Boston Duck Tours, LP v. Super Duck Tours, LLC

In this trademark infringement dispute, defendant-appellant Super Duck Tours (“Super Duck”) appeals a preliminary injunction granted in favor of plaintiff-appel-lee Boston Duck Tours (“Boston Duck”). Asserting that the phrase “duck tour” and the image of a duck splashing in water are generic, and therefore that there is no likelihood of confusion between Boston Duck’s marks and Super Duck’s marks, appellant requests that we dissolve the injunction requiring the alteration of its trade name and logo. Appellee has also cross-appealed the order of the district court, arguing that the court committed clear error by not enjoining Super Duck from using the term “duck” in its trade name.

After carefully reviewing the record in this dispute, we conclude that the district court committed clear error by finding the phrase “duck tour” nongeneric, and thereby according it too much weight in its likelihood of confusion analysis. Conse*8quently, the court erred in issuing a preliminary injunction, which prevented Super Duck from using the phrase “duck tour” in its own trade name. The district court also clearly erred in finding a likelihood of confusion between Super Duck’s and Boston Duck’s design marks. Therefore, the court also should not have issued an injunction preventing Super Duck from using its design mark. We dismiss Boston Duck’s cross-appeal as moot given our resolution of the other issues.

I.

A. Duck Tours

We recite the background facts here, adding additional information throughout the analysis where appropriate. Both Boston Duck and Super Duck are in the business of offering sightseeing tours via land and water in Boston, using amphibious vehicles commonly referred to as “ducks.” The idea to market such tours to the public was hatched by Melvin Flath shortly after World War II. He purchased DUKWs (pronounced “ducks”), amphibious army vehicles used in the war to function as both trucks and boats, and fitted them with bus seats. His company, based in Wisconsin Dells, Wisconsin, became the first purveyor of such tours in the country.

Recently, these amphibious tours have surfaced in many cities in the United States and abroad, including San Francisco, Philadelphia, Seattle, Chicago, and London. Many of the companies that offer these tours use the phrase “duck tour” in their trade names as a way of describing their services. They also use various logos featuring a cartoon duck and water to advertise and represent their businesses.

B. The Parties

Since 1994, Boston Duck has offered sightseeing tours of the Boston area and Charles River, largely with genuine renovated DUKWs from World War II. Andy Wilson, the founder of Boston Duck, started the company after a visit to Memphis, Tennessee in 1992, where he experienced his first tour. The vehicles, which are painted with a rainbow of colors and marked with the Boston Duck logo,1 take customers to a number of Boston landmarks, including Charles Street, the State House, the North End, Quincy Market, and Newbury Street. Boston Duck has received national and international press coverage as well as numerous awards for its tours. The tours were featured in the World Series Parade after the Boston Red Sox won the World Series in 2004 and after the Patriots’ 2002, 2004, and 2005 Super Bowl wins. In 2006 alone, over 585,000 customers enjoyed a tour operated by Boston Duck.

Begun in 2001, Super Duck is also in the business of providing sightseeing tours by land and water. Rather than using renovated DUKWs, Super Duck uses custom-made amphibious vehicles called Hydra-Terras containing several features that make the vessels virtually unsinkable. Because of the Hydra-Terras’ features, which make them “larger and more modern” than the original DUKW vehicles, Super Duck decided to adopt the trade name SUPER DUCK TOURS in 2001. The name and the company’s super-hero theme seek to portray services that are “bigger and newer”; the company essentially combined the word “super” with the description of the service offered, “duck tours.” The “super” theme was also reflected in the company’s advertisements and website, *9which contain a parody of Superman (“It’s a bus. It’s a boat. It’s a Super Duck!”) as well as its logo, which consists of a white cartoon duck with an orange bill, muscular arms, and a cape.

Super Duck opened its business in Portland, Maine in 2001 and operated exclusively there until 2003, when the company ceased its operations in Maine and, seeking a larger New England market, started planning operations in Boston. Unable to obtain the necessary approvals and Hackney licenses to begin operations on its own, Super Duck purchased New England Tours in 2006 and thereafter obtained the necessary permits and licenses to conduct its tours. In late 2006 and 2007, Super Duck took steps to begin its Boston operation, including launching its website, advertising its services in local publications, and making substantial capital and resource investments in its operations. In May 2007, Super Duck entered into an agreement with Discover Boston, an independent company that offers trolley tours, whereby Discover Boston agreed to sell tickets for Super Duck tours in exchange for a sales commission. Super Duck introduced its tour operations the same month; the tours avoid the Back Bay area, a focus of Boston Duck’s tours, instead concentrating on the Boston waterfront area.

C. The Trademarks

Boston Duck owns several state and federal trademark registrations for the composite2 word mark BOSTON DUCK TOURS and a composite design mark consisting of the company’s name and logo in connection with its sightseeing tour services. For both of these marks, Boston Duck’s federal registrations on the Principal Register3 are subject to disclaimers for the terms “duck” and “tours,” meaning that it does not possess exclusive rights to use either term separate and apart from its full, registered mark. See, e.g., Dena Corp. v. Belvedere Int’l, Inc., 950 F.2d 1555, 1560 (Fed.Cir.1991). Boston Duck also owns two federal trademark registrations, one for the composite word mark BOSTON DUCK TOURS and the second for its composite design mark, in connection with “clothing, namely, sweatshirts, T-shirts, golf shirts, sweaters, visors, and hats.”4 Boston Duck claims a first-use date of 1993 for three of its four registrations; the fourth, the word mark in connection with apparel, claims a first-use date of 1995. Super Duck does not dispute that Boston Duck maintains trademark priority rights in the Boston area with respect to its composite marks.

Super Duck maintains one federal registration for the word mark SUPER DUCK TOURS on the Supplemental Register,5 in *10connection with its tour services. Super Duck filed for the mark in 2001 on the Principal Register, but its application was rejected by the United States Patent and Trademark Office (the “PTO”), which ruled that the phrase was merely descriptive, and therefore not entitled to protection unless Super Duck could establish secondary meaning. As a result, Super Duck agreed to register the mark on the Supplemental Register, and further, to disclaim exclusive use of the phrase “duck tours.” The mark has been registered since July 2003, and Super Duck claims a first-use date of 2001.

D. Procedural History

On July 2, 2007, Boston Duck filed a complaint against Super Duck Tours, alleging federal trademark infringement in violation of Section 32(1) of the Lanham Act, codified at 15 U.S.C. § 1114, federal and state unfair competition, tortious interference with prospective business relationships, and a flock of other state and federal claims. In connection with these claims, Boston Duck asked the court for a temporary restraining order and preliminary injunction to prevent Super Duck from using its trademark or logo, or any other mark confusingly similar to Boston Duck’s mark BOSTON DUCK TOURS, in connection with its services. Further, it sought to enjoin Super Duck from interfering with Boston Duck’s relations with prospective customers.

After receiving affidavits and exhibits from both parties, the district court held a non-testimonial hearing on the motion for a preliminary injunction on July 11. Two days later, the court entered its judgment, granting Boston Duck’s motion in part and denying it in part. Specifically, the court found the term “duck tours” to be nongeneric for amphibious sightseeing tours in the Boston area, and therefore capable of trademark protection. The court used the eight-part likelihood of confusion test established in Pignons S.A. de Mecanique de Precision v. Polaroid Corp., 657 F.2d 482, 487 (1st Cir.1981) (the “Pignons factors” or “Pignons analysis”),6 and concluded that Boston Duck had established a likelihood of success on the merits of its trademark infringement claim. Accordingly, the court enjoined Super Duck from “using the term ‘duck tours’ or a cartoon duck as a trademark ... in association with its sightseeing tour service in the greater Boston area.” Further, it enjoined Super Duck from using the term “duck” in a two-word or three-word trademark in conjunction with either “Boston” or “tours.”

The court denied Boston Duck’s preliminary injunction request on the tortious interference claim, finding that the tortious conduct alleged by Boston Duck was in fact committed by Discover Boston, an independent agency unrelated to Super Duck. Finally, pursuant to Fed.R.Civ.P. 65(c), the court required Boston Duck to post a bond in the amount of $100,000 as security for any costs and damages that may be incurred or suffered by Super Duck should it be determined that the injunction was improvidently granted.

Both Boston Duck and Super Duck challenge on appeal aspects of the district *11court’s judgment. Super Duck challenges the district court’s preliminary injunction, arguing that the court erred by not holding that the phrase “duck tours” and the image of a cartoon duck are generic for the services both companies provide. Super Duck further claims that because of this error, the court gave undue weight to the term “duck tours” and the image of a cartoon duck when comparing the parties’ respective marks, and thereby improperly found a likelihood of confusion between the marks BOSTON DUCK TOURS and SUPER DUCK TOURS as well as the parties’ design marks. Boston Duck cross-appeals, arguing that the district court erred by not enjoining Super Duck from using the word “duck” or “ducks” in any part of its trademark in conjunction with its services.

II.

When deciding a motion for a preliminary injunction, a district court weighs several factors: “(1) the plaintiffs likelihood of success on the merits; (2) the potential for irreparable harm in the absence of an injunction; (8) whether issuing an injunction will burden the defendants less than denying an injunction would burden the plaintiffs; and (4) the effect, if any, on the public interest.” United States v. Weikert, 504 F.3d 1, 5 (1st Cir.2007). The first factor, the plaintiffs likelihood of success, is “the touchstone of the preliminary injunction inquiry.” Philip Morris, Inc. v. Harshbarger, 159 F.3d 670, 674 (1st Cir.1998). “[I]f the moving party cannot demonstrate that he is likely to succeed in his quest, the remaining factors become matters of idle curiosity.” New Comm Wireless Servs., Inc. v. SprintCom, Inc., 287 F.3d 1, 9 (1st Cir.2002).

On appeal, we review the grant of a preliminary injunction for abuse of discretion. Weikert, 504 F.3d at 6. Within that framework, findings of fact are reviewed for clear error and issues of law are reviewed de novo. Wine and Spirits Retailers, Inc. v. Rhode Island, 418 F.3d 36, 46 (1st Cir.2005). “[W]e will set aside a district court’s ruling on a preliminary injunction motion only if the court clearly erred in assessing the facts, misapprehended the applicable legal principles, or otherwise is shown to have abused its discretion.” Id.

III.

Before addressing the specific arguments raised on appeal, we set forth as essential background some basic principles of trademark law related to the use of generic terms as marks. Even though Super Duck has acknowledged that the mark BOSTON DUCK TOURS is entitled to trademark protection, thereby conceding the first element of Boston Duck’s trademark infringement claim, see Borinquen Biscuit Corp. v. M.V. Trading Corp., 443 F.3d 112, 116 (1st Cir.2006), we discuss these background trademark principles because they inform our genericism analysis of the phrase “duck tours.”

Although trademarks are not explicitly authorized by the United States Constitution,7 our federal legal system offers statutory protection for entities that adopt and use trademarks to advertise and market *12their goods and services.8 See generally 15 U.S.C. §§ 1051-1129 (the Lanham Act). This protection is justified by the substantial benefits that trademarks offer for both consumers and businesses. Serving to distinguish and identify goods, as well as their sources, trademarks concisely impart information to consumers, reducing their search costs and allowing them to make decisions that more closely coincide with their preferences. See Ty Inc. v. Perryman, 306 F.3d 509, 510 (7th Cir.2002) (“The consumer who knows at a glance whose brand he is being asked to buy knows whom to hold responsible if the brand disappoints and whose product to buy in the future if the brand pleases.”). Because consumers rely heavily on trademarks when making choices, businesses also have an incentive to maintain product quality, lest they lose disappointed consumers. See id.

The considerable reliance on trademarks by consumers also creates an incentive for other competing, and typically less successful businesses, “to pass off their inferior brand as the successful brand by adopting a confusingly similar trademark, in effect appropriating the goodwill created by the producer of the successful brand.” Id. Trademark law is designed, in part, to prevent these “passing-off’ practices and the consumer confusion that results from it. See id. Trademark infringement law is specifically targeted to address this concern.

To succeed on a claim of trademark infringement, a plaintiff must establish (1) that its mark is entitled to trademark protection, and (2) that the allegedly infringing use is likely to cause consumer confusion. Borinquen Biscuit, 443 F.3d at 116. We have interpreted “likely confusion” to mean “more than the theoretical possibility of confusion.” Int’l Ass’n of Machinists and Aerospace Workers, AFL-CIO v. Winship Green Nursing Ctr., 103 F.3d 196, 200 (1st Cir.1996). In other words, the allegedly infringing conduct must create “a likelihood of confounding an appreciable number of reasonably prudent purchasers exercising ordinary care.” Id. at 201.

A mark is entitled to trademark protection if it is capable of functioning as a source-identifier of goods. Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 769, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992). Trademark law categorizes proposed marks along a spectrum of distinctiveness,9 based on their capacity to serve such a source-identifying function. A mark is classified as: (1) generic (least distinctive), (2) descriptive, (3) suggestive, (4) arbitrary, or (5) fanciful (most distinctive). Id. at 768, 112 S.Ct. 2753 (citing Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4, 9 (2d Cir.1976)). Marks classified as suggestive—COP— COPPERTONE screen10—arbitrary— *13APPLE for computers 11—or fanciful— EXXON for petroleum products12—are all considered inherently distinctive because they have the capacity to serve a source—identifying function upon first use. As the Fourth Circuit has recited:

The more distinctive the trade mark is, the greater its influence in stimulating sales, its hold on the memory of purchaser and the likelihood of associating similar designations on other goods with the same source. If the trade mark is a coined word, such as Kodak, it is more probable that all goods on which a similar designation is used will be regarded as emanating from the same source than when the trade-mark is one in common use on the variety of goods, such as “Gold Seal” or “Excelsior”.

Arrow Distilleries v. Globe Brewing Co., 117 F.2d 347, 349 (4th Cir.1941) (quoting Restatement (First) of Torts § 731 cmt. e (1938)) (quotation marks omitted). Accordingly, suggestive, arbitrary, and fanciful terms may be registered as trademarks on the Principal Register under the Lan-ham Act without producing any actual evidence of their source-identifying attributes or the public’s perception of these marks.

Descriptive marks are those that “convey! ] an immediate idea of the ingredients, qualities or characteristics of the goods” to which they are attached, see Equine Techs., Inc. v. Equitechnology, Inc., 68 F.3d 542, 544 (1st Cir.1995) (quoting Blinded Veterans Ass’n v. Blinded Am. Veterans Found., 872 F.2d 1035, 1040 (D.C.Cir.1989)), such as SPORTS ILLUSTRATED for a sports magazine. Because descriptive marks are not inherently capable of serving as source-identifiers, such marks may only be registered on the Principal Register after the owner has provided sufficient evidence to establish that the public associates the term or phrase not only with a specific feature or quality, but also with a single commercial source. See 2 McCarthy, supra, § 15:1. When a descriptive phrase becomes associated with a single commercial source, the phrase is said to have “acquired distinctiveness” or “secondary meaning,” and therefore functions as a trademark. See 15 U.S.C. § 1052(f) (“[N]othing [herein] shall prevent the registration of a mark used by the applicant which has become distinctive of the applicant’s goods in commerce.”). For example, when the public perceives the phrase SPORTS ILLUSTRATED as a particular sports magazine in addition to its primary meaning as a description of a specific feature or element, the phrase has “acquired distinctiveness” or “secondary meaning” and may receive trademark protection. “A mark is, therefore, considered distinctive (and, thus, eligible for trademark protection) if it either is inherently distinctive or exhibits acquired distinctiveness gained through secondary meaning.” Borinquen Biscuit, 443 F.3d at 116-17.

Finally, a generic term, such as “car” or “pizza,” is one that does not have capacity as a source-identifier because it *14designates the class, or “genus” of goods. See Abercrombie & Fitch, 537 F.2d at 9; Colt Def. LLC v. Bushmaster Firearms, Inc., 486 F.3d 701, 705 (1st Cir.2007) (“ ‘A generic term is one that does not distinguish the goods of one producer from the goods of others. Instead, it is one that either by definition or through common use has come to be understood as referring to the genus of which the particular product is a species.’ ” (quoting Keebler Co. v. Rovira Biscuit Corp., 624 F.2d 366, 373-74 (1st Cir.1980))). Rather than answering the question “where do you come from?”, a generic term merely explains “what are you?” Id. (quoting 2 McCarthy, supra, § 12:1).

Because they serve primarily to describe products rather than identify their sources, generic terms are incapable of becoming trademarks, at least in connection with the products that they designate. See Park ‘N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189, 193-94, 105 S.Ct. 658, 83 L.Ed.2d 582 (1985). Awarding trademark rights to any user of the term, especially the first user, would harm competitors and consumers alike. Competitors unable to use a common term that describes or designates their product are at a significant disadvantage communicating to potential customers the nature and characteristics of the product. See Vanessa Bowman Pierce, If it Walks like a Duck and Quacks like a Duck, Shouldn’t it be a Duck?: How a “Functional” Approach Ameliorates the Discontinuity Between the “Primary Significance” Tests for Genericness and Secondary Meaning, 37 N.M. L.Rev. 147, 154 (2007). Likewise, consumers will be forced either to pay a higher price to purchase the desired goods from the seller who owns the generic term as a trademark or expend additional time investigating the alternative products available. Id. Therefore, in accord with the primary justifications for protecting trademarks—to aid competition and lower consumers’ search costs—the law does not grant any party exclusive rights to use generic terms as trademarks.

We turn now to Super Duck’s several challenges to the conclusions of the district court.

IV.

We begin with Boston Duck’s trademark infringement claim against Super Duck based on the parties’ word marks. On appeal, Super Duck challenges the district court’s conclusion that Super Duck’s use of its mark SUPER DUCK TOURS is likely to cause confusion among the relevant consumer public with Boston Duck’s use of its mark BOSTON DUCK TOURS.13 Specifically, it argues that the court erred by finding that the phrase “duck tours” is nongeneric for amphibious sightseeing tours and therefore capable of receiving trademark protection. Because of this erroneous conclusion, Super Duck asserts that the district court gave undue weight to the phrase “duck tours” in the parties’ respective marks when analyzing the relevant Pignons factors. For example, rather than focusing on the nongeneric portions of each party’s mark, “BOSTON” and “SUPER,” respectively, when comparing the marks, the court treated each element in the parties’ composite marks equally when comparing them with one another. If the district court had deemed “duck tours” generic as the law requires, Super Duck contends, it would have de-emphasized that element in each party’s *15mark and found that the differences between the composite marks significantly outweighed the similarities. Further, the court’s analysis of other Pignons factors, including Super Duck’s intent in adopting its mark and the consideration of evidence of actual confusion, would also have been altered in favor of Super Duck. As a result, the district court would not have found a likelihood of confusion between the marks and granted an injunction in favor of Boston Duck.

A. The Eight-Factor Pignons Test

To determine whether the district court’s conclusion regarding the second infringement requirement—a likelihood of confusion between the marks—was clearly erroneous, we must assess the district court’s Pignons analysis of the composite marks BOSTON DUCK TOURS and SUPER DUCK TOURS. We review the district court’s factual findings for each Pignons factor for clear error. See Equine Techs., 68 F.3d at 546. Then we assess whether the court’s overall conclusion— that Boston Duck was able to show a likelihood of confusion between the marks—is erroneous. “The determination as to whether a likelihood of confusion exists is a question of fact, which we review only for clear error.” Id.

Where the district court’s analysis is based on an incorrect legal premise, or it has not addressed a relevant and required issue, we could remand the case to the district court for further consideration in light of the legal errors we have identified. However, where the underlying facts are largely undisputed, as they are in this ease, we think that course of action would be a waste of judicial resources and incompatible with the urgency of the issues before us. See Guglietti v. Sec’y of Health & Human Servs., 900 F.2d 397, 399-400 (1st Cir.1990) (choosing not to remand where the facts were not in dispute and the district court was in no better position to decide the matter); Kos Pharm., Inc. v. Andrx Corp., 369 F.3d 700, 712 (3rd Cir.2004) (noting that it would be a waste of judicial resources to remand a trademark action where the district court committed legal error, but the facts necessary to make the necessary likelihood of confusion determination are undisputed); Opticians Ass’n of Am. v. Indep. Opticians of Am., 920 F.2d 187, 194-98 (3d Cir.1990) (addressing the remaining elements necessary to support a preliminary injunction without remand where the district court had erroneously found that plaintiff was unlikely to succeed on the merits of its claim). Where the district court’s analysis reveals a legal error, we address those conclusions de novo. On the remaining factors—the similarity of the goods, the relationship between the parties’ channels of trade, the relationship between the parties’ advertising, and the classes of prospective purchasers—we defer to the district court’s analysis.

1. Strength of the Plaintiffs Mark

Although the eight-factor Pignons analysis typically begins with an evaluation of the similarities and differences between the marks at issue, there is no necessary sequence in the eight-factor analysis. In this case, Super Duck has asserted that an element of Boston Duck’s mark, “duck tours,” is generic. This genericism challenge implicates the strength of Boston Duck’s mark. Moreover, for the reasons asserted by Super Duck, if the district court’s conclusion on the genericism question was clearly erroneous, the court’s analysis of several other Pignons factors was necessarily flawed as well. Thus, we begin our discussion with the district court’s evaluation of the phrase “duck tours” and, consequently, its analysis of *16the strength of Boston Duck’s composite trademark BOSTON DUCK TOURS.

a. The district court’s genericism analysis

Evaluating the strength of Boston Duck’s composite mark, the district court limited its discussion exclusively to the phrase “duck tours,” assessing both the phrase’s conceptual strength—its inherent distinctiveness—and its commercial strength.14 The court found the phrase “duck tours” was nongeneric. In reaching this conclusion, the court relied exclusively on a dictionary definition of “duck,” which did not include any reference to DUKWs or amphibious vehicles. See 2 McCarthy, supra, § 12:13 (listing “dictionary definitions” as one of several factors a court should consider when determining whether a phrase is generic). The court’s analysis of the genericism issue consisted of the following:

Unlike the examples of “wool felt” and “crab house” identified by the defendant, “duck tours” does not clearly refer to a broad, general category of services. Rather, the word “duck” is a play on the World WAR II amphibious vehicles, DUKWs, which were the distinctive predecessors of the vehicles used by Plaintiff, Boston Duck Tours. The tour itself does not involve ducks, as creatures, in any way. Contrary to what the name “duck tours” may imply, the tour does not involve either duck watching or duck hunting. Thus, unlike “wool felt” or “crab house” in which the generic terms referred exactly to what the public would assume the words to mean, the link here is more attenuated.

Although it did not classify “duck tours” as descriptive, suggestive, or fanciful, it concluded that the conwas capable of receiving trademark protection.15

The district court was correct in analyzing both the conceptual and commercial strength of plaintiffs mark in this case.16 Although we have tended to focus *17our strength analysis on “practical matters [related to a mark’s perception in the marketplace] and not the legal classification of the mark,” Attrezzi, LLC v. Maytag Corp. 436 F.3d 32, 40 (1st Cir.2006); see Beacon Mut. Ins. Co. v. OneBeacon Ins. Group, 376 F.3d 8, 19 (1st Cir.2004), even going so far as to say that consideration of the inherent distinctiveness of a mark was “a proposition [not] supported by any First Circuit case law and its logic is not apparent to us,” Attrezzi, 436 F.3d at 40, the conceptual strength of the overall mark, and its components, is indeed relevant where a composite mark contains an arguably generic phrase.17 If we were to measure BOSTON DUCK TOURS’ strength only on the basis of practical, commercial factors, we would risk affording greater protection to the mark than is warranted. Specifically, we would risk recognizing commercial strength that is the result of the descriptive force of the unprotectable generic element “duck tours.”18 See Colt Def., 486 F.3d at 705 (“Generic terms, being the least distinctive, receive no trademark protection.”). For this reason, our statement in Attrezzi, where we determined the strength of a protectable, unitary mark, does not apply to this case.19

b. Is the phrase “duck tours” generic?

To assess the district court’s conclusion on the strength of plaintiffs mark BOSTON DUCK TOURS, we must begin our analysis with a discussion of whether the phrase “duck tours” is generic in connection with the services offered by the parties.20 See Am. Cyanamid Corp. v. *18Connaught Labs., Inc., 800 F.2d 306, 308 (2d Cir.1986) (noting that for the same reasons that trademark law does not grant protection to generic words or phrases, it also does not extend protection to a generic component of a trademark). Although the presence of a generic term or phrase in a full mark, such as “pizza” in DOMINO’S PIZZA, will not “render the entire mark invalid, its presence does affect the analysis of whether a competitor’s mark containing the same component is likely to create confusion.” Id. As noted above, a generic term is one whose “primary significance ... to the relevant public” is as an identifier of the nature of a good, rather than its source. Colt Def., 486 F.3d at 705; see also 15 U.S.C. § 1064(3). A question of fact, 2 McCarthy, supra, § 12:12, “the district court’s finding [that the phrase is nongeneric] ... [is] subject to review only for clear error,” Boston Beer Co. v. Slesar Bros. Brewing Co., 9 F.3d 175, 180 (1st Cir.1993).

Recognizing the deference owed to the district court and the inherent time pressures a court faces when deciding complex matters at the preliminary injunction phase, we nevertheless conclude that the court committed clear error by finding that the phrase “duck tours” is nongeneric. In undertaking its analysis, the district court did not identify and apply the appropriate legal standard in this case. As a result, its factual conclusion on the generi-cism question was erroneous.

i. The Legal Standard

In undertaking the genericism analysis, the court overlooked relevant evidence in the record on the genericism question. Tasked with determining the “primary significance” of the phrase at issue “to the relevant public,” the law instructs courts to consider several factors when undertaking this analysis: (1) consumer surveys; (2) the use of the term in media publications, (3) use of the term by competitors in the industry, (4) purchaser testimony concerning the term; and (5) the plaintiffs use of the term. Colt Def, 486 F.3d at 705. Nevertheless, as noted above, the court relied exclusively on one dictionary definition of “duck” to reach the conclusion that “duck tours” is nongeneric. Although the presence or absence of a word in the dictionary, and its corresponding meaning(s), is evidence of how the public perceives a term, Surgicenters of Am., Inc. v. Med. Dental Surgeries, Co., 601 F.2d 1011, 1015 n. 11 (9th Cir.1979) (“While not determinative, dictionary definitions are relevant and often persuasive in determining how a term is understood by the consuming public, the ultimate test of whether a trademark is generic .... ”), it is only one of many factors to consider. The touchstone of the analysis remains the phrase’s primary significance to the relevant public.

In addition, the district court focused on the separate terms “duck” and “tours,” not the entire phrase. See Filipino Yellow Pages, Inc. v. Asian Journal Publ’ns., Inc., 198 F.3d 1143, 1150 (9th Cir.1999) (finding that the genericism analysis should be conducted by looking at an entire mark rather than dissecting it into its smaller parts). Rather than analyzing each word separately, thereby altering the context of the inquiry, the court should have considered “duck tours” as a unified phrase. As courts have found in related contexts, a complete phrase may signify *19something different than the sum of its parts. See AmCan Enters., Inc. v. Renzi, 32 F.3d 233, 234 (7th Cir.1994) (“yellow pages” generic for a business telephone directory); American Express Co. v. Mastercard Int'l. Inc., 685 F.Supp. 76, 78 (S.D.N.Y.1988) (“gold card” generic for credit card services). The district court’s decision to focus exclusively on the individual elements of the phrase, and each word’s respective dictionary definition,21 removed the inquiry from its proper context and tainted the overall analysis.

ii. Application

Because the district court misapprehended the legal framework governing the genericism inquiry, its ultimate factual conclusion that “duck tours” is nongeneric was necessarily erroneous as well. In reaching its conclusion, the district court did not consider evidence in the record suggesting that the phrase “duck tour” is generic for the amphibious, sightseeing tours offered by both parties. Specifically, the court did not consider three types of evidence typically considered integral to the genericism determination: uses (1) by the media and other third parties, (2) within the industry generally, and (3) by Boston Duck itself.

With respect to the first type of relevant evidence, the district court overlooked articles in the media and other third-party sources that use the phrase “duck tours” generically to refer to amphibious, sightseeing tours. For example, a Boston Globe article from 2002 notes that rather than walking around Portland, Maine on foot, a visitor may use “town trolleys, duck tours, and harbor cruises.” Beth Green-berg, Tattoos and Views The Best-Known Artists Craft Their Designs in Some of the Region’s Most Interesting Cities, Boston Globe, Oct. 13, 2002, at M13. Another Boston Globe article published in 1997 used the phrase generically, but in a different context. In asserting that the Boston mayor needed to manage traffic issues better in the city, the author wrote: “It is important that city government remembers that it owns the streets. Tour buses, ersatz trolleys, and amphibious duck tours need to have their routes and parking carefully examined.” Lawrence S. DiCara, Running Against Himself; Mayor Meni-no’s Real Race will be to Forge a Legacy after Reelection Tuesday, Boston Globe, Nov. 2, 1997, at Cl. These articles from Boston join many others from around the country which use “duck tours” in a generic manner. See, e.g., Schwan’s IP, LLC v. Kraft Pizza Co., 460 F.3d 971, 975 (8th Cir.2006) (discussing the relevance of newspaper articles using the phrase “brick oven” to name a type of pizza rather than a specific brand).

Second, the court overlooked the widespread generic use of “duck” and “duck tours” by other companies around the country that provide the same amphibious sight-seeing services. According to the record, there are at least 36 companies in operation today that provide tour services in cities around the United States and the world. Of these 36 tours described in the record, 32 include the term “duck” in their company or trade name, and more than 10 use both the terms “duck” and “tour(s)” in their trade name. Many of these companies characterize their own services generically as “duck tours.” Taken together, this evidence indicates that when consum*20ers hear the term “duck tours,” they associate it primarily with a product rather than a source. See Colt Def., 486 F.3d at 706 (“This evidence is probative of generic use because ‘[t]he more members of the public see a term used by competitors in the field, the less likely they will be to identify the term with one particular producer.’ ” (quoting Classic Foods Int’l Corp. Kettle Foods, Inc., 468 F.Supp.2d 1181, 1190 (C.D.Cal.2007)) (alteration in original)); Schwan’s IP, 460 F.3d at 975 (finding generic use of a term by a company’s competitors is indicative of public perception).

Finally, the district court did not consider Boston Duck’s own generic use of the phrase “duck tour,” which provides strong evidence against its claim that the term is primarily associated with its company rather than the services it provides. See Retail Servs., Inc. v. Freebies Publ’g, 364 F.3d 535, 545 (4th Cir.2004) (“[E]vi-dence of the owner’s generic use, in particular, ‘is strong evidence of genericness.’ ” (quoting 2 McCarthy, supra, § 12:13)). On its website, the company notes that “[cjontrary to local belief, the unique idea of a[d]uck [t]our did not originate in Boston. Duck operations have been in existence in the Midwest for decades, and in fact, continue to thrive. What we did, however, is take a unique product and improve and enhance it, while at the same time bringing it to a major metropolitan city.” See Boston Duck Tours, http://www. bostonducktours.com/ company—history— main.html (last visited March 31, 2008). Although Boston Duck is quick to point out the introductory clause to its website statement, which asserts that the local population believes Boston Duck is the only provider of “duck tours,” this speculative statement about what the Boston population believes is largely irrelevant for purposes of our analysis. The two sentences, and the website more generally, are published by Boston Duck to inform consumers in the Boston area and around the world about the company’s history and services. Accordingly, the site instructs these customers that the phrase “duck tours” is the common phrase used to describe amphibious sightseeing tours rather than Boston Duck’s product alone.

In addition, several articles submitted by Boston Duck to the district court to establish the company’s fame and recognition provide evidence of the company’s generic use of the phrase “duck tours” and an unwitting acknowledgment of the generic nature of the phrase. In a Business Week article from 1998, the author states that “[a]t the time [when the founder of Boston Duck came up with the idea], there were only two other Duck tours in the country.” Edith Hill Updike, A Nice Business Built on Being Nice, Business Week, Sept. 14, 1998, at 10. In a similar article describing the history of Boston Duck in Boston Business Journal, the writer states that “[d]uck tours have been popular for decades in the Lake Country of Wisconsin and in southern resort towns.” Lauren John, Rough Road to Amphibious Landing on the Charles, Boston Business Journal, June 30, 1995.

Indeed, all of this evidence, especially the widespread use of “duck” and “duck tours” by companies in the industry, indicates that no other “commonly used” and effective “alternative[s]” denote the sightseeing services both parties offer. See A.J. Canfield Co. v. Honickman, 808 F.2d 291, 305-06 (3d Cir.1986). Although alternatives do exist, such as “amphibious tours,” none have been widely adopted in the industry. Unlike “duck tours,” they fail to describe the services accurately and concisely. See Bayer Co. v. United Drug Co., 272 F. 505, 511 (S.D.N.Y.1921) (finding “Aspirin” generic in part because the pub-*21lie did not accept the alternative names to describe the product); Ty Inc. v. Softbelly’s, Inc., 353 F.3d 528, 532 (7th Cir.2003) (“[B]ecause ‘beanies’ is so much shorter and punchier than the alternatives that have emerged so far for designating the product, such as ‘plush beanbag animals,’ Ty may be fighting a losing war to keep its ‘Beanies’ trademark from becoming ‘beanies’ a generic term.”). Because of the lack of adequate alternatives, “duck tours”— the most common and accepted term to describe the services at issue—may not be appropriated for one party’s exclusive use. To grant Boston Duck exclusive rights to use the phrase in the Boston area would be to erect a barrier of entry into the marketplace, thereby preventing other entities, such as Super Duck, from calling their product by its name. Super Duck, as well as other potential competitors, would be placed at a significant market disadvantage. See Devan R. Desai & Sandra L. Rierson, Confronting the Genericism Conundrum, 28 Cardozo L.Rev. 1789, 1851 (2007).

Boston Duck’s Counter-Arguments

Seeking to convince us that the district court’s nongenericism determination was not clearly erroneous, Boston Duck raises a raft of arguments in support of its position that “duck tours” is nongen-eric. Boston Duck asserts that the significant evidence of actual confusion by consumers in the marketplace between the marks BOSTON DUCK TOURS and SUPER DUCK TOURS indicates that the phrase “duck tour” possesses source-identifying attributes. Confusion could not arise, Boston Duck argues, unless consumers associate “duck tours” with Boston Duck’s services specifically. Second, Boston Duck argues that the numerous records showing that the PTO has processed trademark applications for services similar those provided by both parties without requiring a disclaimer of the term “duck” or the phrase “duck tour” suggest that the terms are neither generic nor descriptive. Third, Boston Duck claims that Super Duck should not be permitted to claim that “duck tours” is generic now given its previous stance before the PTO that neither the term “duck” nor the phrase “super duck tours” was generic or descriptive. Finally, Boston Duck argues that the dictionary definitions indicate that “duck” is nongen-eric. We address each argument in turn.

(a) Consumer Confusion

Although the documentation provided by Boston Duck establishes numerous instances of consumer confusion between the services provided by the parties, the confusion alone suggests little about whether the term “duck tours” is generic. The confusion that Boston Duck cites can easily be attributed to Boston Duck’s status as the exclusive purveyor of amphibious tours in the Boston area for over a decade rather than the source-identifying attributes of the phrase “duck tours.” Although trademark law is designed to prevent consumer confusion between inherently distinctive marks or descriptive marks that have acquired secondary meaning, it is not intended to prevent confusion between two similar, generic marks, see Blinded Veterans Ass’n v. Blinded Am. Veterans Found., 872 F.2d 1035, 1043-45 (D.C.Cir.1989), or, relatedly, between marks when one mark has acquired a “de facto secondary meaning” through its exclusive use of a generic term that causes customers to associate the term with that specific source, see Am. Online, Inc. v. AT & T Corp., 243 F.3d 812, 822 (4th Cir.2001) (“[T]he repeated use of ordinary words functioning within the heartland of their ordinary meaning, and not distinctively, cannot give AOL a proprietary right over those words, even if an association devel*22ops between the words and AOL.”); 1 McCarthy, supra, § 7:66. Boston Duck’s evidence of actual confusion, therefore, has little probative value on the question of whether the phrase “duck tours” is generic.

(b) PTO Treatment

Boston Duck overstates the weight a court must give to the decisions of the PTO regarding disclaimers. Although the PTO Examining Attorney is obligated to consider whether a disclaimer is appropriate for each application reviewed, the decision remains a discretionary choice rather than a conclusive legal determination. See In re Creative Goldsmiths of Wash., Inc., 229 U.S.P.Q. 766, 768 (T.T.A.B.1986) (“[W]e conclude that it is within the discretion of an Examining Attorney to require the disclaimer of an unregistrable component (such as a common descriptive, or generic, name) of a composite mark sought to be registered on the Principal Register under the provisions of Section 2(f).”); see also Trademark Manual of Examining Procedure, § 1213(a)-(c) (5th ed.2007). Therefore, a decision by the PTO to either require a disclaimer or not is merely a single piece of evidence in the court’s overall generi-cism analysis. See In re Nat’l Data Corp., 753 F.2d 1056, 1059 (Fed.Cir.1985) (“The power of the PTO to accept or require disclaimers is discretionary under the statute, and its practice over the years has been far from consistent. Thus, it is inappropriate to give the presence or absence of a disclaimer any legal significance.” (internal citation omitted)).

Moreover, although Boston Duck invokes several relevant trademark applications where no disclaimer of “duck” or “duck tours” was required, other applications include such disclaimers. For example, an eqtity named Liberty Duck Tours, LLC22 maintains several pending applications for the phrase LIBERTY DUCK TOURS that disclaim exclusive use of the phrase “duck tours.”23 Further, because “[t]he crucial date for the determination of genericness is the date on which the alleged infringer entered the market with the disputed mark or term,” Yellow Cab Co. of Sacramento v. Yellow Cab of Elk Grove, Inc., 419 F.3d 925, 928 (9th Cir.2005)—here, late 2006 and early 2007—the evidence put forward by Boston Duck of the PTO’s past treatment of applications has only limited relevance to the district court’s inquiry. The PTO’s most recent decision on this issue is its treatment of the applications for LIBERTY DUCK TOURS, which were considered by the PTO only last year, and which, at least in connection with the sightseeing services at issue in this case, contain a disclaimer for the phrase “duck tours.”

(c) Estoppel

Super Duck is not foreclosed from arguing now that the term “duck” or the phrase “duck tours” is generic because it argued to the contrary several years ago in front of the PTO. That contrary argument is simply another factor in the evi-*23dentiary mix. See Keebler, 624 F.2d at 375 n. 7 (allowing the defendant to argue in an infringement action that plaintiffs mark was generic despite a previous attempt to register the mark under the Puerto Rico trademark statute); Am. Rice, Inc. v. H.I.T. Corp., 231 U.S.P.Q. 793, 798 (T.T.A.B.1986) (“It is clear that while these earlier positions [of the party] may be considered as evidence, no equitable estop-pel may be derived from earlier inconsistent positions.” (internal citations omitted)).

(d) Dictionary Definitions

As already noted, the dictionary definitions in the record are not conclusive on the question of public perception. Boston Duck offered the definitions of “duck” from several common dictionaries, none of which includes a reference to amphibious World War II vehicles. Conversely, Super Duck offered a single definition of “duck” from the Oxford English Reference Dictionary, which states that a duck is an “amphibious landing-craft.” Because dictionary definitions are not conclusive indicators of overall public perception, and the specific dictionary definitions offered by the parties in this case do not significantly favor one party over the other, this factor does not materially help Boston Duck to convince us that the district court did not commit clear error in its genericism determination.

c. Strength of the Composite Mark

Having concluded that the phrase “duck tours” is generic for the parties’ services, we remain obligated to assess the strength of plaintiffs composite mark BOSTON DUCK TOURS. Evaluating the conceptual strength first, we conclude that the mark BOSTON DUCK TOURS is highly descriptive of the services Boston Duck offers. The mark combines “duck tours,” a generic phrase entitled to no trademark protection at all, with “Boston,” a weak, descriptive term that deserves minimal protection. Thus, conceptually, BOSTON DUCK TOURS as a composite mark is weak. However, BOSTON DUCK TOURS has obtained significant secondary meaning as a result of its continuous use in the Boston area for the past 13 years. This secondary meaning overcomes the mark’s initial weakness. 2 McCarthy, supra, § 11:80 (“[A] descriptive mark, upon attaining secondary meaning may be protected just as if it had been ‘strong’ and arbitrary or fanciful at its inception.”).

The evidence of secondary meaning also supports the commercial strength of Boston Duck’s mark, which is assessed by examining “the length of time the mark has been used, the trademark holder’s renown in the industry, the potency of the mark in the product field (as measured by the number of similar registered marks), and the trademark holder’s efforts to promote and protect the mark,” Borinquen Biscuit, 443 F.3d at 121; see 2 McCarthy, supra, § 11:82 (“When determining the commercial or marketplace strength of a mark, the courts look to the same kind of evidence of real world recognition of the mark as is used to decide the presence or absence of secondary meaning to determine whether a noninherently distinctive designation is or is not a valid mark.”).

Boston Duck has received significant national and international press in connection with its services, including coverage in U.S. Airways Attach, W Magazine, Cosmopolitan UK, and the Christian Science Monitor, as well as numerous awards since the mark was first used in 1994. For example, in 1996 Boston Duck won the Boston Magazine award for Best Tour in Boston. As the exclusive purveyor of duck tour services in the Boston area for over a decade, the company has become well *24known, and its revenue and number of customers have grown substantially. The number of ducks used by the company has grown from four in 1994 to twenty-four in 2006. Moreover, 585,000 customers enjoyed a duck tour operated by Boston Duck in 2006. Further, although a number of other entities around the country use the phrase “duck tours” in their trade names to describe similar services, none operate within the greater Boston area. Together, these facts indicate that the composite mark BOSTON DUCK TOURS is reasonably strong overall as an identifier of Boston Duck’s tour services in the Boston area, although part of the mark— “duck tours”—is entitled to no trademark protection at all.

2. Similarity of the Marks

The degree of similarity between two marks—-here, BOSTON DUCK TOURS and SUPER DUCK TOURS—is determined by analyzing their sight, sound, and meaning. Volkswagenwerk Aktiengesellsehaft v. Wheeler, 814 F.2d 812, 817 (1st Cir.1987). Although “similarity is determined on the basis of the total effect of the designation, rather than a comparison of individual features,” Pignons, 657 F.2d at 487; see I.P. Lund Trading ApS v. Kohler Co., 163 F.3d 27, 43 (1st Cir.1998), we give less weight to the generic portions of the parties’ respective, composite marks. 4 McCarthy, supra, § 23:49 (“If a common portion of the two conflicting marks is a public domain generic name, the emphasis of enquiry should be upon the confusing similarity of the nongeneric portion, with the ultimate issue determined by the confusing similarity of the total impression of both marks.”); see Banff, Ltd. v. Federated Dep’t Stores, Inc., 841 F.2d 486, 491 (2d Cir.1988) (finding a likelihood of confusion between B WEAR and BEE WEAR for womens’ clothing); Am. Cyanamid, 800 F.2d at 309 (HibVAX does not infringe HIB-IMUNE because the nongeneric portions of each mark are sufficiently dissimilar); cf. Beacon Mut., 376 F.3d at 19 (finding similarity between two composite marks where the dominant elements of each mark were very similar). Accordingly, we compare “the nongeneric components of a mark ... in the context of the overall composite mark.” Banff, 841 F.2d at 491.24

*25Having mistakenly concluded that the phrase “duck tours” was nongeneric, the district court placed undue emphasis on the similarity of that phrase in the parties’ marks and concluded that the differences between the composite marks was “slight.” Analyzing the similarities ourselves with the correct genericism conclusion in mind, we focus our inquiry, albeit not exclusively, on the similarities and differences between the words “Boston” and “Super,” the non-generic elements of each mark. Allowing “duck tours” to largely drop out of the analysis, we conclude that the parties’ marks are reasonably, although not completely, dissimilar. Although both “Boston” and “Super” are two-syllable words, and they are each followed by a two-word generic phrase, the words look and sound completely different. In addition, they have different meanings, describing a different feature of their products.

3. Similarity of the Goods

The district court concluded that the parties “offer virtually identical services: amphibious sightseeing tours of the greater Boston area.” It found that although slight differences exist between the products, including their routes, the differences are negligible as compared to the overall similarities. Super Duck does not challenge this conclusion, which is amply supported by the record.

A Relationship Between the Parties’ Channels of Trade, 5. Relationship Between the Parties’ Advertising, and 6. The Classes of Prospective Purchasers

After considering these three interrelated factors together, see Beacon Mut., 876 F.3d at 19, the district court found significant overlap between the companies’ channels of trade, advertising, and prospective consumers. Super Duck does not dispute that the companies are close competitors, and the court’s findings are supported by the record.

7. Evidence of Actual Confusion

“While evidence of actual confusion is ‘often deemed the best evidence of possible future confusion,’ proof of actual confusion is not essential to finding likelihood of confusion.” Borinquen Biscuit, 443 F.3d at 120 (quoting Attrezzi, 436 F.3d at 40); see Volkswagenwerk, 814 F.2d at 818. Nevertheless, as already noted, Boston Duck has documented in the record several dozen instances of actual confusion involving customers who were confused by the parties or believed that they were affiliated. Having found the phrase “duck tours” nongeneric, the district court gave considerable weight to this evidence of actual confusion as supporting Boston Duck’s claim of infringement.

Given our conclusion that “duck tours” is generic, the value of the evidence of actual confusion documented by Boston Duck is diminished substantially for the likelihood of confusion analysis. As a matter of law, the confusion was largely, if not exclusively, the product of Boston Duck’s lack of competition in its early years and the use of a generic phrase to describe its product. As such, it is not the type of confusion that warrants trademark protection. We therefore treat that evidence as largely irrelevant in reaching our overall determination.

8. Defendant’s Intent in Adopting its Mark

Whether the defendant acted in bad faith in adopting its mark is another *26factor to be considered in determining the likelihood of consumer confusion. Attrezzi, 436 F.3d at 40. The district court concluded that Super Duck’s intent in selecting the mark SUPER DUCK TOURS was “at least, suspect.” It explained this conclusion by stating that “[Super Duck] knowingly and intentionally entered the Boston market offering a service, mark and logo nearly identical to that of Boston Duck Tours, an established and successful enterprise.”

The court’s conclusion regarding Super Duck’s intent was clear error because it was based on the erroneous premise that the marks SUPER DUCK TOURS and BOSTON DUCK TOURS are highly similar. As already noted in our discussion of the similarity of the marks factor, the marks are reasonably dissimilar because the shared element in each party’s mark is the generic phrase “duck tours.” Because this foundational element of the court’s analysis was incorrect, its conclusion that Super Duck’s intent was “at least, suspect” was necessarily mistaken as well.

Super Duck adopted a name that combined a generic phrase describing its product with a significantly different descriptive term. Although there are similarities between the marks, logos, and services, and Super Duck benefited to some degree from Boston Duck introducing the Boston population to the duck tour product, these effects are natural consequences of having competition in the marketplace. See Nora Beverages, Inc. v. Perrier Group of Am., Inc., 269 F.3d 114, 124 (2nd Cir.2001) (“While intentional copying can raise a presumption of consumer confusion, ‘[t]he intent to compete by imitating the successful features of another’s product is vastly different from the intent to deceive purchasers as to the source of the product.’ ”) (quoting Streetwise Maps, Inc. v. Van-Dam, Inc., 159 F.3d 739, 745 (2d Cir.1998) (internal citation omitted)). Moreover, Super Duck created its mark and logo years before it entered the Boston market, when it began its operations in Portland, Maine.

B. Conclusion on the Likelihood of Confusion Analysis

Considering all of the factors in the Pignons analysis, we conclude that the district court’s finding that Boston Duck was likely to succeed on the merits of its trademark infringement claim is clearly erroneous. Although the companies are direct competitors because their products, potential customers, channels of trade, and advertising all significantly overlap, these factors alone are insufficient to establish a successful claim of infringement. See, e.g., Prof'l Golfers Ass’n of Am. v. Bankers Life & Cas. Co., 514 F.2d 665, 669 (5th Cir.1975) (“[D]irect competition is not the sine qua non of trademark infringement; rather the gist of the [trademark] action lies in the likelihood of confusion to the public.”). Boston Duck has failed to demonstrate bad faith on the part of Super Duck in adopting the mark SUPER DUCK TOURS. Its evidence of actual trademark confusion between the parties’ marks has little probative value because the confusion shown is largely based on Boston Duck’s lack of competition in its early years and each party’s use of the generic phrase “duck tours” in its mark.25 Where the parties are direct competitors, the most important factor in the Pignons analysis is the similarity-of-marks inquiry. See McNeil Nutritionals, LLC v. Heartland Sweeteners, LLC, 511 F.3d 350, 367 (3d Cir.2007) (“ ‘[W]hen goods are directly competing [as is undisputed in the instant case], both precedent and common *27sense counsel that the similarity of the marks takes on great prominence.’ ”) (quoting A & H Sportswear, Inc. v. Victoria’s Secret Stores, Inc., 237 F.3d 198, 214 (3rd Cir.2000)). Here, the parties’ marks—with greater emphasis placed on the words “Boston” and “Super”—are substantially dissimilar, based on sight, sound, and meaning. To the extent any similarity exists between the composite marks, that similarity is a result of each party’s decision to use a generic phrase to describe its product. Accordingly, we conclude that the Pignons factors do not establish a likelihood of confusion between the parties’ composite word marks. That conclusion means that Boston Duck has failed to establish a likelihood of success on the merits of its trademark infringement claim.

Y.

We turn now to the infringement analysis of the parties’ composite design marks.26 In reviewing Boston Duck’s petition for a preliminary injunction, the district court also found that Boston Duck was likely to succeed on its trademark infringement claim relating to the parties’ design marks.27 Accordingly, the court enjoined Super Duck from using, in connection with its services, any design that contained a cartoon duck in water.

A. The Eight-Factor Pignons Test

Without evaluating each Pignons factor separately, the court found that the parties’ design marks were “strikingly] similar[ ]” and, consequently, that Super Duck’s decision to choose a logo “nearly identical” to that of Boston Duck’s called into question its intent in choosing its mark. Relying on these findings, the court found a likelihood of confusion between the parties’ design marks. We disagree with the court’s analysis and conclusions.

Although several of the eight factors warranted independent analysis directed specifically at the parties’ design marks— including the similarity of the marks, the evidence of actual confusion, and the strength of Boston Duck’s mark—the only factor the court considered separately was the similarity of the designs. On this factor, the court limited its focus to a single element within each mark—the central image of a cartoon duck splashing in water— to the exclusion of other elements, including the colors present in each image, the background images, the content and design of the textual elements in each mark, and their overall appearance. In the absence of a discussion of the strength of Boston Duck’s mark, the similarities between the parties’ composite design marks, and evidence of actual confusion by the district court, we address these factors de novo, and then evaluate our conclusions in conjunction with the district court’s findings and our own earlier discussion of the remaining Pignons factors.

1. Strength of Plaintiff’s Mark

Again, we begin our analysis with the strength of plaintiffs composite design mark, including an inquiry into whether the central element of that mark—the image of a duck splashing in water—is generic in connection with plaintiffs duck tour services. As noted above, the strength determination, including the important genericism component, weigh *28heavily in our consideration of the other Pignons factors.

a. Genericism

The district court did not subject Boston Duck’s design to a genericism analysis in the context of assessing the strength of plaintiffs mark. Just as a word or phrase may be classified as generic, so too can an image or logo that describes a category of goods or services rather than a specific source. See 2 McCarthy, supra, § 12:34 (“A mark that consists of an illustration of the goods themselves may be ‘generic,’ and hence incapable of serving a trademark function at all.”); Kendall-Jackson Winery, Ltd. v. E. & J. Gallo Winery, 150 F.3d 1042, 1049 (9th Cir.1998) (finding that an image of a grape leaf on a wine bottle was not itself a protectable mark because it cannot serve to distinguish between brands). Accordingly, Super Duck argued that the image of a cartoon duck in water is generic for duck tour services and therefore these images should play a minimal role in the Pignons analysis of the design marks. The district court, however, did not address this issue directly; instead, it compared the marks with the underlying assumption that all of the elements of Boston Duck’s logo were nongeneric.

On appeal, Super Duck contests the district court’s implicit conclusion, arguing that the court ignored evidence establishing that virtually every duck tour company in the country uses a cartoon duck in its logo, and many of those use a cartoon duck in conjunction with an image of water. From this evidence, Super Duck argues that the image of a cartoon duck in water merely describes duck tour services rather than a specific provider thereof, and the image, therefore, is generic.

Although many duck tour providers use the image of a duck in water in connection with their products, many others do not.28 Other than industry use, there is no evidence in the record establishing how the consumer population perceives the image in question. Therefore, we have little basis upon which to conclude that the image of a duck in water is viewed by the public as merely synonymous with the phrase “duck tour,” such that it has no source-identifying capacity at all. Moreover, the evidence in the record suggests that there are effective alternatives to the image of a duck in water to identify a source of duck tour services. Accordingly, to grant an entity the exclusive use of such a design, thereby preventing other entities from using a design confusingly similar to it, would not give the first user a significant competitive advantage. Based on these factors, we cannot conclude that the district court clearly erred in its implicit finding that the design is nongeneric.29 See 2 McCarthy, supra, § 12:20 (“[T]here is only a fine line between describing [a product] and naming [it].”).

However, although we have insufficient evidence to conclude that the image of a duck splashing in water is generic for duck tour services, we do conclude that the image is highly descriptive for the parties’ services, and hence the image has weak source-identifying attributes. See Law*29rence v. P.E. Sharpless Co., 203 F. 762 (E.D.Pa.1913) (finding the image of a cow descriptive for dairy products); In re Eight Ball, Inc., 217 U.S.P.Q. 1183 (T.T.A.B.1983) (finding the picture of an eight ball and a pool cue merely descriptive of billiard parlor services); see also 1 McCarthy, supra, § 7:36 (listing relevant cases). Accordingly, less weight should be given to this specific similarity in the similarity-of-the-marks analysis.

b. Strength of the Composite Mark

Examining the overall strength of plaintiffs composite design mark, which consists of the duck image in conjunction with other design elements and Boston Duck’s trade name, we conclude that the conceptual and commercial strength of this design mark are similar to that of Boston Duck’s word mark BOSTON DUCK TOURS. Given its widespread recognition and continued use in the Boston area for over a decade, the composite design mark is reasonably strong, despite the fact that it contains elements that have an inherently weak capacity as source-identifiers.

2. Similarity of the Marks

As with word marks, when comparing design marks “similarity is determined on the basis of the total effect of the designation, rather than a comparison of individual features.” Pignons, 657 F.2d at 487 (quoting Alpha Indus., Inc. v. Alpha Steel Tube & Shapes, Inc., 616 F.2d 440, 444 (9th Cir.1980)); see 4 McCarthy, supra, § 23:25 (“Obviously, for picture and design marks (as opposed to word marks), similarity of appearance is controlling.”). Even if elements of each party’s mark overlap, or are visually similar, the marks as a whole may still create a distinct commercial impression, especially if the similarities are limited to generic or descriptive elements. See McNeil Nutritionals, 511 F.3d at 359 (“[F]orceful and distinctive design features should be weighed more heavily because they are more likely to impact the overall impression.”); see also First Sav. Bank, F.S.B. v. First Bank Sys., Inc., 101 F.3d 645, 655 (10th Cir.1996) (“When the primary term is weakly protected to begin with, minor alterations may effectively negate any confusing similarity between the two marks.”). Further, the marks should be viewed “sequentially [as if one were] in the context of the marketplace” rather than “in a side-by-side comparison.” Louis Vuitton Malletier v. Dooney & Bourke, Inc., 454 F.3d 108, 117 (2d Cir.2006).

Considering the parties’ respective designs as a whole, as they have been depicted in the record, including the colors used, the text, and the various design elements contained therein, we conclude that the parties’ design marks are substantially dissimilar. Although each contains the central image of a cartoon duck in water, the ducks are significantly different in overall appearance. Further, the other elements depicted in conjunction with the duck are substantially dissimilar. Together, these differences indicate that the designs are associated with separate, unrelated sources rather than the same one.

To highlight the differences between the marks, we briefly describe the features of each. Boston Duck’s logo consists of a purple background enclosed within a black-bordered circle. Encircling the top of the background is the name BOSTON DUCK TOURS in black, in a sans-serif font.30 Finally, a yellow cartoon duck with *30a camouflage hat sits in the center of the circle, flapping its wings into some water and causing the blue-and-white-tipped water to splash outside of the contained circle.

By contrast, Super Duck’s logo has a blue background, denoting water, and contains several orange elements as an accent. At the top portion of the logo, and getting smaller across the logo to suggest motion from the left portion of the logo to the right portion, the trade name Super Duck Tours is printed in orange, in a serif font. At the center of the logo is a white cartoon duck, shaped to depict an actual duck tour vehicle, and simulating the move from land to water that a tour vehicle makes. The duck is wearing an orange cape and holding up an orange flag with one of its two muscular arms. Passengers crowd together on the deck of the duck, which has “Super Duck Tours” emblazoned on its side. As the duck enters the water, white pockets of water splash into the air.

The many specific differences between the two marks reaffirm our conclusion about the central enquiry: if consumers viewed these two marks sequentially in the marketplace, they would not consider them substantially similar.

3. Evidence of Actual Confusion

Although the record contains evidence that customers confused the parties’ services, and perhaps their trade names, we are unable to discern any evidence of actual confusion traceable specifically to the parties’ logos. In any event, as we noted above, evidence of actual confusion is not necessary to succeed on a trademark infringement action. See Volkswagenwerk, 814 F.2d at 818.

B. Conclusion on the Likelihood of Confusion Analysis

We consider the above three factors in conjunction with our analysis of the other five Pignons factors undertaken in our discussion of the parties’ word marks. Because those five factors focus on the parties themselves, or the goods and services at issue, rather than the marks, the analysis is identical. See supra Part IV.A. Four of those factors—the similarity of the goods; the relationship between the parties’ channels of trade; the relationship between the parties’ advertising; and the classes of prospective purchasers—weigh in favor of Boston Duck. As noted above, so too does the strength of the mark. However, the absence of bad faith by Super Duck weighs against a finding of infringement. And, most importantly, the designs’ overall dissimilarity weighs strongly against a finding of likely confusion. See McNeil Nutritionals, 511 F.3d at 367 (noting that the similarity-of-the-marks factor becomes more important where the products at issue directly compete with one another). Visually, the marks are substantially different. The greatest similarity between them—the image of a cartoon duck in water—is of minimal significance because of its highly descriptive nature. Thus, we conclude that the district court’s determination that there is a likelihood of confusion between Boston Duck’s design mark and Super Duck’s design mark is clearly erroneous. Again, that conclusion means that Boston Duck has failed to establish a likelihood of success on the merits of its infringement claim.

VI.

Although we have concluded that we must reverse the decision of the district court, we have had the luxury of time in reaching our decision in this difficult case-a luxury that the district court did not have. Despite those constraints, the district court issued a thoughtful opinion *31which allowed the parties to focus their arguments in a manner that has been helpful to us.

For the reasons stated herein, Boston Duck failed to establish a likelihood of success on its claims of trademark infringement. Accordingly, the district court should not have issued a preliminary injunction against Super Duck precluding the use of its word and design marks.31 We reverse the district court’s order and remand the matter to the district court for further proceedings consistent with this opinion. Costs are awarded in favor of Appellant Super Duck Tours.

So ordered.