372 F.3d 1330
Fed. Cir.
2004
SCHALL, Circuit Judge.

Defendant-appellant in this trademark infringement suit, ICON Health and Fitness, Inc. (“ICON”), appeals from the order of the United States District Court for the Western District of Washington granting a preliminary injunction in favor of plaintiff-appellee, The Nautilus Group, Inc. (“Nautilus”). Nautilus Group, Inc. v. ICON Health & Fitness, Inc., 308 F.Supp.2d 1208 (W.D.Wash.2003) (“Preliminary Injunction Order”). The preliminary injunction, which has been stayed pending this appeal, bars ICON from using its “CrossBow” trademark to sell its exercise equipment. Nautilus is the holder of the “Bowflex” trademark, which it uses to market its patented exercise machines of the same name. Because we do not find that the district court applied the incorrect law, relied on clearly erroneous factual findings, or otherwise abused its discretion in finding a likelihood of confusion to' exist between the Bowflex and CrossBow marks, we affirm the grant of the preliminary injunction.

BACKGROUND

I.

Nautilus and ICON are direct competitors in the market for home exercise equipment. Both produce resistance training systems that use bendable rods. In Nautilus’s product, the Bowflex exerciser, the rods are arranged vertically; when a user pulls on a cable connected to the upper ends of the rods, the rods curve or “bow” outward. The rods are structured to resist this outward movement, requiring exertion by the user.

Nautilus has continuously produced the Bowflex exerciser since 1984, and holds two patents on related technology. 1 Nautilus registered the trademark for the “Bowflex” brand in 1986. 2 Since 1993, when Nautilus began to directly market its Bowflex exerciser to consumers, the company has invested in excess of $233 million in promotion efforts. Nautilus’s advertising has focused on infomercials, the Internet, and print publications. As a result, over 780,000 machines have been sold, with total revenues to Nautilus of $900 million.

In 2002, ICON, another large exercise equipment manufacturer, introduced a competing rod-based resistance training system it called “CrossBow.” In the CrossBow exerciser, the bendable rods are arranged horizontally so that they bend downward, in an inverted U-shape, rather than outward, as in Bowflex’s vertical rod configuration. According to ICON, the product’s resemblance to the medieval crossbow weapon inspired the brand name. In ICON’S two-line logo for the machine, the “o” in “Cross” is replaced by a circular crosshairs. Beneath the mark, in smaller type, appears an additional line, “by Weider,” signaling the machine’s source. 3 Since it introduced the CrossBow exerciser, ICON has spent $13 million on marketing in many of the same advertising channels as Nautilus, including infomercials and the Internet.

II.

Nautilus commenced a suit for patent infringement against ICON on December 3, 2002. Thereafter, in April of 2003, the district court granted Nautilus’s motion for leave to amend its suit to include a claim of trademark infringement. On May 16, 2003, the district court granted ICON’S motion for summary judgment of non-infringement of the Nautilus patents. This order became the subject of Nautilus’s first appeal to this court. Nautilus Group, Inc. v. ICON Health & Fitness, Inc., 82 Fed. Appx. 691 (Fed.Cir.2003) (unpublished opinion). 4

While its appeal of the patent infringement issue was pending, Nautilus moved for a preliminary injunction to suspend ICON’S use of the CrossBow trademark. The district court, applying pertinent Ninth Circuit law, determined that Nautilus had established likelihood of confusion between the marks such that irreparable injury could be presumed. The preliminary injunction was granted on July 14, 2003. Before Nautilus posted the required bond, ICON moved for reconsideration or for a stay of the injunction pending appeal. Both motions were denied by the district court, and this appeal followed. 5 We have jurisdiction pursuant to 28 U.S.C. § 1295(a)(1).

ANALYSIS

I.

Any “word, name, symbol, or device” may be treated as a “trademark” if it is “used by a person” to “identify and distinguish his or her goods, including a unique product, from those manufactured or sold by others and to indicate the source of the goods....” 15 U.S.C. § 1127 (2000). A trademark holder may bring an action for infringement to prevent or recover for the unauthorized use of its mark. Id. § 1114. In a trademark case, we apply the law of the applicable regional circuit, in this case, the Ninth Circuit. See Payless Shoesource, Inc. v. Reebok Int’l Ltd., 998 F.2d 985, 987-88 (Fed.Cir.1993). Under Ninth Circuit law, the core element of trademark infringement is whether the reasonably prudent consumer is likely to be confused as to the origin of the good or service bearing one of the marks. Dreamwerks Prod. Group, Inc. v. SKG Studio, 142 F.3d 1127, 1129 (9th Cir.1998).

In the Ninth Circuit, a district court may grant a preliminary injunction in a trademark case when the plaintiff demonstrates “either (1) a combination of ‘probable success on the merits’ and ‘the possibility of irreparable injury’ or (2) the existence of ‘serious questions going to the merits’ and ‘that the balance of hardships tips sharply in his favor.’ ” GoTo.com, Inc. v. Walt Disney Co., 202 F.3d 1199, 1205 (9th Cir.2000) (quoting Sardi’s Restaurant Corp. v. Sardie, 755 F.2d 719, 723 (9th Cir.1985)). Probable success on the merits requires a showing that the similarity of the marks, among other factors, has created a likelihood of confusion as to the source or origin of the goods. Ocean Garden, Inc. v. Marktrade Co., 953 F.2d 500, 506 (9th Cir.1991). If a likelihood of confusion between the marks is found, irreparable injury to the plaintiff may be presumed. See GoTo.com, 202 F.3d at 1205; Vision Sports, Inc. v. Melville Corp., 888 F.2d 609, 612 n. 3 (9th Cir.1989).

In accordance with Ninth Circuit law, “[a] district court’s decision to grant a motion for a preliminary injunction will be upheld unless the court ‘applied the incorrect law, relied on clearly erroneous factual findings, or otherwise abused its discretion.’ ” Ocean Garden, 953 F.2d at 502 (quoting NEC Electronics v. CAL Circuit Abco, 810 F.2d 1506, 1508 (9th Cir.1987)); see also GoTo.com, 202 F.3d at 1204. As far as the facts are concerned, “[i]f the district court’s account of the evidence is plausible in light of the record viewed in its entirety, we may not reverse even though convinced that had we been sitting as the trier of fact, we would have weighed the evidence differently.” Interstellar Starship Servs. v. Epix, Inc., 304 F.3d 936, 941 (9th Cir.2002) (citing Anderson v. Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985)). The Ninth Circuit views the ultimate conclusion of likelihood of confusion as a mixed question of law and fact, and it reviews that conclusion under the clearly erroneous standard. Levi Strauss & Co. v. Blue Bell, Inc., 778 F.2d 1352, 1355 (9th Cir.1985) (en banc).

II.

In determining that Nautilus had established a likelihood of confusion be tween its mark and ICON’S, the district court applied the eight factors set forth in AMF, Inc. v. Sleekcraft Boats, 599 F.2d 341, 348-49 (9th Cir.1979). The Sleekcraft factors include (1) the similarity of the marks, (2) the relatedness or proximity of the two companies’ products or services, (3) the strength of the registered mark, (4) the marketing channels used, (5) the degree of care likely to be exercised by the purchaser in selecting goods, (6) the accused infringers’ intent in selecting the mark, (7) evidence of actual confusion, and (8) the likelihood of expansion in product lines. See also Interstellar Starship, 304 F.3d at 942. “This eight factor test is pliant, and the relative import of each factor is case specific.” Id.

The district court began its analysis by considering together two complementary Sleekcraft factors: the similarity of the Bowflex and CrossBow marks, and the similarity of the associated products. The court observed that aside from the shared use of “bow,” the marks’ respective appearances and meanings were dissimilar. On the other hand, the court viewed the competing products as virtually interchangeable. Because the close similarity in the products was likely to create confusion, the district court reasoned that less similarity between the marks was necessary. Thus, despite acknowledged differences between the marks, Nautilus was deemed to prevail on these two Sleekcraft factors, similarity of marks and relatedness of goods.

The next Sleekcraft factor considered by the district court was the strength of the registered mark. The district court found the Bowflex mark to be “suggestive,” because it “convey[ed] an impression of a good but require[d] the exercise of some imagination and perception to reach a conclusion as to the product’s nature.” Preliminary Injunction Order, slip op. at 5. Although the district court did not view Bowflex as a particularly creative mark, it found that Nautilus had invested significantly in the brand name, so that it was now broadly recognized. For this reason, the district court determined that strength of the registered mark to be a factor weighing in Nautilus’s favor.

The court found that the fourth Sleek-craft factor, similarity of marketing channels, also weighed in favor of Nautilus. Both companies utilize identical marketing channels, including infomercials, print ads, and the Internet, to promote their products. The district court rejected ICON’S suggestion that its use of comparative advertising in these venues prevented any confusion.

Nautilus sought to establish actual confusion, another Sleekcraft factor, by presenting evidence of several telephone calls to Nautilus’s sales center from customers interested in ICON’S product. Although the court noted that these consumers might have been confused by the similarity of the products, the suggestion of any actual confusion was, in the court’s view, sufficient to confirm the possibility of future confusion.

From circumstantial evidence, the court next surmised that this potential confusion might have been intentionally created by ICON: “By naming its product ‘CrossBow,’ Defendant invited comparison and reference to the BowFlex.” Id. at 7. In the court’s opinion, ICON had unfairly capitalized on Nautilus’s successful marketing of “bow” in the context of exercise equipment. Intent to confuse was, then, another Sleekcraft factor that was found to weigh against ICON.

The degree of care in purchasing was the only factor that, in the court’s opinion, supported ICON. Because the Nautilus and ICON machines vary in price from hundreds to thousands of dollars, the district court reasoned that it would be fair to expect that customers would carefully investigate the range of available products before purchasing. The presumptive high degree of care in purchasing would minimize the possibility that consumers would buy a CrossBow based solely on their mistaken association of that machine with Nautilus’s Bowflex product, the court concluded.

Balancing the seven Sleekcraft factors it had considered, the court determined that a likelihood of confusion existed between the Bowflex and CrossBow marks:

The marks are somewhat similar, and there is evidence of actual confusion. Moreover, BowFlex is a strong mark, and there is circumstantial evidence that CrossBow chose its name, incorporating “bow,” to take advantage of BowFlex’s strong mark. Although consumers may be likely to take some care in buying an exercise machine, the possibility of confusion is increased because of the use of the same advertising channels and the similarity of the products. Analysis of the above factors favors BowFlex, and demonstrates Plaintiffs probable success on the merits.

Id. at 9. Because likelihood of confusion had been found, the court could presume that irreparable injury to Nautilus would result if injunctive relief were not granted. A preliminary injunction against ICON was therefore entered.

III.

On appeal, ICON makes several related arguments for vacating the preliminary injunction. Its primary focus is on the findings made by the district court with respect to its alleged intent to confuse, evidence of actual confusion, the strength of the Bowflex mark, and the similarity between the two marks. ICON also challenges the manner in which the Sleekcraft factors were balanced to find likelihood of confusion in what ICON views as the unique context of this case. We address each of ICON’S arguments in turn.

A.

We first consider ICON’S challenge to the district court’s finding that ICON’S choice of the CrossBow as its brand name evinced an intent to confuse consumers. After considering circumstantial evidence, the court concluded that “[wjhile CrossBow’s intent may predominantly have been to compete,” it was likely that “Defendant chose to incorporate the term ‘bow’ in an attempt to capture some of the market of BowFlex.” Id. at 8.

ICON argues that intent to confuse should not have been found. It asserts that the CrossBow brand name was chosen because the shape of the machine resembled the weapon of the same name, not because ICON wanted to benefit from goodwill associated with the Bowflex mark. ICON contend that this explanation is supported by the dissimilarity of the two marks. Had it intended to capitalize on the recognition of the Bowflex name, ICON states, it would have chosen a more similar term. As additional proof of its lack of intent, ICON relies on the denotation of source, “By Weider,” that appears next to the CrossBow mark on its machines. In ICON’S view, the addition of an explicit designation of source should negate any inference of intent to copy. Nautilus responds that because circumstantial evidence of intent to copy is sufficient to suggest intentional infringement, it was not clear error for the district court to find intent to infringe on the facts of this case.

“The law has long been established that if an infringer ‘adopts his designation with the intent of deriving benefit from the reputation of the trade-mark or trade name, its intent may be sufficient to justify the inference that there are confusing similarities.’ ” Brookfield Communications, Inc. v. W. Coast Entm’t Corp., 174 F.3d 1036, 1059 (9th Cir.1999) (quoting Pac. Telesis v. Int’l Telesis Communications, 994 F.2d 1364, 1369 (9th Cir.1993) (citation omitted)). Evidence of copying is not required to show intent; instead, it may be inferred when a similar mark is adopted by a defendant who had actual or constructive knowledge of the plaintiffs mark. Id.

Recently, however, the Ninth Circuit may have de-emphasized somewhat the role of the intent factor in the likelihood of confusion analysis. In Brookfield Communications, the defendant selected an Internet domain name that was similar to the plaintiffs mark. While the defendant was aware of the plaintiffs mark when it chose its domain name, the court found that it was unclear whether the selection had been made with intent to confuse. Because this factor did not favor either the plaintiff or the defendant, the court suggested that it was not of great assistance in determining likelihood of confusion. Id. at 1059 (“This factor is only relevant to the extent that it bears upon the likelihood that consumers will be confused by the alleged infringer’s mark (or to the extent that a court wishes to consider it as an equitable consideration).”). Subsequently, in GoTo.com, Inc. v. Walt Disney Co., the Ninth Circuit reiterated the “minimal importance” of the intent factor by suggesting that “even if ... [the defendant] was as innocent as a fawn with no intent to copy ... it would prove nothing since no intent is necessary to demonstrate a likelihood of confusion.” 202 F.3d at 1208 (emphasis added).

We agree with ICON that the district court’s analysis on this point was concluso-ry, and without more, could not justify placing the intent factor in Nautilus’s column for purposes of granting a preliminary injunction. This is not an error, however, that requires reversal of the preliminary injunction. For this reason, we see no need to consider further ICON’S specific intent.

B.

We address next ICON’S challenge to the sufficiency of the evidence presented to demonstrate actual confusion. Nautilus submitted transcripts from four calls to its customer service center to prove that actual confusion had resulted from ICON’S use of the CrossBow brand name. ICON rejoined that any consumer confusion suggested by these calls was a consequence of the virtual identity of the products, not the alleged similarity of the brand names. It further argued that evidence relating to four calls out of thousands received could not be indicative of any broader confusion. The court agreed that it was unclear “where exactly this customer confusion originates.” Preliminary Injunction Order, slip op. at 8. It nevertheless concluded, however, “that Plaintiff has produced evidence of confusion” such that this factor should be weighed in its favor. Id. at 9.

On appeal, ICON raises essentially the same arguments it made in the district court. First, it contends that because the source of the customers’ alleged confusion could not be determined, the evidence should not have been used to assess whether the Bowflex and CrossBow marks are confusingly similar. Second, it contends that de minimis evidence of just four phone calls out of thousands is, as a matter of law, insufficient. Nautilus argues in response that any showing of actual confusion is highly persuasive as to the likelihood of future confusion, particularly when the allegedly infringing mark has only been recently introduced.

As a general proposition, “a showing of actual confusion among significant numbers of consumers provides strong support for likelihood of confusion.” Playboy Enters., Inc. v. Netscape Communications Corp., 354 F.3d 1020, 1026 (9th Cir.2004); see also Official Airline Guides, Inc. v. Goss, 6 F.3d 1385, 1393 (9th Cir.1993); Sleekcraft, 599 F.2d at 352 (explaining that “evidence that use of the two marks has already led to confusion is persuasive proof that future confusion is likely”). This confusion must be rooted in the identity of the marks; confusion that arises solely from similarities in functionality cannot be captured and protected as part of trademark law. See Shakey’s Inc. v. Covalt, 704 F.2d 426, 431-32 (9th Cir.1983) (refusing to “condemn the ... defendants for physical similarities” between their product and plaintiffs because “ ‘functional’ similarities alone ... cannot be a valid basis for an unfair competition claim”). Finally, “[a] reasonable juror may ... find de minimis evidence of actual confusion unpersuasive as to the ultimate issue of likelihood of confusion.” Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1150 (9th Cir.2002). This is appropriate because trademark infringement is only actionable when a mark is likely to “confuse an appreciable number of people as to the source of the product.” Id. at 1151 (emphasis added).

We agree with ICON that the relatively small number of calls presented by Nautilus renders this evidence too unreliable to establish actual confusion. We also agree that the ambiguity as to the cause of the confusion — similarity of marks versus similarity of products — detracts from the value of this evidence as an indicator of likelihood of confusion. In fact, a closer look at the call transcripts indicates that, if any confusion existed, it arose out of the similarities in the appearance of the CrossBow exerciser and the BowFlex machine. 6 Such confusion would be expected, given that Nautilus has been the only source for this type of exercise system for the last twenty years. This market dominance would lead one to expect that consumers might associate any machine of this type with the Bowflex exerciser, regardless of that machine’s own brand name.

We agree with ICON that it was improper for the district court to consider this scant and ambiguous evidence of actual confusion in Nautilus’s favor for purposes of granting a preliminary injunction. Accordingly, as with the intent to confuse factor, we merely decline to weigh this factor in favor of Nautilus at this early stage in the proceeding.

C.

ICON next challenges the district court’s conclusion as to the strength of Nautilus’s “Bowflex” mark. The court found that, as a result of Nautilus’s significant investment in advertising, “Bowflex” has become “a strong mark.” Preliminary Injunction Order, slip op. at 6. According to the court, this strength merited a greater degree of trademark protection. Id.

“The scope of the trademark protection that we give marks depends upon the strength of the mark, with stronger marks receiving greater protection than weak ones.” Entrepreneur Media, 279 F.3d at 1141 (citing Miss World (UK) Ltd. v. Mrs. Am. Pageants, Inc., 856 F.2d 1445, 1448 (9th Cir.1988)). A strong mark is one that is inherently distinctive, such that it operates to clearly identify goods as originating with a particular source. See 15 U.S.C. § 1127. In conformity with that background principle, marks have traditionally been categorized along the following range of increasing distinctiveness: (1) generic, (2) descriptive, (3) suggestive, (4) arbitrary, or (5) fanciful. See Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992) (citing Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4, 9 (2d Cir.1976)).

A mark that is fanciful, arbitrary, or suggestive is considered to be inherently distinctive, and thus automatically qualifies for trademark protection under 15 U.S.C. § 1051. See KP Permanent Make-Up, Inc. v. Lasting Impression I, Inc., 328 F.3d 1061, 1067 (9th Cir.2003) (explaining that such marks “naturally ‘serve[ ] to identify a particular source of a product’ ” and are thus entitled to trademark protection (citations omitted)); Genesee Brewing Co., Inc. v. Stroh Brewing Co., 124 F.3d 137, 143 (2d Cir.1997). A fanciful mark is a non-dictionary word concocted by the trademark holder for its product, whereas an arbitrary mark is a known word used in an unexpected or uncommon way. Interstellar Starship, 304 F.3d at 943 n. 6. “A suggestive mark is one for which ‘a consumer must use imagination or any type of multistage reasoning to understand the mark’s significance ... the mark does not describe the product’s features, but suggests them.’ ” Entrepreneur Media, 279 F.3d at 1142 (quoting KendallrJackson Winery, Ltd. v. E. & J. Gallo Winery, 150 F.3d 1042, 1047 n. 8 (9th Cir.1998)). “[UJnlike arbitrary or fanciful marks which are typically strong, suggestive marks are presumptively weak,” but can be strengthened by achieving “actual marketplace recognition.” Brookfield Communications, 174 F.3d at 1058.

These creative marks are distinguished from descriptive marks, which “define qualities or characteristics of a product in a straightforward way that requires no exercise of the imagination to be understood.” Kendalh-Jackson Winery, 150 F.3d at 1047 n. 8. “Because they tend to consist of common words that might be the only way to describe a category of goods,” a mark that is descriptive of its underlying product is only entitled to trademark protection if it has acquired “secondary meaning” in the minds of consumers. Japan Telecom, Inc. v. Japan Telecom Am. Inc., 287 F.3d 866, 872 (9th Cir.2002). To possess secondary meaning, the term must have “become distinctive of the applicant’s goods in commerce.” Two Pesos, 505 U.S. at 769, 112 S.Ct. 2753 (quoting 15 U.S.C. § 1052(e), (f)); see also Official Airline Guides, 6 F.3d at 1391 (“A mark acquires secondary meaning if customers associate the mark with a particular source.”). 7

The line between descriptive and suggestive marks can be difficult to draw. The Ninth Circuit has suggested that courts look to “the imaginativeness involved in the suggestion, that is, how immediate and direct is the thought process from the mark to the particular product. If the mental leap between the word and the product’s attribute is not almost instantaneous, this strongly indicates suggestiveness, not direct descriptiveness.” Self-Realization Fellowship Church v. Ananda Church of Self-Realization, 59 F.3d 902, 911 (9th Cir.1995) (internal quotation marks and citation omitted); see, e.g., Entrepreneur Media, 279 F.3d at 1142 (holding that the “ENTREPRENEUR” trademark for a monthly magazine geared towards small business owners was descriptive because it “describe[d] both the subject matter and the intended audience of the magazine and programs” in such a manner that “an entirely unimaginative, literal-minded person would understand the significance of the reference”); Brookfield Communications, 174 F.3d at 1058 (finding the term “Movie-Buff” to be suggestive because “it require[d] a mental leap from the mark to the product” and did not merely “describe ... [the] product or its purpose”). As another measure of a mark’s descriptiveness, courts may look to use by competitors of the same term, because “if there are numerous synonyms for a common trademarked word, others will have less need to use the trademarked term.” Entrepreneur Media, 279 F.3d at 1143. See also Playboy Enters., 354 F.3d at 1028 n. 33 (“[Ejxtensive third-party use of a mark might tend to show that consumers are likely to associate the mark with companies and meanings other than the mark-holder’s.”).

Fanciful, arbitrary, suggestive, and descriptive words are all capable of receiving trademark protection; generic terms, on the other hand, are not. Filipino Yellow Pages, Inc. v. Asian Journal Publs., Inc., 198 F.3d 1143, 1147 (9th Cir.1999). Generic terms describe a category of products, and therefore cannot signal any particular source: “A [tradejmark answers the buyer’s questions “Who are you? Where do you come from? Who vouches for you?’ But the [generic] name of a product answers the question ‘What are you?’ ” Official Airline Guides, 6 F.3d at 1391 (quoting 1 J. McCarthy, Trademarks and Unfair Competition § 12.01 (3d ed.1992)). It would be improper to extend trademark protection to a product’s generic name because “it would effectively ‘grant [the] owner of the mark a monopoly, since a competitor could not describe his goods as what they are.’ ” Filipino Yellow Pages, 198 F.3d at 1151 (quoting Surgicenters of Am., Inc. v. Med. Dental Surgeries, Co., 601 F.2d 1011, 1017 (9th Cir.1979)). If, however, a generic term is combined with another word or words to form a distinctive term, the resulting compound can be trademarked. See Cal. Cooler, Inc. v. Loretto Winery, Ltd., 774 F.2d 1451, 1455 (9th Cir.1985) (holding that the validity of a composite term “is not judged by an examination of its parts”); Park ‘N Fly, Inc. v. Dollar Park & Fly, 718 F.2d 327, 331 (9th Cir.1983) (rejecting the argument that because “park” and “fly” were both ordinary words, their combination was also generic).

In the case at hand, the district court applied these guiding principles to find Bowflex to be suggestive, rather than descriptive. Preliminary Injunction Order, slip op. at 6. Because Nautilus had “made a significant investment” in its mark to ensure that “the public in general associates resilient rod exercise machines with Bowflex,” the court found that the presumption of weakness applied to suggestive marks was overcome in this case. Id.

On appeal, ICON argues that the “bow” component of the “Bowflex” mark is either generic, or descriptive without secondary meaning. According to ICON, the term “bow” describes a fundamental and distinctive characteristic common to the exercise machines that are the focus of this case— the use of resilient rods that bend in a bow-like fashion to produce resistance. ICON therefore asserts that the word, “bow,” is generic. In the alternative, ICON contends that “bow” is at best descriptive, because it merely “describe[s] a product, [but does] not inherently identify a particular source.” Two Pesos, 505 U.S. at 769, 112 S.Ct. 2753. ICON contends that granting trademark protection to a descriptive term like “bow” would have the effect of granting Nautilus protection, via trademark law,'over the functional characteristics of its product. ICON suggests that to avoid this result, the words and images that describe those characteristics must remain freely available for use by Nautilus’s competitors.

For its part, Nautilus maintains that its mark, “Bowflex,” is suggestive. First, it contends that the “validity of a trademark is to be determined by viewing the trademark as a whole,” Cal. Cooler, 774 F.2d at 1455, not by its component parts. Second, Nautilus explains that the term, “bow,” “suggests” key features of the product, rather than just “describing” them:

[T]he BowFlex mark calls to mind the image of an archer repeatedly bending a bow by pulling the draw string back and forth. This image does not directly describe the BowFlex exercise machine. Rather, the image of an archer drawing a bowstring back and forth is a metaphor for the resistance principle employed by the BowFlex exercise machine, which creates resistance in the manner of a bow resisting the archer’s pull on the bowstring, by providing resilient rods connected to cables, that the user pulls upon, thereby bending the rod which resists the pull on the cable. This metaphor is an intermediate stage of, reasoning between the trademark and the property of the product that it suggests, which the consumer bridges by use of his or her imagination.

In the event that Bowflex is found to be merely “descriptive,” Nautilus submits that there is extensive evidence of secondary meaning, achieved through twenty years of successful marketing.

We begin with ICON’S argument that, in the context of this case, “bow” is generic. The word “bow” could only be found to be generic if its primary significance were “to describe the type of product rather than the producer.” Anti-Monopoly, Inc. v. General Mills Fun Group, 611 F.2d 296, 304 (9th Cir.1979). Applying the “who-are-you/what-are-you” test, see Filipino Yellow Pages, 198 F.3d at 1147 (citing Official Airline Guides, 6 F.3d at 1391), it is clear that “bow” does not answer the question “what are you?” hypothetically posed by one looking at a resistance rod exercise machine. As the district court found, multiple, more plausible alternatives exist for describing this category of exercise equipment, such as, for example, “rod exerciser” or “rod bender.” In our view, the term “bow” is not necessarily used to describe “the genus of which the particular product is a species.” Abercrombie & Fitch, 537 F.2d at 9.

It is a closer question whether the term “bow,” as applied to this particular class of exercise equipment, is descriptive. Bound by our limited scope of review, we cannot say that the district court’s finding that the mark was suggestive was the result of an incorrect application of the law, was made in reliance on erroneous factual findings, or was an abuse of the court’s discretion. See Ocean Garden, 953 F.2d at 502.

To determine if a mark is descriptive or suggestive, we look to “how immediate and direct ... the thought process [is] from the mark to the particular product.” Japan Telecom, 287 F.3d at 873 (quoting Self-Realization Fellowship Church, 59 F.3d at 911). We do not think it was error for the district court to find that “Bowflex” differed from descriptive marks such as “Japan Telecom,” for a telecommunications business catering to Japanese clientele, id.; “ENTREPRENEUR,” for a magazine targeted to small business owners, Entrepreneur Media, 279 F.3d at 1142; or “micro colors,” for the pigment used in permanent makeup, KP Permanent Make-Up, 328 F.3d at 1071. In each of these cases, the marks left little, if anything, to the imagination. On the other hand, because the “mental leap” between “Bowflex” and the curving rods it describes does not strike us as “almost instantaneous,” see Self-Realization Fellowship Church, 59 F.3d at 911, we conclude that the district court reasonably found that “Bowflex” was a suggestive mark. As with terms such as “Movie-Buff,” for use in the entertainment industry, Brookfield Communications, 174 F.3d at 1058; “GoTo.com,” for an Internet search engine, GoTo.com, 202 F.3d at 1207; and “ACCURIDE,” for a slide mechanism used in drawers, Accuride, 871 F.2d at 1536; “Bowflex” is not a particularly creative or distinctive mark, but it can at least be said that it requires “imagination or any type of multistage reasoning to understand the mark’s significance.” Entrepreneur Media, 279 F.3d at 1142 (quoting Kendall-Jackson Winery, 150 F.3d at 1047 n. 8). We therefore cannot say that, for the purposes of determining whether a preliminary injunction should issue, the court clearly erred in preliminarily finding Bowflex to be a suggestive mark. For this reason, we decline to disturb the weighing of Sleekcraft’s strength of the mark factor in Nautilus’s favor.

Further, in the context of our review of the district court’s grant of the preliminary injunction, we do not think the court clearly erred in finding that Nautilus has strengthened a presumptively weak suggestive mark through its advertising. Sufficient evidence was presented below to demonstrate that the Bowflex mark enjoys a high level of brand recognition.

D.

We next address ICON’S challenge to the district court’s finding of similarity between the marks. Acknowledging that the meaning, appearance, and sound of the marks differed, the district court found that “Bowflex” and “CrossBow” were only “somewhat similar” marks. Preliminary Injunction Order, slip op. at 4. Because Nautilus’s and ICON’S products were identical, however, this lesser showing was considered nonetheless sufficient for a finding of likelihood of confusion.

On appeal, ICON’S first argument goes to the analysis required under Sleekcraft and its progeny for similarity of the marks at issue. Relying on Ocean Garden, Inc. v. Marktrade Co., ICON asserts that in addition to considering the Sleekcraft factors, a court must ascertain whether any resulting likelihood of confusion was directly caused by the similarity of the marks. To fulfill this requirement, ICON would have a court make an explicit “determination” finding a “causal link” between the similarity of two marks and any ensuing confusion. The failure of the district court in this case to make such a specific finding amounts to reversible error, ICON asserts.

Second, ICON argues that, setting aside the marks’ shared component, “bow,” the remainder of the marks, “flex” and “cross,” are entirely dissimilar. This distinction, along with other differences in appearance, sound, and meaning, should, in ICON’S view, operate to prevent a finding of similarity. Without similarity, argues ICON, Nautilus’s claim of trademark infringement must fail.

Finally, ICON argues that if any likelihood of confusion does exist, it stems from (i) the similarity of the machines and/or (ii) Nautilus’s longstanding dominance in this particular industry — not from any closeness of the “Crossbow” and “Bowflex” marks. In ICON’S view, confusion based predominantly on similar functionality cannot be protected through trademark law.

To the extent that ICON contends that the district court erred because it failed to perform a separate causation analysis, its position does not find support in the law. The Ninth Circuit has never held that, in addition to considering the Sleekcraft factors, a court must satisfy itself that a separate causal relationship exists between the similarity of the marks and consumer confusion. The statement made by the Ocean Garden court, upon which ICON heavily relies, does not indicate otherwise. In Ocean Garden, the court required the plaintiff to “prove probable success on the claim that there is a likelihood of confusion as to the source or origin of the goods caused by the similarity of the marks.” 953 F.2d at 506 (emphasis added). Although this does imply that “similarity of the marks” is critical to an infringement inquiry, it does not suggest that a separate finding of causation, tied solely to the similarity of the marks, is required. Instead, the matter of “similarity of the marks” is already taken into account as one of the Sleekcraft factors, and the Ocean Garden court cited with approval the use of the Sleekcraft factors. See id. at. 506 n. 2. At no point did the Ocean Garden court indicate that a causal link between mark similarity and consumer confusion must be established in the manner ICON now suggests. ICON is attempting to inject into the court’s mere choice of language unintended (and unprecedented) substance. It was not error for the court in this case to decline to do more than systematically review the eight Sleekcraft factors.

On the other hand, ICON is correct to the extent it argues that the marks must be similar for a likelihood of confusion to be found. Indeed, “[t]he core element of trademark infringement is the likelihood of confusion, i.e., whether the similarity of the marks is likely to confuse customers about the source of the products.” E. & J. Gallo Winery v. Gallo Cattle Co., 967 F.2d 1280, 1290 (9th Cir.1992); see GoTo.com, 202 F.3d at 1205 (“The first Sleekcraft factor — the similarity of the marks — has always been considered a critical question in the likelihood-of-confusion analysis.”). In general, “the greater the similarity between the two marks at issue, the greater the likelihood of confusion.” E. & J. Gallo Winery, 967 F.2d at 1289.

The similarity of two marks is assessed through a comparison of their sight, sound, and meaning. Official Airline Guides, 6 F.3d at 1392. In conducting this comparison, “[similarities weigh more heavily than differences.” Entrepreneur Media, 279 F.3d at 1144. Here, as the district court noted, “CrossBow” and “Bowflex” do not appear or sound alike, aside from their shared use of “bow.” At the same time, the meanings suggested by the marks are divergent. Thus, the district court concluded that “the marks are somewhat similar,” but not “necessarily confusing.” Preliminary Injunction Order, slip op. at 4. In other words, while similarity of the marks was not a factor heavily in Nautilus’s favor, the marks were not found to be so distinct as to be deemed dissimilar and incapable of causing confusion.

We disagree with ICON that the court’s balancing was clearly erroneous or an incorrect application of the law. Because the marks share the term “bow,” they are at least somewhat similar. ICON takes the view that because “bow” is a common term, any comparison of the marks should focus more on the distinctive “flex” and “cross” components. 8 Like the district court, we decline to compare the marks in parts, rather than in their entirety. In judging similarity, “marks must be considered in their entirety and as they appear in the marketplace.” Official Airline Guides, 6 F.3d at 1392. The way in which the district court framed its analysis was consistent with this standard.

We turn now to ICON’S final argument, that a finding of trademark infringement is not proper where likelihood of confusion is premised on a combination of identical functionality of products and only somewhat similar marks. That relatedness of goods or services is itself an “important” Sleekcraft factor, see Brookfield Communications, 174 F.3d at 1054, renders this argument somewhat dubious. In practice, these two Sleekcraft factors, similarity of the marks and relatedness of the goods and services, are often interdependent. Entrepreneur Media, 279 F.3d at 1147 (“[T]he more closely related the goods are, the more likely consumers will be confused by similar marks.”); see also Thane Int’l, Inc. v. Trek Bicycle Corp., 305 F.3d 894, 903-904 (9th Cir.2002). The parallel treatment of these factors is particularly appropriate when the mark and the product or service combine to form an impression on the consumer as to the source of the goods. For example, if a consumer encounters two related goods or services within the same market, less similarity between the marks would be required for confusion of that consumer to be likely. On the other hand, even if two marks are identical, if they are encountered in different contexts, the consumer can often easily distinguish between the two products. For these reasons, “[a] diminished standard of similarity is ... applied when comparing the marks of closely related goods.” Official Airline Guides, 6 F.3d at 1392. Application of this principle to our case, where the products are essentially identical, leads to the conclusion that the district court was free to accept a lesser showing of similarity between the marks. See Preliminary Injunction Order, slip op. at 5. At this early stage in the proceeding, we see no legal error in the court’s analysis that requires our reversal.

E.

Finally, we consider ICON’S broader challenge to the “likelihood of confusion” finding made by the district court. ICON argues that the district court did not appropriately balance the totality of the Sleekcraft factors to find likelihood of confusion.

In general, the Ninth Circuit has cautioned against use of the Sleekcraft factors as a “scorecard — whether a party ‘wins’ a majority of the factors is not the point.” Thane Int% 305 F.3d at 901. Instead, the application of this multifactor test should remain flexible and “pliant.” Interstellar Starship, 304 F.3d at 942. There is “ ‘no mechanistic formula or list [that] can set forth in advance’ the variety of elements that comprise the market context from which likelihood of confusion must be determined.” Dr. Seuss Enters., L.P. v. Penguin Books USA, Inc., 109 F.3d 1394, 1404 (9th Cir.1997) (quoting Restatement (Third) of Unfair Competition § 21 cmt. a (1995)). “Some factors are much more important than others, and the relative importance of each individual factor will be case specific.” Brookfield Communications., 174 F.3d at 1054. In general, the factors should not be “rigidly weighed; we do not count beans.” Dreamwerks, 142 F.3d at 1129. To the extent that the Ninth Circuit imposes any consistent requirement on district courts in this area, it is simply that, within the context of each specific case, they must “consider what each factor, and — more importantly — what the analysis as a whole reveals about the ultimate question before [them]: the likelihood of consumer confusion as to the origin of the product or service bearing the allegedly infringing mark.” Entrepreneur Media, 279 F.3d at 1141.

Our review at this stage in the proceeding is limited. See Ocean Garden, 953 F.2d at 502 (“A district court’s decision to grant a motion for a preliminary injunction will be upheld unless the court ‘applied the incorrect law, relied on clearly erroneous factual findings, or otherwise abused its discretion.’ ” (internal citation omitted)). Setting aside the intent to confuse and evidence of actual confusion factors, there remains sufficient evidence, under the Sleekcraft factors, to justify the preliminary injunction in Nautilus’s favor. The degree of similarity of the marks, proximity of the products and services, strength of the Bowflex mark, and similarity of marketing channels are all factors that the court found to favor Nautilus. We have sustained the district court’s finding with respect to the strength of the Bowflex mark. At the same time, ICON does not challenge the court’s conclusions with respect to the similarity of the marketing channels. On similarity of the marks, we agree with ICON that the marks are not that similar, but we recognize, as the district court did, that this weakness is to some extent compensated for by the identity of the products and the strength of Nautilus’s mark. Accordingly, we are not prepared to say that the district court abused its discretion in finding likelihood of confusion. See Vision Sports, 888 F.2d at 616 (concluding that even if some Sleek-craft factors are successfully challenged, a preliminary injunction “must be upheld on the basis of other factors” if those other' factors are on their own sufficient to support a finding of likelihood of confusion).

CONCLUSION

For the foregoing reasons, the decision of the district court granting a preliminary injunction in favor of Nautilus is

AFFIRMED.

COSTS

Each party shall bear its own costs.

Notes

1

. U.S. Patent No. 4,620,704, "Universal Exercising Machine,” and its continuation, U.S. Patent No. 4,725,057, of the same name.

2

. The original trademark registration received by Nautilus is for the term "BOW-FLEX,” but Nautilus has since procured additional trademarks on this term including "Bowflex,” "BowFlex,” and "BOW FLEX.”

3

. In actuality, ICON, not Weider, is the manufacturer and source of the CrossBow machine. It licenses the "Weider” mark, which has been in use in the exercise equipment industry since 1936.

4

. In a non-precedential opinion, this court reversed the judgment of non-infringement after concluding that it was based on an erroneous claim construction. The case was remanded to the district court for a determination of infringement under the correct claim construction.

5

. As noted above, the preliminary injunction has been stayed pending this appeal.

6

. We think it unlikely that any of the four calls submitted by Nautilus reflect evidence of actual confusion resulting from the alleged closeness of the marks. In each case, the customer was either confused by the similar machinery, or indicated that he or she was aware of the distinctions between the companies.

In Exhibit A, Call No. 18, for example, it is obvious that the consumer’s confusion is rooted in the similarity of the equipment mechanism; in fact, he does not even know the name of ICON'S product until prompted by the Nautilus Representative:

[Nautilus Representative ("NR”) ]: The Bowflex is the one machine on TV with that leg attachment. It’s going to be $34.00 a month. The total price on that would be 999 plus 129 shipping and handling.
[Caller ("C”) ]: I saw the same thing a year ago about 499.
[NR]: I wouldn’t say the same thing.
[C]: Well, it's Bowflex 8 240 pounds and everything just like this one.
[NR]: It's not the Bowflex, though ... it's the crossbow.
[C]: Well that’s the better one than the Bowflex, right?
[NR]: No.
[C]: Sure looks like it. All right, thanks.
[NR]: Whatever.

A similar conversation occurred with the customer in Call No. 49, Exhibit D. In that call, the customer thought that she had seen the BowFlex machine at Sears; evidently, Sears sells only the CrossBow exerciser. After the Nautilus representative pointed this out to her, she stated, "Oh that makes sense.” The Nautilus representative proceeded to make clear this distinction by explaining the differences between the structure and operation of CrossBow and Bowflex. As in Exhibit A, at no point in the ensuing conversation did the customer or the representative suggest that the confusion stemmed from anything but the similar appearance and machinery of the exercise systems.

Exhibit B also does not suggest confusion over the marks. In Call No. 23, the Nautilus representative placed a follow-up sales call to a customer who had received some information on Bowflex. The customer demonstrated that she was aware of the difference between CrossBow and Bowflex by explaining that she had already ordered CrossBow. Before the call was terminated, the customer casually asked the Nautilus representative whether the company was "connected with the Cross Bow.” When the Nautilus representative replied in the negative, the customer stated, "Oh, I thought it was the same maker!,] Wider or Weider,” and admitted that she did not know these were two different companies. This admission shows only that the customer was confused by the source designation and the relationship of the companies, not by the closeness of the "Bowflex” and "CrossBow” marks.

The same is true of Call No. 24, Exhibit C. The customer called in reference to a CrossBow offer:

[C]: I almost bought from Sears an advantage you know that model?
[NR]: No, unfortunately I don't, sir.
[C]: How come you don't know? Do you work for Bowflex or do you work for them? It’s a Bowflex product.... Sears must be sanctioned because you guys allowed them to sell it.

Midway through the call, the customer independently realized he had erroneously called Bowflex instead of CrossBow, and explained, "Oh, I’m sorry. You know what it is I thought this was Cross Bow. I'm sorry.... Cross Bow’s the one that I'm pissed off at.... I’m sorry, I'm sorry to bother you. I’ll call the wrong people [sic].” This call also does not prove actual confusion because this customer knew of and understood the distinction between the companies and their respective marks.

7

. Secondary meaning can be established through "direct consumer testimony; survey evidence; exclusivity, manner, and length of use of a mark; amount and manner of advertising; amount of sales and number of customers; established place in the market; and proof of intentional copying by the defendant.” Filipino Yellow Pages, Inc. v. Asian Journal Publs., Inc., 198 F.3d 1143, 1151 (9th Cir.1999); see also Accuride Int’l, Inc. v. Accuride Corp., 871 F.2d 1531, 1536 (9th Cir.1989) ("A descriptive or suggestive mark may be strengthened by such factors as extensive advertising, length of exclusive use, public recognition, and uniqueness.”).

8

. ICON’S position is that because "bow” is either generic, or descriptive without secondary meaning, for these types of exercise machines, it must be freely available for use by Nautilus's competitors to describe their machines. For this reason, ICON argues that any analysis of similarity must not rely on that shared term. As explained above, we do not accept ICON'S argument that, in the context of this case, "bow” is generic or descriptive; we therefore conclude that the import of "bow” cannot be discounted in a comparison of the two marks.