13. This analogy is unhelpful because Goodyear's reasoning regarding corporate designators does not apply with equal force to domain names. As the Supreme Court explained in Goodyear, the use of a corporate designation had no source identifying value because it "only indicates that the parties have formed an association or partnership to deal in [particular] goods, either to produce or to sell them." 128 U.S. at 602. 9 S.Ct. 166. By contrast, adding a TLD such as ".com" to a generic SLD does more than indicate that a company offers services via the internet; it indicates a unique domain name that can only be owned by one entity. In this respect, unlike a corporate designation, a TLD that functions as part of a domain name does have source identifying significance.
Defendants further argue that the public understands that a mark comprised of a generic SLD combined with a TLD is generic for that class of goods or services; however, they provide no evidence to support this position other than citations to the aforementioned Federal Circuit decisions. Def. Mem. at 13. As will be discussed below, defendants' evidence shows that the public understands that such a mark represents a unique domain name indicating to consumers that the proprietor of the domain name provides goods or services relating to the generic term. Blair Rep. at 14 (arguing there is a "tendency for [some survey] respondents to think that any DOT-COM name is a brand").
Next, citing Advertise.com, Inc. v. AOL Adver., Inc., 616 F.3d 974. 980 (9th Cir. 2010), defendants raise the policy argument that recognizing the source identifying significance of TLDs would create "a per se rule — in contravention of the Lanham Act — that the combination of `.com' with any generic term renders it protectable." Def. Mem. at 2, 17. In Advertise.com, the Ninth Circuit addressed AOL's motion for a preliminary injunction against the registrant of ADVERTISE.COM on the basis that the mark was confusingly similar to AOL's mark ADVERTISING.COM, ultimately holding that the combination of a generic SLD and a TLD was not eligible for protection. The Court understands defendants to be invoking Advertise.com's conclusion to argue that, if trademark law recognizes TLDs as source identifying, the addition of a TLD to a generic SLD would always result in a protectable mark. This argument overreaches. Acknowledging that combining a TLD with a generic SLD can produce a source identifying domain name is not tantamount to finding that all domain name marks are protectable. Rather, a generic SLD combined with a TLD creates a descriptive mark that is eligible for protection only upon a showing of acquired distinctiveness. Importantly, acquired distinctiveness is a much higher bar than uniqueness and requires an evidentiary showing that "in the minds of the public, the primary significance of a ... term is to identify the source of the product rather than the product itself." Inwood Labs., Inc. v. Ives Labs., 456 U.S. 844, 851 n.11, 102 S.Ct. 2182, 72 L.Ed.2d 606 (1982). In the trademark context, "source" does not refer to the location where a good or service may be found, e.g., at the website associated with a domain name, but to the "producer." Kellogg, 305 U.S. at 118, 59 S.Ct. 109. Therefore, domain name marks composed of a generic SLD and TLD, will be eligible for protection only when the applicant can show that "the primary significance" of the mark in the minds of the relevant consumers is the producer. Id. Such a showing is only possible where the owner of the mark has developed strong brand recognition.
The second policy concern raised by defendants is that granting trademark protections to domain names with a generic SLD would prevent competitors from using the generic term in their domain names, hampering their ability to communicate the nature of their services. Def. Opp. at 25. This argument again echoes the Ninth Circuit's analysis in Advertise.com, which reasoned that granting protection to such a mark would "grant the trademark holder rights over far more intellectual property than the domain name itself," permitting mark holders to monopolize a wide swath of domain names, including those comprised of the generic SLD at issue and any other TLD (e.g., "advertise.net"; "advertise.biz"; "advertise.org") as well other domain names that contain the generic word (e.g., "localadvertise.com"; "advertiseonline.com"). See 616 F.3d at 980-81. This argument, although initially alarming, does not withstand scrutiny.
The most obvious refutation of the monopolization concern is that domain names with a descriptive SLD, such as "steelbuilding," are already eligible for protection upon a showing of secondary meaning. Steelbuilding.com, 415 F.3d at 1299; Oppedahl & Larson, 373 F.3d at 1173. Moreover, although the USPTO has registered marks with what it determined are descriptive SLDs, such as WORKOUT.COM. ENTERTAINMENT.COM, and WEATHER.COM, this has not stopped competitors from using the words "workout," "entertainment," or "weather" in their domain names. To the contrary, such related domain names abound and many, such as MIRACLEWORK-OUT.COM. WWW.GOLIVE-ENTERTAINMENT.COM. and CAMPERSWEATHER.COM, have actually been afforded trademark protection by being registered on the Principal Register.6
In addition, the descriptive nature of domain name marks with a generic SLD will significantly limit the protection they receive, thereby safeguarding competition and public use. It is axiomatic that "descriptive terms qualify for registration as trademarks only after taking on secondary meaning...with the registrant getting an exclusive right not in the original, descriptive sense, but only in the secondary one associated with the markholder's goods." KP Permanent Make-Up, 543 U.S. at 122, 125 S.Ct. 542. Beyond the circumscribed protection afforded to descriptive marks, competitors are also protected by the likelihood of confusion standard. As the Supreme Court emphasized in KP Permanent Make-Up, the party charging infringement bears the burden of proving that a competitor's use of a mark is likely to confuse consumers. Id. at 118, 125 S.Ct. 542. This is a heavy burden for a plaintiff because likelihood of confusion rests on nine factors, which include the source identifying strength of the plaintiff's mark, the degree of similarity between the marks, and the defendant's intent.7 See, e.g., H. Jay Spiegel & Assocs., P.C. v. Spiegel, 652 F.Supp.2d 639, 650 (E.D. Va. 2009), aff'd, 400 Fed.Appx. 757 (4th Cir. 2010) (finding, on a motion for summary judgment, that the record was inconclusive as to whether SPIEGELLAW.COM was confusingly similar to SPIEGELAW.COM). Likelihood of confusion is particularly difficult to prove for descriptive marks because they are considered "weak" marks, see Shakespeare Co., 110 F.3d at 239-40, and "when the common element between two marks is a word which is `weak,' the likelihood of confusion between the marks is reduced." Pizzeria Uno Corp. v. Temple, 566 F.Supp. 385, 396 (D.S.C. 1983), aff'd, 747 F.2d 1522 (4th Cir. 1984). Moreover, even if the party charging infringement succeeds in establishing likelihood of confusion, the accused party can defend itself by demonstrating fair use.8 See KP Permanent Make-Up, 543 U.S. at 121, 125 S.Ct. 542 ("some possibility of consumer confusion must be compatible with fair use"). As the Supreme Court has explained, "[i]f any confusion results" from descriptive fair use "that is a risk the plaintiff accepted when it decided to identify its product with a mark that uses a well known descriptive phrase." Id. (internal citation omitted). This principle is equally true in the context of domain names and will preclude holders of marks comprised of a generic SLD and a TLD from preventing competitors from using the generic term in other domain names.9
Defendants' third policy concern, which again proves more imagined than real, is that granting trademark protection to domain names with generic SLDs would deprive competitors of the right to describe their goods and services as what they are. Def. Mem. at 11-12. As defendants elaborate, "Imagine being forbidden to describe a Chevrolet as a `car' or an `automobile' because Ford or Chrysler or Volvo had trademarked these generic words." Id. at 12 (citing Retail Servs., 364 F.3d at 538). Defendants appear to suggest that plaintiff's competitors need to be able to describe themselves as "booking.coms." Although concerns about monopoly are one of the animating forces behind the prohibition on registering generic marks, because each domain name is unique the Court is unpersuaded that the threat of monopoly applies with equal force to domain names.10 Further, the monopoly argument appears to assume that certain terms must be left in the public commons because they have descriptive value and are needed by consumers and competitors alike; however, no evidence in this record supports the view that domain names are used as descriptive terms for classes of services. To the contrary, the record is replete with evidence that consumers are predisposed to think that a domain name refers to a particular entity. Blair Rep. at 14; In re Hotels.com, 87 U.S.P.Q.2d 1100, 1109 (T.T.A.B. Mar. 24, 2008) (according to the TTAB, "consumers may automatically equate a domain name with a brand name"). By this same logic, plaintiff's competitors, such as Expedia and Travelocity, have no incentive to describe themselves as "booking.coms" because this risks diverting customers to the website of their competitor. In short, there is no evidence in this record indicating that permitting registration of a domain names with a generic SLD would result in the monopolization of descriptive terms that must be left free for public use.
Recognizing that the policy concerns regarding generic terms are a poor fit for marks comprised of a generic SLD and a TLD, the next question is whether the dual purposes of the Lanham Act — protecting consumers and incentivizing brand development — militate for or against protection. Generally, the consumer protection rationale favors trademark protection because brands minimize the information costs of purchasing decisions. Qualitex, 514 U.S. at 164, 115 S.Ct. 1300. Although trademark rights are disfavored when they would cause consumer confusion or impede competition, Am. Online, 243 F.3d at 821, because domain names are inherently unique and the scope of protection afforded to a domain name with a generic SLD will be narrow, the risk of consumer confusion or anticompetitive monopolies is remote. Rather, the evidence in this record shows that consumers are primed to perceive a domain name as a brand which militates for, not against, trademark protection for domain names. In addition, because online goods and services are a significant and ever-growing part of the economy, granting trademarks to producers who primarily offer goods and services online and brand themselves based on their domain name favors the interest of consumers by limiting the prospect of deception and confusion. Incidentally, this also protects the good will generated by producers, often at great effort and expense, and thereby incentivizes brand development.11 In sum, the rationales animating the Lanham Act are aligned with the conclusions that TLDs are generally source identifying and that a mark composed of a generic SLD and a TLD is a descriptive mark eligible for protection upon a showing of acquired distinctiveness.
iii. Evidence of Public Understanding Regarding Genericness
The Court now considers evidence of the public's understanding of BOOKING.COM, which may include "purchaser testimony, customer surveys, dictionary listings, newspapers, and other publications." Retail Servs., 247 F.Supp.2d at 826. The most striking feature of the evidence in this record is the absence of evidence that consumers or producers use the term "booking.com" to describe the genera of services at issue, that is, hotel and travel reservation services. Pl. Mem. at 12. Defendants point to no purchaser testimony, consumer surveys, newspaper articles, websites, or other publications demonstrating that either the consuming public or plaintiff's competitors refer to travel and hotel reservation services offered online as "booking.coms." See Dial-a-Mattress, 240 F.3d at 1346. Instead, they rely heavily on a statement from the Federal Circuit that use "is irrelevant" and "the correct inquiry is whether the relevant public would understand, when hearing the term `mattress.com,' that it refers to online mattress stores." Def. Mem. at 13 (citing 1800Mattress.com, 586 F.3d at 1364 (emphasis added)); see also Def. Opp. at 7, 18; Def. Reply at 13. In reaching this conclusion, the Federal Circuit cited H. Marvin Ginn v. Int'l Ass'n of Fire Chiefs, Inc.'s genericness test, which asks whether the "relevant public primarily use or understand the term sought to be protected to refer to the genus of goods or services in question." 782 F.2d 987, 989-90 (Fed. Cir. 1986) (emphasis added). But 1800Mattress.com is not controlling authority, and the Fourth Circuit has not adopted H. Marvin Ginn's test for genericness. In addition, even under this test, whether a mark is used to refer to a genus is certainly relevant; it simply is not dispositive. Importantly, in this de novo proceeding, the evidence before the Court indicates that "[l]inguistic understanding is not some further mental condition"; rather, in the words of Ludwig Wittgenstein, "meaning is use." Leslie Rep., Pl. Ex. 2 [Dkt. No. 64-2] ¶ 78 (citing Philosophical Investigations § 43 (1953)).12 Accordingly, the absence of evidence indicating that the consuming public uses the term BOOKING.COM to refer to a class of services, is highly relevant.
What evidence defendants have produced shows that the types of services offered by plaintiff are routinely referred to as "booking website(s)," "booking site(s)," etc. Def. Mem. at 21. According to defendants, "these same meanings are immediately conveyed by the term `booking.com,'" id. at 22; however, they offer no support for the argument that any composite term that communicates the same meaning as a generic term is itself generic. Id. Further, because domain names are unique, the Court is unpersuaded that BOOKING.COM has the same meaning as "a booking website" or "booking websites," both of which could refer to any number of websites. Cf. Dial-a-Mattress, 240 F.3d at 1346 ("[A] phone number is not literally a genus or a class name.").
Defendants also point to fifteen third-party websites that include "booking.com" or "bookings.com." Def. Mem. at 23; PTO-0148-65 (identifying examples such as "dreamvacationbooking.com"; "vacationhomebooking.com"; and "bhutanbookings.com"). This evidence is unpersuasive because including the characters "b-o-o-k-i-n-g-.-c-o-m" in a longer domain name is not the equivalent of describing one's service as a "booking.com." Indeed, a brief review of these websites indicates that they do not describe themselves as such.13 Moreover, accepting the defendants' logic would result in privileging trademark protection for long SLDs over short ones simply because a shorter domain name's set of characters is likely to be included in longer domain names. This is not the USPTO's practice, as evidenced by its registration of marks like WORKOUT.COM, ENTERTAINMENT.COM, and WEATHER.COM notwithstanding the multitude of other domain names that uses these strings of characters. Therefore, the Court finds that defendants' list of domain names does not constitute evidence that BOOKING.COM is used to refer to a genus of services.14
Conversely, plaintiffs have adduced persuasive evidence that the consuming public understands BOOKING.COM to be a specific brand, not a generic name for online booking services. In particular, plaintiff produced a Teflon survey which revealed that 74.8 percent of respondents identified BOOKING.COM as a brand name. Poret Rep. at 29. Teflon surveys, which are the "most widely used survey format to resolve a genericness challenge," 2 McCarthy on Trademarks at § 12:16, provide survey respondents with a primer on the distinction between the generic or common names and trademark or brand names, and then present respondents with a series of names, which they are asked to identify as common or brand names. Plaintiff's survey was conducted by Hal Poret, a statistician with experience administering over 200 consumer surveys regarding trademarks. The survey was administered online to 400 respondents from March to April 2016. Poret Rep., Pl. Ex. 1 [Dkt. No. 64-1] at 8, 18, 25.
The survey began by explaining the distinction between "brand names" and "common names" and provided consumers with examples of three brand names (TOYOTA, CHASE, and STAPLES.COM) and three common names (AUTOMOBILE, BANK, and OFFICESUPPLIES.COM). Id. at 9. The survey then tested consumer's understanding of the distinction between common and brand names by asking them to identify whether KELLOGG and CEREAL were common or brand names. Id. at 10. Respondents who correctly answered that KELLOGG is a brand name and CEREAL is a common name continued with the survey while those who did not were excluded. Id. at 11. Following that initial screening, respondents were shown a series of terms, one at a time, and for each term were asked to answer the following question:
• "Do you think this is a ...
• Brand name
• Common name
• Don't know"
Id. The list of terms and product descriptions shown to respondents were
• The term at issue:
• "BOOKING.COM (Hotel and other lodging reservation services)"
• Three brand names:
• "ETRADE.COM (Stock and investor broker services)"
• "PEPSI (Cola and other soft drinks)"
• "SHUTTERFLY (Photo-sharing and photo gift services)"
• Three common names
• "SPORTING GOODS (Products used in sports and other physical activity)"
• "WASHINGMACHINE.COM (Review and sales of washing machines)"
• "SUPERMARKET (Retail sale of food and other groceries)"
Id. at 11-13. There were four separate rotations in which these terms were presented to the respondents, in each of which the terms were ordered differently and with BOOKING.COM placed in a different position on the rotation "so as not to bias the responses to the term BOOKING.COM." Id. at 13-14. There were also two versions of the survey language, one in which the phase "brand name" always preceded the phrase "common name" (i.e., "This survey is about brand names and common names.") and one in which this order was reversed. Id. at 9, 15. The following table displays the proportion of respondents who identified each trademark as a brand name versus a common name, compared to BOOKING.COM:
BOOKING.COM PEPSI ETRADE.COM SHUTTERFLY
Brand name 74.8% 99.3% 96.8% 96.8%
Common name 23.8% 0.8% 3.0% 3.0%
Don't know 1.5% 0.0% 0.3% 0.3%
Id. at 28. The following table displays the proportion of respondents who identified each generic term as a brand name versus a common name, compared to BOOKING.COM:
BOOKING.COM SUPER-MARKET SPORTING WASHINGMACHINE.
Brand name 74.8% 0.0% 0.5% 33.0%
Common name 23.8% 100.0% 99.5% 60.8%
Don't know 1.5% 0.0% 0.0% 6.3%
Id. Poret concluded that in his opinion "these results strongly establish that BOOKING.COM is not perceived by consumers to be a generic or common name." Id. at 29; see also E. I. DuPont de Nemours & Co. v. Yoshida Int'l, Inc., 393 F.Supp. 502, 527 (E.D.N.Y. 1975) (finding that survey results indicating that 68% of consumers viewed Teflon as a brand name rebutted the claim that the mark was generic).
Defendants argue that plaintiff's Teflon survey should be excluded pursuant to Hunt Masters, 240 F.3d at 255, where the Fourth Circuit held that in a genericness inquiry consumer surveys are not relevant "where a term was commonly used prior to its association with the products at issue" whereas surveys are relevant where the term at issue "began life as a `coined term.'" Def. Opp. at 12. Although "booking" is not a coined term, BOOKING.COM arguably is. More importantly, defendants have presented no evidence that BOOKING.COM is in the category of marks for which Hunt Masters forecloses reliance on consumer surveys, namely that it "was commonly used prior to its association with the products at issue." Id. Moreover, the Court finds that because domain names marks are relatively new to trademark law, public understanding is highly relevant to understanding how these marks are perceived. As a result, this case is not on all fours with Hunt Masters and plaintiff's Teflon survey, which sheds light on how the composite mark BOOKING.COM is understood by consumers, is highly relevant.15
Defendants next attack the methodological soundness of plaintiff's survey as a basis for excluding the report. Def. Opp. at 13. Defendants rely primarily on a comment by Poret, posted in a blog by a colleague in 2009, in which he remarked that because consumers often assume that domain names have source identifying significance, surveys testing TLD marks should be composed exclusively or primarily of TLD marks. Def. Ex. A at PTO-00366. During his deposition, Poret explained that he has since revised his views. Recognizing that consumers conventionally encounter an array of marks, including TLD and non-TLD marks, Poret is now of the opinion that it is unnatural to test brand recognition with only TLD marks. Poret Tr., Pl. Ex. A [Dkt. No. 72-1] at 18:8-19:21. Defendants make much of this change of opinion but point to no case law, scholarly authority, or principled justifications for conducting Teflon surveys comprised exclusively or even primarily of TLD marks. Def. Mem. at 26-27; Def. Opp. at 14-15.
Defendants also point to three alleged methodological flaws in Poret's survey identified by their expert, Dr. Edward Blair. Blair Rep., Def. Ex. B [Dkt. No. 61-3]. First, Dr. Blair contends that the survey population is under-inclusive because it was restricted to consumers who search for or make hotel or travel reservations online but plaintiff's trademark applications also reference services offered in person. Id. at 11-13. Poret's supplemental expert response explains that the survey focused on consumers who used online reservation services because the USPTO determined that BOOKING.COM was generic for a website service, thus "measuring consumer perception of BOOKING.COM in the online context squarely tested the context in which the chance of consumers understanding a mark ending in `.COM' to be generic was greatest." Poret Supp. Rep., Def. Ex. D [Dkt. No. 61-5] at 3.
Second, although Dr. Blair concedes that the survey explained and tested the distinction between dot-com brand names and common names, he contends that this educational component was insufficient because it did not focus specifically on dot-com names and respondents were not tested on their ability to distinguish between dot-com brand names and dot-com common names. Blair Rep. at 5-6, 14. Observing that 33% of respondents incorrectly identified WASHINGMACHINE.COM, one of the test terms, as a brand name, Dr. Blair contends that the educational portion of the survey was ineffective and respondents were predisposed to think that any dot-com name was a brand name. Id. Without conceding that this is a flaw in the survey design, Poret explains that one can control for this predisposition by removing the respondents who answered that WASHINGMACHINE.COM is a brand name. Poret Supp. Rep. at 4. Even with that adjustment, of the remaining respondents, 65 percent identified BOOKING.COM as a brand name. Id. at 5.
Dr. Blair's third critique is that the responses varied based on the order in which the marks were presented, which he posits is an indication that respondents did not understand the distinction between dot-com brand names and common names and were answering based on context rather than actual knowledge. Blair Rep. at 19. Poret acknowledges that the results exhibit order effects but explains that all Teflon surveys have order effects, irrespective of whether they test dot-com or other marks. Poret Supp. Rep. at 8. "[T]he very reason that [Teflon] surveys include various orderings of the terms," he elaborates, "is because it is well known and expected that responses to terms often vary in this manner" and "providing various orderings is designed to control for this phenomenon." Id. (emphasis in original).
The Court is persuaded by Poret's responses and finds that Dr. Blair's critiques do not undermine the veracity of the survey results. "[N]o survey is perfect." Selchow & Righter Co. v. Decipher, Inc., 598 F.Supp. 1489, 1502 (E.D. Va. 1984). Poret's decision to limit the survey to online consumers was reasonable, the method used to instruct them on the distinction between generic and brand names was sufficient, and the ordering effects are, as Poret explains, both expected and appropriately controlled for by the survey design. In short, plaintiff's Teflon survey is reliable. It also provides the only actual evidence of consumers' understanding of BOOKING.COM, because defendants "had an equal opportunity to conduct [their] own survey but chose not to" do so. Selchow & Righter, 598 F.Supp. at 1503.
Numerous courts agree that "direct consumer evidence, e.g., consumer surveys and testimony is preferable to indirect forms of evidence" such as dictionaries, trade journals, and other publications. See, e.g., Berner Int'l Corp. v. Mars Sales Co., 987 F.2d 975, 982 (3d Cir. 1993). Even the Federal Circuit, the source of authority upon which the USPTO principally relies, has held that "consumer surveys may be a preferred method of proving genericness." BellSouth Corp. v. DataNational Corp., 60 F.3d 1565, 1570 (Fed. Cir. 1995). Therefore, the Court declines defendants' invitation to rely on theoretical and indirect sources of consumer understanding, such as dictionary definitions, over plaintiff's Teflon survey.
In sum, defendants have not met their burden of proving by clear evidence that BOOKING.COM is generic. To the contrary, the Court finds that the relevant consuming public primarily understands that BOOKING.COM does not refer to a genus, rather it is descriptive of services involving "booking" available at that domain name. Dial-a-Mattress, 240 F.3d at 1346 (finding that 1-8-8-8-M-A-T-R-E-S-S "immediately conveys the impression that a service relating to mattresses is available by calling the telephone number"). And, because "booking" is a broad enough term to refer to both hotel and travel reservation services, the Court finds that BOOKING.COM is descriptive of both the Class 39 and Class 43 services described in plaintiff's applications.
3. Acquired Distinctiveness
As with any descriptive mark, BOOKING.COM is eligible for protection only upon a showing of secondary meaning or acquired distinctiveness. See Steelbuilding.com, 415 F.3d at 1299. To make this showing, the burden shifts to the applicant to demonstrate 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." Inwood Labs., 456 U.S. at 851 n.11, 102 S.Ct. 2182. "Saying that a trademark has acquired secondary meaning is shorthand for saying that a descriptive mark has become sufficiently distinctive to establish a mental association in buyers' minds between the alleged mark and a single source of the product." Retail Servs., 364 F.3d at 539 (citation and internal quotation marks omitted); see also Sara Lee, 81 F.3d at 464 (noting that "secondary meaning" exists when, "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" (internal quotation marks omitted)); Perini Corp. v. Perini Constr., Inc., 915 F.2d 121, 125 (4th Cir. 1990) ("Secondary meaning is the consuming public's understanding that the mark, when used in context, refers, not to what the descriptive word ordinarily describes, but to the particular business that the mark is meant to identify.").
Proof of secondary meaning requires a "rigorous evidentiary" showing and courts consider six factors: "(1) advertising expenditures; (2) consumer studies linking the mark to a source; (3) sales success; (4) unsolicited media coverage of the product; (5) attempts to plagiarize the mark; and (6) the length and exclusivity of the mark's use." Perini, 915 F.2d at 125. Secondary meaning exists if a "substantial portion" of the relevant consuming public associates the term with the particular business, id., and the applicant bears the burden of proof, U.S. Search, LLC v. U.S. Search.com, Inc., 300 F.3d 517, 525-26 (4th Cir. 2002).
As explained above, applications containing multiple classes are treated as separate applications, 3 McCarthy on Trademarks § 19:56.50; therefore, the Court's analysis of the evidence of secondary meaning must assess Class 39 and 43 independently. But, in the interest of efficiency, the Court will begin by summarizing the evidence in the record.
a. Evidence of Acquired Distinctiveness
With respect to advertising, plaintiff has submitted evidence of the number of visual impressions of BOOKING.COM by consumers. Although the secondary meaning test refers to "advertising expenditures," Perini, 915 F.2d at 125, the Court is satisfied that the number of visual impressions is equally, if not more, probative of secondary meaning because it more closely approximates the number of consumers who have been exposed to a brand. Plaintiff aired BOOKING.COM branded television commercials that received 1.3 billion visual impressions from U.S. customers in 2015 and 1.1 billion impressions in 2016. Pl. Ex. A, Dunlap Decl. ¶ 9a. Its internet advertisements during these years received 212 million and 1.34 billion visual impressions from U.S. customers, respectively. Id. ¶ 9c. And its 2015 movie theater advertisements received approximately 40 million visual impressions from U.S. customers. Id. ¶ 9b. This is compelling evidence of plaintiff's advertising efforts and is considerably more wide-reaching than the evidence used to satisfy the first factor in other Fourth Circuit cases. See Resorts of Pinehurst, Inc. v. Pinehurst Nat'l Corp., 148 F.3d 417, 421-22 (4th Cir. 1998) (finding that the markholder had made "considerable advertising efforts and expenditure of money toward developing a reputation and goodwill" for its mark through a "nationwide marketing campaign" that involved "placing advertisements in numerous national golf publications such as Golf and Golf Digest magazines" and aggressively "seeking out major professional golf tournaments").16
As to the second factor, plaintiff cites two surveys. The first is a 2012 JD Power & Associates survey recognizing plaintiff as having the highest customer satisfaction rate of any travel site in the United States. Pl. Mem. ¶ 27; Pl. Opp. at 14. Defendants argue that the survey is entitled to little weight because it does not reveal anything about what consumers understand BOOKING.COM to mean. Def. Opp. at 26. Plaintiff contends that the survey is an admissible form of evidence, Pl. Opp. at 14, but admissibility is not the problem. Surveys such as the JD Power & Associates survey are designed to gauge the relative popularity of a product not the source identifying effect of the mark. As a result, defendants are correct in arguing that the JD Power & Associates survey is not probative of secondary meaning and is entitled to minimal weight. Chase Fed. Sav. & Loan Ass'n v. Chase Manhattan Fin. Servs. Inc., 681 F.Supp. 771, 780-81 (S.D. Fla. 1987) (explaining that survey evidence is entitled to "slight weight" when it "derive[s] from questions that were not asked in a specific, limited and probative context,... and did not probe the primary significance of the term" (internal quotation marks omitted)).
Plaintiff also relies on the Teflon survey conducted by Poret, which indicated that 74.8 percent of consumers of online travel services recognize BOOKING.COM as a brand. Pl. Mem. ¶ 4; id. at 29. Although primarily used to determine whether a mark is generic, Teflon surveys are also a generally accepted way of measuring secondary meaning. See, e.g., Firefly Digital, Inc. v. Google, Inc., No. CIV.A. 6:10-0133, 2011 WL 6160222, at *5 (W.D. La. July 7, 2011); Schwan's IP, LLC v. Kraft Pizza Co., 379 F.Supp.2d 1016, 1024 (D. Minn. 2005), aff'd, 460 F.3d 971 (8th Cir. 2006); March Madness Athletic Ass'n, L.L.C. v. Netfire, Inc., 310 F.Supp.2d 786, 809 (N.D. Tex. 2003), aff'd, 120 Fed.Appx. 540, 2005 WL 147264 (5th Cir. 2005); see also Innovation Ventures, LLC v. NVE, Inc., 90 F.Supp.3d 703, 720 (E.D. Mich. 2015) (finding a Teflon survey probative of the strength of a protectable mark). In this Circuit, "survey evidence is generally thought to be the most direct and persuasive way of establishing secondary meaning." U.S. Search, LLC, 300 F.3d at 526 n.13. And Professor McCarthy has identified survey evidence as one of a handful of types of direct evidence of consumer understanding — along with consumer testimony — as compared to the other secondary meaning factors, which offer circumstantial evidence of brand recognition. 2 McCarthy on Trademarks § 15:30. Because plaintiff's Teflon survey is the only evidence in the record that speaks directly to how consumers understand plaintiff's mark, it weighs heavily in the secondary meaning analysis and the survey's finding that approximately three out of four consumers recognize BOOKING.COM as a brand indicates strong brand awareness. See IDV N. Am., Inc. v. S & M Brands, Inc., 26 F.Supp.2d 815, 823 (E.D. Va. 1998) (holding that BAILEYS liqueurs had secondary meaning, based in part upon a "51% consumer awareness rating").
The third factor, record of sales success, is also well-established. Plaintiff's public filings reflect that its U.S. customers conduct billions of dollars' worth of transactions each year, Pl. Mem. ¶ 13,17 and, as of 2013, plaintiff's total transaction value, both in the United States and abroad, was over $8 billion, see Prakke Decl., A2522, which is substantially higher than the sales success in other cases where courts in this Circuit have found secondary meaning. See, e.g., Worsham Sprinkler Co. v. Wes Worsham Fire Prot., LLC, 419 F.Supp.2d 861, 869-72 (E.D. Va. 2006) (citing annual revenues averaging $18-20 million); IDV N. Am., Inc., 26 F.Supp.2d at 823 (finding that BAILEYS liqueurs' $1 billion in sales over the course of a decade supported a finding of secondary meaning). In addition, plaintiff's mobile app, which can be used to search for hotels and make reservations, was downloaded approximately 1 million times in 2014, 1.9 million times in 2015, and 2.5 million times in 2016, Pl. Ex. A, Dunlap Decl. ¶ 7, which offers additional, circumstantial evidence of sales success and consumer brand recognition.
The fourth factor is unsolicited media coverage. In 2015 and 2016 the number of news articles published in the United States referencing BOOKING.COM was over 600 and 650, respectively. Pl. Ex. A, Dunlap Decl. ¶ 10. This compares very favorably to other cases where courts have found that media coverage demonstrated that a brand had achieved public prominence. Washington Speakers Bureau, Inc. v. Leading Authorities, Inc., 33 F.Supp.2d 488, 496-97 (E.D. Va. 1999), aff'd, 217 F.3d 843 (4th Cir. 2000) (relying on evidence that a "news database search offered by [the markholder] disclosed hundreds of articles specifically referring to [the service] and its activities").
Plaintiff identifies no evidence of the fifth factor, attempts to plagiarize the mark. Pl. Mem. at 29. But, a party need not prove all six factors and the Fourth Circuit has concluded that secondary meaning can exist even when "no attempts to plagiarize the mark were shown." Perini, 915 F.2d at 126.
With respect to the sixth factor, length and exclusivity of use, plaintiff, which has been offering "online hotel reservation service" since 1996, operated from "1996 to June 2006 using the mark BOOKINGS. In June 2006, [plaintiff] modified its mark to BOOKING.COM and has been providing services under that mark since then." Prakke Decl. A2522. Eleven years of uninterrupted use, in conjunction with the other factors, weighs in favor of secondary meaning. See Teaching Co. Ltd. P'ship v. Unapix Entm't, Inc., 87 F.Supp.2d 567, 579-80 (E.D. Va. 2000) (finding that secondary meaning existed in a mark that had been used without interruption for eight years).
In addition, there is no evidence in the record that "any other party offering travel agency services refers to itself as a `Booking.com.'" Pl. Ex. C, Moskin Decl. ¶ 12. As previously discussed in the genericness evaluation, defendants point to fifteen third-party websites that include "booking.com" or "bookings.com," Def. Mem. at 23, and one might argue that this is evidence that plaintiff has not enjoyed exclusive use. This argument fails because the mere existence of a registered domain name or even a website does not equate to its use as a "mark." "[A] domain name does not become a trademark or service mark unless it is also used to identify and distinguish the source of goods or services." 1 McCarthy on Trademarks § 7:17.50. Out of the millions of domain names, only a fraction play the role of a mark. Id. Indeed, the websites associated with the domain names cited by the defendants identify their services not by reference to their domain name but by phrases such as "Dream Vacation Booking" and "Vacation Home Booking." Further, as explained above, these websites are not actually referring to themselves as "booking.coms," therefore they are not using the term either descriptively or as a mark.
Finally, plaintiff has adduced evidence of its substantial social media following. As of 2016, over 5 million consumers had "liked" BOOKING.COM on Facebook and over 100,000 "followed" BOOKING.COM on Twitter. Pl. Ex. A, Dunlap Decl. ¶ 12c.18 Although this evidence does not directly relate to any of the Perini factors, those factors are non-exhaustive, Shammas v. Rea, 978 F.Supp.2d 599, 612 (E.D. Va. 2013), and, just as unsolicited media coverage offers circumstantial evidence of consumer awareness of a brand, the size of a producer's social media following is indicative of the number of consumers who are familiar with a brand, interested in receiving additional information about it, and presumably tend to feel goodwill toward the producer.
In the face of this evidence, defendants argue that "although [p]laintiff has provided documents related to its commercial success, they do not demonstrate actual market recognition of `booking.com' as a source indicator." Def. Mem. at 30. This argument ignores the direct evidence of consumer understanding established by plaintiff's Teflon survey and appears to challenge the very nature of the secondary meaning test, which acknowledges that five of the six factors — advertising expenditures, sales success, media coverage, attempts to plagiarize, and exclusivity of use — are all circumstantial evidence. 2 McCarthy on Trademarks § 15:30. Professor McCarthy acknowledges direct evidence "is not a requirement and secondary meaning can be, and most often is, proven by circumstantial evidence." Id. In addition, it defies logic to suggest that billions of consumer impressions through advertising, billions of dollars in sales, and over 1,000 newspaper articles have no bearing on whether consumers understand BOOKING.COM to be a source of reservation services.19
b. Class Specific Analysis
Having summarized the evidence of secondary meaning, the next step is to consider what this evidence means for the two classes of marks set forth in plaintiff's applications. Unfortunately, the evidence does not clearly differentiate between Class 39 — travel agency services — and Class 43 — hotel reservation services. Plaintiff's evidence often speaks simply of BOOKING.COM, and, where it does differentiate, it refers only to plaintiff's hotel reservation services. For example, the Dunlap Declaration, which is the source of plaintiff's evidence regarding advertising, sales success, and unsolicited media coverage, describes plaintiff as "the worldwide leader in online accommodation reservation services" and reports that plaintiff enables customers to make reservations at over "1,027,450 hotels and accommodation providers throughout the world." Pl. Ex. A, Dunlap Decl. ¶¶ 3, 5. Other than referencing Booking.com's receipt of the "World's Leading Online Travel Agency Website" award in 2014 and 2015, Dunlap's declaration makes no reference to travel agency services. In addition, the Prakke Declaration, which establishes the length and exclusivity of use, also portrays Booking.com as an "online hotel reservation service," explaining that since 1996 plaintiff has been "providing hotels and consumers alike with an online hotel reservation service through which hotels all over the world can advertise their rooms for reservation and through which consumers all over the world can make reservations." A2522. Likewise, plaintiff's Teflon survey characterized Booking.com as providing "[h]otel and other lodging reservation services." Pl. Ex. 1, [Dkt. No. 64-1] at B-000055. In light of the total absence of evidence that either the consuming public, or even Booking.com's officers, associate BOOKING.COM with travel agency services, plaintiff has failed to carry its burden of establishing secondary meaning as to Class 39.20
Conversely, the record demonstrates strong evidence of secondary meaning for Class 43 on five of the six secondary meaning factors: Plaintiff has established the existence of an extensive nationwide advertising campaign; a strong public perception that BOOKING.COM is a brand identifier, as evidenced by the Teflon survey; robust consumer sales; voluminous unsolicited media coverage; and a decade of exclusive use. This evidence is more than sufficient to demonstrate that "in the minds of the public, the primary significance of" BOOKING.COM "is to identify the source of the product rather than the product itself," Sara Lee, 81 F.3d at 464, and that plaintiff's mark is entitled to protection for the services identified in Class 43, as a descriptive mark.21
The question of whether a TLD has source identifying significance is a question of first impression in this Circuit. After carefully reviewing the Federal Circuit's precedent on this issue, the purposes of the Lanham Act, and the competition-protecting features built into the structure of trademark law, the Court has concluded both that a TLD generally has source identifying significance and that a mark composed of a generic SLD and a TLD is usually a descriptive mark eligible for protection upon a showing of secondary meaning. Applying these holdings to the facts of this case, the Court holds that BOOKING.COM is a descriptive mark and that plaintiff has carried its burden of demonstrating the mark's secondary meaning as to the hotel reservation services described in Class 43 but not as to the travel agency services recited in Class 39.
For these reasons, in an order to be issued with this Memorandum Opinion, plaintiff's Motion for Summary Judgment will be granted in part and denied in part, defendants' Motion for Summary Judgment will be granted in part and denied in part, the USPTO will be ordered to register BOOKING.COM as to the Class 43 services in the '998 Application and '097 Application,22 and the Court will remand applications '365 and '366 to the USPTO for further administrative proceedings consistent with the findings and conclusions of this Memorandum Opinion to determine whether the design and color elements in those two applications, in combination with the protectable word mark, are eligible for protection as to Class 43 services.23