OPINION

Marketquest Group, Inc. (Marketquest) appeals the district court’s grant of summary judgment in favor of Norwood Promotional Products, LLC (Norwood), BIC Corp., and BIC USA, Inc. (collectively, Defendants) on Marketquest’s trademark infringement claims. The district court held that Defendants’ uses of Market-quest’s trademarks “All-in-One” and “The Write Choice” were completely protected by the fair use defense. We reverse and remand.

FACTS AND PRIOR PROCEEDINGS

Marketquest produces and sells promotional products, and has used its United States Patent and Trademark Office registered trademarks “All-in-One” and “The Write Choice” since 1999 and 2000, respectively. In 2009, BIC Corp. and BIC USA, Inc. (collectively, BIC) acquired Norwood, a promotional products company, and in 2010 Norwood published a promotional products catalogue for 2011 that featured the phrase “All-in-One” on the cover of and in the catalogue. The 2011 catalogue consolidated all of Norwood’s eight “hard goods” catalogues “in one” catalogue, whereas they were previously published in separate catalogues. In 2010, BIC also used the phrase “The WRITE Pen Choice for 30 Years” in advertising and packaging for its pens, in connection with its thirtieth anniversary promotion.

Marketquest filed the operative First Amended Complaint (FAC) against Defendants on May 5, 2011, alleging infringement of Marketquest’s “All-in-One” and “The Write Choice” trademarks. On August 26, 2011, Marketquest moved for a preliminary injunction. The arguments and evidence submitted by Marketquest in support of its motion pertained only to Defendants’ use of “All-in-One,” and not “The Write Choice,” so the district court deemed Marketquest’s request for a preliminary injunction as to Defendants’ use of “The Write Choice” waived. The district court denied the requested injunction regarding Defendants’ use of “All-in-One” after concluding that Marketquest was unlikely to succeed on the merits because Defendants were likely to succeed in asserting a fair use defense.

Discovery proceeded and the parties filed cross-motions for summary judgment. The district court granted summary judgment for Defendants, holding that there was “some likelihood of confusion and therefore the potential for trademark infringement liability,” but that further analysis of likelihood of confusion was unnecessary because fair use provided Defendants a complete defense to allegations of infringement of both the “All-in-One” and “The Write Choice” trademarks. Market-quest timely appealed.

STANDARD OF REVIEW

We review the district court’s grant of summary judgment de novo. KP Permanent Make-Up, Inc. v. Lasting Impression I, Inc., 408 F.3d 596, 602 (9th Cir. 2005) (KP Permanent II). We view the evidence in the light most favorable to Marketquest and determine “whether there are any genuine issues of material fact and whether the district court correctly applied the relevant substantive law.” Id. We are mindful that “summary judgment is generally disfavored in the trademark arena” due to “the intensely factual nature of trademark disputes.” Id. (quoting Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1140 (9th Cir. 2002)).

ANALYSIS

I. Marketquest’s pleading was adequate to support a cause of action for trademark infringement under a reverse confusion theory of likely confusion.

The Lanham Act provides a cause of action for the owner of a registered trademark against any person who, without consent of the owner, uses the trademark in commerce in connection with the sale or advertising of goods or services, when such use is likely to cause confusion. 15 U.S.C. § 1114 (1). The validity of Marketquest’s trademarks is not disputed in this appeal. Thus, the question is whether there is a likelihood of confusion; that is, whether Defendants’ “actual practice^ were] likely to produce confusion in the minds of consumers about the origin of the goods ... in question.” KP Permanent Make-Up, Inc. v. Lasting Impression I, Inc., 543 U.S. 111, 117 , 125 S.Ct. 542 , 160 L.Ed.2d 440 (2004) (KP Permanent I).

We have recognized two theories of consumer confusion that support a claim of trademark infringement: forward confusion and reverse confusion. Surfvivor Media, Inc. v. Survivor Prods., 406 F.3d 625, 630 (9th Cir. 2005). “Forward confusion occurs when consumers believe that goods bearing the junior mark came from, or were sponsored by, the senior mark holder.” Id. For example, consumers would experience forward confusion if they believed that Defendants’ 2011 cata-logue came from Marketquest because it featured the phrase “AU-in-One.” “By contrast, reverse confusion occurs when consumers dealing with the senior mark holder believe that they are doing business with the junior one.” Id. For example, consumers would experience reverse confusion if they did business with Market-quest, but believed that they were doing business with Defendants, because they had come to associate the words “All-in-One” with Defendants.

Marketquest argues that this is a “reverse confusion case,” while Defendants counter that Marketquest did not adequately plead reverse confusion. Our circuit has not previously addressed the pleading standard required to state a cause of action for trademark infringement under a reverse confusion theory. We now hold that reverse confusion is not a separate claim that must be specifically pleaded, but instead is a theory of likely confusion that may be alleged by itself or in addition to forward confusion. Accord Dorpan, S.L. v. Hotel Melia, Inc., 728 F.3d 55 , 65 n.12 (1st Cir. 2013) (“‘Reverse confusion’ is not a separate legal claim requiring separate pleading. Rather, it is a descriptive term referring to certain circumstances that can give rise to a likelihood of confusion.”). Thus, when reverse confusion is compatible with the theory of infringement alleged in the complaint, a plaintiff need not specifically plead it.

Defendants cite Surfvivor, 406 F.3d at 631-34 , and Murray v. Cable National Broadcasting Co., 86 F.3d 858, 861 (9th Cir. 1996), to support their contention that strict pleading is required when someone sues for reverse confusion, but these cases are distinguishable. In Surfvi-vor we held that the plaintiff raised no cognizable forward confusion claim because it failed to reference forward confusion in its complaint. 406 F.3d at 631 . Instead, the plaintiff only raised a cognizable trademark infringement claim under a reverse confusion theory. Id. Only reverse confusion was plausible in that case; Surfvivor, a maker of beach-themed products sold only in Hawaii, alleged that consumers thought that its products were sponsored by the well-known, national reality show Survivor. Id. at 629 . Facts indicating forward confusion (i.e., that consumers would think the well-known, national show Survivor came from the small, Hawaiian Surfvivor brand) were not alleged and would be highly unlikely. Surfvivor therefore stands for the proposition that when (1) a plaintiffs trademark infringement claim is based on a reverse confusion theory, (2) the plaintiff did not plead examples of forward confusion, and (3) only reverse confusion is plausible based on the allegations in the complaint, a plaintiff does not allege a cognizable trademark infringement claim based on forward confusion.

Defendants cite Murray for the proposition that to plead reverse confusion, a plaintiff must allege that the defendant “saturated the market with advertising,” or allege actual reverse confusion from customers. See 86 F.3d at 861 . However, Murray was decided before we recognized reverse confusion as a theory of trademark infringement, and we concluded that such recognition was unnecessary because Murray did not allege any cognizable trademark infringement claim, whether under what we now refer to as reverse or forward confusion theories. Id. Thus, Murray did not set out a specific pleading standard for reverse confusion; there was no likelihood of any type of confusion, and we merely listed shortcomings in Murray’s allegations. To advance a reverse confusion theory, a plaintiff may allege that the defendant flooded the market with advertising, or that actual instances of reverse confusion occurred. However, the allegations will vary in individual cases.

Applying these principles to Mark-etquest’s FAC, we first note that Market-quest did not use the words “reverse confusion,” allege that Defendants saturated the market, or allege instances of actual reverse confusion. The FAC generally alleged that customers were confused “as to whether some affiliation, connection, or association exist[ed]” among Defendants and Marketquest, and specifically alleged that there were actual instances of forward confusion (i.e., that consumers thought that Defendants’ goods came from Market-quest). Marketquest did not allege instances of actual reverse confusion until its motion for a preliminary injunction. The district court’s orders denying the preliminary injunction and granting summary judgment recognized Marketquest’s reverse confusion theory.

We hold that reverse confusion is a theory of confusion available to Marketquest, and it did not foreclose this theory by failing to plead it with particularity in its FAC. This case is distinguishable from Surfvivor because, in that case, the theory of confusion not specifically pleaded in the complaint was implausible. See 406 F.3d at 629 . Here, BIC is clearly the larger, more widely-known entity, and has acquired some smaller promotional products companies, such as Norwood. It is therefore plausible that consumers would associate Marketquest’s marks with Defendants, or think that BIC had acquired Marketquest. Marketquest’s general allegation that there was confusion “as to whether some affiliation, connection,- or association existed]” among the parties fairly encompasses this possibility, and Marketquest should not have been prevented from ex ploring possible instances of reverse confusion during discovery. Thus, the district court was correct to consider the possibility of reverse confusion as the case proceeded.

II. Consideration of the intent factor in the likelihood of confusion analysis varies with the type of confusion being considered.

In trademark infringement cases we assess likelihood of consumer confusion by considering the Sleekcraft factors. Fortune Dynamic, Inc. v. Victoria’s Secret Stores Brand Mgmt., Inc., 618 F.3d 1025, 1030 (9th Cir. 2010) (citing AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 348-49 (9th Cir. 1979)). The Sleekcraft “analysis is pliant, illustrative rather than exhaustive, and best understood as simply providing helpful guideposts.” Id. (internal quotation marks omitted). “[T]he relative importance of each individual factor will be case-specific.” Id. at 1031 (quoting Brookfield Commc’ns, Inc. v. W. Coast Entm’t Corp., 174 F.3d 1036, 1054 (9th Cir. 1999)).

Our assessment of the Sleekcraft intent factor (“the defendant’s intent in selecting the mark”) is different when we consider a forward confusion theory than it is when we consider a reverse confusion theory, because the relevance of intent varies with the underlying theory of confusion. See Cohn v. Petsmart, Inc., 281 F.3d 837 , 841 n.5 (9th Cir. 2002); see also Commerce Nat’l Ins. Servs., Inc. v. Commerce Ins. Agency, Inc., 214 F.3d 432, 444 (3d Cir. 2000). When considering forward confusion, we ask “whether defendant in adopting its mark intended to capitalize on plaintiffs good will.” Fortune Dynamic, 618 F.3d at 1043 . However, in the case of reverse confusion, typically “neither junior nor senior user wishes to siphon off the other’s goodwill.” Dreamwerks Prod. Grp., Inc. v. SKG Studio, 142 F.3d 1127, 1130 (9th Cir. 1998).

Marketquest argues that mere knowledge of the senior user’s mark establishes intent to confuse when considering reverse confusion, while Defendants argue that only evidence of a deliberate intent to push the plaintiff out of the market should suffice. However, no one type of evidence is required to establish intent in trademark infringement cases under either a forward or reverse theory of confusion. Indeed, “an intent to confuse consumers is not required for a finding of trademark infringement.” Brookfield Commc’ns, 174 F.3d at 1059 . The Sleekcraft factors are to be applied flexibly, knowing that “not all of the factors are of equal importance or applicable in every case.” KP Permanent II, 408 F.3d at 608 .

Thus, when a court applies Sleekcraft in a case that presents reverse confusion, and the intent factor is relevant, it may consider several indicia of intent. At one extreme, intent could be shown through evidence that a defendant deliberately intended to push the plaintiff out of the market by flooding the market with advertising to create reverse confusion. See Freedom Card, Inc. v. JPMorgan Chase & Co., 432 F.3d 463, 479 (3d Cir. 2005). Intent could also be shown by evidence that, for example, the defendant knew of the mark, should have known of the mark, intended to copy the plaintiff, failed to conduct a reasonably adequate trademark search, or otherwise culpably disregarded the risk of reverse confusion. See, e.g., Surfvivor, 406 F.3d at 634 ; Cohn, 281 F.3d at 843 . The tenor of the intent inquiry shifts when considering reverse confusion due to the shift in the theory of confusion, but no specific type of evidence is necessary to establish intent, and the importance of intent and evidence presented will vary by case.

III. The district court erred by granting summary judgment in favor of Defendants based upon on the fair use defense regarding their use of “All-in-One.”

Applying the “classic fair use” defense, “[a] junior user is always entitled to use a descriptive term in good faith in its primary, descriptive sense other than as a trademark.” Cairns v. Franklin Mint Co., 292 F.3d 1139, 1150 (9th Cir. 2002). A defendant must show that its use is (1) other than as a trademark, (2) descriptive of the defendant’s goods, and (3) in good faith. 15 U.S.C. § 1115 (b)(4). Additionally, “the degree of customer confusion [is] a factor in evaluating fair use.” KP Permanent II, 408 F.3d at 609 .

The district court considered the elements of the fair use defense, and concluded that Defendants’ use of “All-in-One” in connection with the 2011 catalogue was completely protected by the fair use defense. As discussed below, genuine issues of material fact exist regarding the elements of fair use in this case, thereby precluding summary judgment. While summary judgment on the fair use defense in a trademark case is possible, we reiterate that “summary judgment is generally disfavored” in trademark cases, due to “the intensely factual nature of trademark disputes.” Fortune Dynamic, 618 F.3d at 1031 .

A. Non-catalogue uses of All-in-One

Marketquest first argues that the district court erred by not specifically analyzing all of the uses of “All-in-One” employed by Defendants, since the fair use analysis often varies when a defendant uses the same mark in different ways. The “other uses” of “All-in-One” included (1) promotional materials that featured an image of the 2011 catalogue; (2) promotional materials that directed customers to look for products or information in “the 2011 Norwood All in ONE catalogue”; and (3) an online advertisement that said “Put Your Drinkware Needs ... in a Norwood ALL in ONE Basket,” which included a photo of a basket containing several different types of drinkware. Defendants respond that there was no need to conduct a design-by-design review because all of these uses connected to the 2011 cata-logue, and there is no basis for the claim that the district court did not consider all the evidence, even if other uses were not specifically referenced by the district court.

It appears from its summary judgment order that the district court focused on Defendants’ use of “All-in-One” on the 2011 catalogue, and perhaps did not consider other uses. While a design-by-design review of promotional materials that merely included a picture of the 2011 catalogue was not required, references to “the 2011 Norwood All in ONE catalogue” and “a Norwood ALL in ONE Basket” are sufficiently distinct to require analysis for fair use. These uses are considered below, along with the 2011 catalogue use.

B. Use other than as a trademark.

A fair use must be a use other than as a trademark. 15 U.S.C. § 1115 (b)(4). A trademark is used “to identify and distinguish ... goods ... from those manufactured or sold by others and to indicate the source of the goods.” Id. § 1127. “To determine whether a term is being used as a mark, we look for indications that the term is being used to associate it with a manufacturer,” and “whether the term is used as a symbol to attract public attention.” Fortune Dynamic, 618 F.3d at 1040 (internal quotation marks omitted). We also consider “whether the allegedly infringing user undertook precautionary measures ... to minimize the risk that the term will be understood in its trademark sense.” Id. (internal quotation marks omitted).

A genuine issue of fact exists regarding whether Defendants used “All-in-One” as a trademark. Defendants did take “precautionary measures” when featuring “All-in-One” on the 2011 catalogue: Nor-wood was printed at the top in large, bold, capital letters with a trademark symbol, while “All-in-One” was located further down on the page, in smaller letters, without a trademark symbol, and positioned as a heading over a list of all the products consolidated “in one” catalogue. This suggests that Norwood was used to indicate the source of the goods, rather than “All-in-One” (although it is possible for more than one trademark to appear on a cata-logue cover). However, when considering all of Defendants’ uses of “All-in-One,” a jury could potentially find trademark use. The “precautionary measures” listed above were not present when Defendants referred to “the 2011 Norwood All in ONE catalogue” and “a Norwood ALL in ONE Basket.” In these uses, there is no obvious distinction between Norwood and “All-in-One,” and both could reasonably be understood to indicate source.

C. Descriptive use.

To prevail on fair use, a defendant must show that it used the mark “in its primary, descriptive sense.” Fortune Dynamic, 618 F.3d at 1041 (quoting Brother Records, Inc. v. Jardine, 318 F.3d 900, 906 (9th Cir. 2003) (alteration omitted)); see 15 U.S.C. § 1115 (b)(4). While “we accept some flexibility in what counts as descriptive,” Fortune Dynamic, 618 F.3d at 1042 , “the scope of the fair use defense varies with ... the descriptive purity of the defendant’s use and whether there are other words available to do the describing.” Id. at 1041 . Even when “there [is] some evidence of descriptive use, [a jury] could still reasonably conclude that [a defendant’s] lack of ‘precautionary measures’ ” outweighs such evidence. Id. at 1042 .

There is a strong argument that Defendants’ use of “All-in-One” on the 2011 catalogue was descriptive, because it was used as a heading for a list of the products consolidated “all in one” cata-logue. Moreover, as discussed above, Defendants took “precautionary measures” on their catalogue cover by using a design that indicated descriptive use. However, Defendants’ other uses of “All-in-One” were arguably not descriptive, and “precautionary measures” were not taken with these uses. While Defendants’ use of “All-in-One” as a heading on the 2011 catalogue strongly indicates descriptive use, such use is not apparent in decontextualized references to “the 2011 Norwood All in ONE catalogue.” Additionally, “Put Your Drink-ware Needs ... in a Norwood ALL in ONE Basket” does not fall under the descriptive use explanation that Defendants advance because it does not refer to a consolidated catalogue. It may descriptively refer to consolidating drinkware in a basket, but the “descriptive purity” of such use is questionable because it is unclear if the basket is literal or suggestive. See id. at 1041 . Uses of “All-in-One” in ways that stripped it of its possible descriptive meaning undermine Defendants’ descriptive use argument, such that a finder of fact could determine that the use was not descriptive. Moreover, a finder of fact could determine that Defendants “had at [their] disposal a number of alternative words [or phrases] that could adequately capture [their] goal,” limiting the scope of the fair use defense in this case. Id. at 1042 .

D. In good faith.

A defendant asserting fair use must also show that it used the mark in good faith. 15 U.S.C. § 1115 (b)(4). When considering forward confusion, this element “involves the same issue as the intent factor in the likelihood of confusion analysis”; that is, “whether defendant in adopting its mark intended to capitalize on plaintiffs good will.” Fortune Dynamic, 618 F.3d at 1043 . The shift in focus discussed in Part II above for assessing intent when considering likelihood of confusion under a reverse confusion theory generally applies when considering good faith as part of the fair use defense in a case that presents reverse confusion. However, the good faith inquiry differs somewhat from the Sleekcraft intent factor, regardless of the underlying theory of confusion. In fair use, good faith is an element of the defense, not merely a factor to consider when it is relevant in a given case.

As with intent in Sleekcraft, there is no bright-line rule or required piece of evidence to establish good or bad faith. While the focus may differ when considering forward or reverse confusion, generally the same types of evidence will be relevant to this inquiry. This includes evidence such as whether the defendant intended to create confusion, whether forward or reverse; intended to push the plaintiff out of the market; remained ignorant of the plaintiffs mark when it reasonably should have known of the mark; knew of the mark and showed bad faith in its disregard of the plaintiffs rights; or any other evidence relevant to whether the defendant’s claimed “objectively fair” use of the mark was done in good faith. See KP Permanent I, 543 U.S. at 123 , 125 S.Ct. 542 ; see also, e.g., Fortune Dynamic, 618 F.3d at 1043 (holding that a material question of fact existed regarding defendant’s good faith when the plaintiff introduced evidence that the defendant carelessly failed to investigate the trademark at issue, along with expert testimony that a trademark search would have been standard practice in the relevant industry).

Marketquest argues that because this case presents reverse confusion, mere knowledge of Marketquest’s ownership and use of the “All-in-One” mark establishes bad faith on the part of Defendants, and fair use is thus unavailable as a matter of law. That is incorrect. An inference of bad faith does not arise from mere knowledge of a mark when the use is otherwise objectively fair, even in a case presenting reverse confusion. Marketquest also argues that Defendants’ use of two of its marks in the same year supports an inference of bad faith. This fact by itself, coupled with Marketquest’s knowledge of the marks, is thin evidence of bad faith. However, we cannot say on summary judgment that no reasonable finder of fact could infer bad faith from these facts.

E. Degree of consumer confusion.

“The fair use defense only comes into play once the party alleging infringement has shown by a preponderance of the evidence that confusion is likely.” KP Permanent II, 408 F.3d at 608-09 . This is because if there is no likelihood of consumer confusion, then there is no trademark infringement, making an affirmative defense to trademark infringement irrelevant. KP Permanent I , 543 U.S. at 120, 125 S.Ct. 542 . After the plaintiff meets the threshold showing, in the fair use analysis “the degree of customer confusion [is] a factor” to consider. KP Permanent II, 408 F.3d at 609 . However, a defendant raising the defense does not have the burden to negate any likelihood of consumer confusion. KP Permanent I, 543 U.S. at 114 , 125 S.Ct. 542 . Some consumer confusion is compatible with fair use, and when a plaintiff chooses “to identify its product with a mark that uses a well known descriptive phrase” it assumes the risk of some confusion. Id. at 121-22 , 125 S.Ct. 542 .

The district court held that Marketquest met the threshold requirement for fair use by showing that there is some likelihood of confusion, relying upon its previous Sleek-craft analysis in the order denying Mark-etquest’s motion for a preliminary injunction. However, the district court held that any further Sleekcraft analysis was “unnecessary” because fair use provided Defendants a complete defense.

Marketquest argues that the district court’s holding was incomplete because it did not conduct a full Sleekcraft analysis, nor did it consider the additional factors that we stated in KP Permanent II would be relevant,to the jury’s consideration of fair use in that case. Defendants counter that a court may grant summary judgment on the fair use defense without deciding the likelihood of confusion because confusion is not the focus of fair use; the focus is objective fairness,‘and some confusion is accepted.

We emphasize that the degree of consumer confusion is a factor in the fair use analysis, not an element of fair use. See KP Permanent I, 543 U.S. at 118 , 125 S.Ct. 542 (“Congress said nothing about likelihood of confusion in setting out the elements of the fair use defense.”). This factor is useful in considering whether, overall, the use was objectively fair. A use that is likely to confuse consumers, or that has caused actual confusion, is less likely to be objectively fair (although some confusion is permissible). Accord Shakespeare Co. v. Silstar Corp. of Am., 110 F.3d 234, 243 (4th Cir. 1997) (“While it is true that to the degree that confusion is likely, a use is less likely to be found fair, it does not follow that a determination of likely confusion precludes considering the fairness of use.”). The Sleekcraft factors and additional factors that we identified as relevant in KP Permanent II may also be relevant in a given case where fair use is at issue. A court is not required in every case to recite and analyze all the factors identified in Sleekcraft and KP Permanent II one-by-one for a fair use analysis to be complete. A court may do so, but these are merely factors to facilitate a court’s analysis, to the degree they are relevant in a given case.

In this case, the district court referenced its previous Sleekcraft analysis at the preliminary injunction phase. The district court was not required to conduct this analysis again and determine all potential issues of fact as a matter of law before considering summary judgment on fair use. However, because we reverse summary judgment on fair use for the reasons indicated above, we leave it to the district court to determine on remand the relevance of the degree of consumer confusion in this case.

IV. The district court erred by applying the fair use analysis to Defendants’ use of “The Write Choice” after determining that Market-quest presented no evidence of likely confusion.

The district court found that there was “no evidence of actual or potential confusion” resulting from Defendants’ use of “The Write Choice,” and then concluded that Defendants had shown fair use. That fair use analysis was in error because “[t]he fair use defense only comes into play once the party alleging infringement has shown by a preponderance of the evidence that confusion is likely.” KP Permanent II, 408 F.3d at 608-09 . The district court could not properly find here that there was no evidence of confusion, fail to conduct a Sleekcraft analysis, and still conclude that the Defendants qualified for the fair use defense. Thus, we remand for the district court to consider Marketquest’s trademark infringement claim regarding Defendants’ use of “The Write Choice.”

CONCLUSION

We REVERSE and REMAND for proceedings consistent with this opinion.

. "Those factors are: 1) the strength of the mark; 2) proximity or relatedness of the goods; 3) the similarity of the marks; 4) evidence of actual confusion; 5) the marketing . channels used; 6) the degree of care customers are likely to exercise in purchasing the goods; 7) the defendant's intent in selecting the mark; and 8) the likelihood of expansion into other markets.” KP Permanent II, 408 F.3d at 608 . . These factors included “the degree of likely confusion, the strength of the trademark, the descriptive nature of the term for the product or service being offered by [the defendant] and the availability of alternate descriptive terms, the extent of the use of the term prior to the registration of the trademark, and any differences among the times and contexts in which [the defendant] has used the term.” KP Permanent II, 408 F.3d at 609 .