Playboy Enterprises International, Inc. (PEI) appeals from the district court’s grant of summary judgment in favor of Netscape Communications Corporation and Excite, Inc. PEI sued defendants for trademark infringement and dilution. We have jurisdiction pursuant to 28 U.S.C. § 1291 . Because we conclude that genuine issues of material fact preclude summary judgment on both the trademark infringement and dilution claims, we reverse and remand.

I. FACTS

This case involves a practice called “keying” that defendants use on their Internet search engines. Keying allows advertisers to target individuals with certain interests by linking advertisements to pre-identified terms. To take an innocuous example, a person who searches for a term related to *1023 gardening may be a likely customer for a company selling seeds. Thus, a seed company might pay to have its advertisement displayed when searchers enter terms related to gardening. After paying a fee to defendants, that company could have its advertisements appear on the page listing the search results for gardening-related terms: the ad would be “keyed” to gardening-related terms. Advertisements appearing on search result pages are called “banner ads” because they run along the top or side of a page much like a banner. 1

Defendants have various lists of terms to which they key advertisers’ banner ads. Those lists include the one at issue in this case, a list containing terms related to sex and adult-oriented entertainment. Among the over-400 terms in this list are two for which PEI holds trademarks: “playboy” and “playmate.” 2 Defendants require adult-oriented companies to link their ads to this set of words. Thus, when a user types in “playboy,” “playmate,” or one of the other listed terms, those companies’ banner ads appear on the search results page. 3

PEI introduced evidence that the adult-oriented banner ads displayed on defendants’ search results pages are often graphic in nature and are confusingly labeled or not labeled at all. In addition, the parties do not dispute that buttons on the banner ads say “click here.” When a searcher complies, the search results page disappears, and the searcher finds him or herself at the advertiser’s website. PEI presented uncontroverted evidence that defendants monitor “click rates,” the ratio between the number of times searchers click on banner ads and the number of times the ads are shown. Defendants use click rate statistics to convince advertisers to renew their keyword contracts. The higher the click rate, the more successful they deem a banner ad.

PEI sued defendants, asserting that they were using PEI’s marks in a manner that infringed upon and diluted them. The district court denied PEI’s request for a preliminary injunction, and this court affirmed in an unpublished disposition. 4 On remand, the parties filed cross-motions for summary judgment. The district court granted summary judgment in favor of defendants. We reverse.

II. STANDARD OF REVIEW

We review the district court’s grant of summary judgment de novo. 5 Viewing the evidence in the light most favorable to PEI, and drawing all reasonable inferences in PEI’s favor, we must determine whether there are any genuine issues of material fact and whether the district court correctly applied the relevant substantive law. 6 The moving party — in this case, the defendants — bears the “initial *1024 burden of identifying for the court the portions of the materials on file that it believes demonstrate the absence of any genuine issue of material fact.” 7 If the moving party meets its initial burden, the burden shifts to the non-moving party to “set forth, by affidavit or as otherwise provided by Rule 56, 'specific facts showing that there is a genuine issue for trial.’ ” 8 We may not weigh the evidence or determine the truth of the matter but may only determine whether there is a genuine issue for trial. 9

III. DISCUSSION

A. Trademark Infringement

With regard to PEI’s trademark infringement claim, the parties disagree on three points. First, the parties dispute whether a direct or a contributory theory of liability applies to defendants’ actions. We conclude that defendants are potentially liable under one theory and that we need not decide which one. Second, the parties disagree regarding whether PEI has successfully shown that a genuine issue of material fact exists regarding the likelihood of consumer confusion resulting from defendants’ use of PEI’s marks. We conclude that a genuine issue of material fact does exist. Finally, the parties dispute whether any affirmative defenses apply. We conclude that no defenses apply. We will address each dispute in turn.

1. Theory of liability.

Whether the defendants are directly or merely contributorily liable proves to be a tricky question. However, we need not decide that question here. We conclude that defendants are either directly or con-tributorily liable. Under either theory, PEI’s case may proceed. Thus, we need not decide this issue.

2. PEI’s case for trademark infringement.

The “core element of trademark infringement,” the likelihood of confusion, lies at the center of this case. 10 No dispute exists regarding the other requirements set forth by the statute: PEI clearly holds the marks in question and defendants used the marks in commerce 11 without PEI’s permission. 12

PEI’s strongest argument for a likelihood of confusion is for a certain kind of confusion: initial interest confusion. 13 *1025 Initial interest confusion is customer confusion that creates initial interest in a competitor’s product. 14 Although dispelled before an actual sale occurs, initial interest confusion impermissibly capitalizes on the goodwill associated with a mark and is therefore actionable trademark infringement. 15

PEI asserts that, by keying adult-oriented advertisements to PEI’s trademarks, defendants actively create initial interest confusion in the following manner. Because banner advertisements appear immediately after users type in PEI’s marks, PEI asserts that users are likely to be confused regarding the sponsorship of unlabeled banner advertisements. 16 In addition, many of the advertisements instruct users to “click here.” Because of their confusion, users may follow the instruction, believing they will be connected to a PEI cite. Even if they realize “immediately upon accessing” the competitor’s site that they have reached a site “wholly unrelated to” PEI’s, the damage has been done: Through initial consumer confusion, the competitor “will still have gained a customer by appropriating the goodwill that [PEI] has developed in its [ ] mark.” 17

PEI’s theory strongly resembles the theory adopted by this court in Brookfield Communications, Inc. v. West Coast Entertainment Corporation, 18 In Brookfield, a video rental company, West Coast Entertainment Corporation, planned on using “moviebuff.com” as a domain name for its website and using a similar term in the metatags for the site. 19 Brookfield had trademarked the term “MovieBuff,” however, and sued West Coast for trademark infringement. 20 The court ruled in favor of Brookfield. It reasoned that Internet users entering Brookfield’s mark (plus “.com”) or searching for Brookfield’s mark on search engines using metatags, would find themselves at West Coast’s website. Although they might “realize, immediately upon accessing ‘moviebuff.com,’ that they have reached a site operated by West Coast and wholly unrelated to Brookfield,” some customers who were originally seeking Brookfield’s website “may be perfectly content with West Coast’s database (especially as it is offered free of charge).” 21 Because those customers would have found West Coast’s site due to West Coast’s “misappropriation of Brookfield’s goodwill” in its mark, the court concluded that Brookfield withstood summary judgment. 22

In this case, PEI claims that defendants, in conjunction with advertisers, have misappropriated the goodwill of PEI’s marks by leading Internet users to competitors’ websites just as West Coast video misappropriated the goodwill of Brookfield’s mark. Some consumers, initially seeking PEI’s sites, may initially believe that unlabeled banner advertisements are links to PEI’s sites or to sites affiliated with PEI. Once they follow the instructions to “click here,” and they access the site, they may well realize that they are not at a PEI-sponsored site. However, they may be perfectly happy to remain on the competitor’s site, just as the Brookfield court surmised that some searchers initial *1026 ly seeking Brookfield’s site would happily remain on West Coast’s site. The Internet user will have reached the site because of defendants’ use of PEI’s mark. Such use is actionable. 23

Although analogies to Brookfield suggest that PEI will be able to show a likelihood of confusion sufficient to defeat summary judgment, we must test PEI’s theory using this circuit’s well-established eight-factor test for the likelihood of confusion to be certain. Accordingly, we turn to that test now.

The Ninth Circuit employs an eight-factor test, originally set forth in AMF Inc. v. Sleekcraft Boats, 24 to determine the likelihood of confusion. The eight factors are:

In the Internet context, courts must be flexible in applying the factors, as some may not apply. 26 Moreover, some factors are more important than others. For example, a showing of actual confusion among significant numbers of consumers provides strong support for the likelihood of confusion. 27 For that reason, we turn first to an examination of factor four: evidence of actual confusion.

a. Factor I: Evidence of Actual Confusion.

The expert study PEI introduced establishes a strong likelihood of initial interest confusion among consumers. Thus, factor four alone probably suffices to reverse the grant of summary judgment.

PEI’s expert, Dr. Ford, concluded that a statistically significant number of Internet users searching for the terms “playboy” and “playmate” would think that PEI, oí-an affiliate, sponsored banner ads containing adult content that appear on the search results page. When study participants were shown search results for the term “playboy,” 51% believed that PEI sponsored or was otherwise associated with the adult-content banner ad displayed. 28 When shown results for the term “playmate,” 31% held the same belief. Using control groups, Dr. Ford also concluded that for 29% of those participants viewing “playboy” searches and 22% of those viewing “playmate” searches, the confusion stemmed from the targeting of the banner advertisements. The individuals were not confused by random, un-tar-geted advertisements.

Defendants criticize Dr. Ford’s procedures and conclusions. They offer their own interpretations of his data, with signif *1027 icantly lower rates of confusion. Defendants cite cases identifying probabilities of confusion of 7.6% and less as de minimis and then argue that Dr. Ford’s results showed de minimis confusion as well. Their critique of Dr. Ford’s methods and interpretations formed the basis of a motion to exclude his expert testimony and report before the district court. The district court denied that motion, however, and allowed the introduction of the evidence.

Defendants may have valid criticism of Dr. Ford’s methods and conclusions, and their critique may justify reducing the weight eventually afforded Dr. Ford’s expert report. The district court’s evidentia-ry ruling is not before us on appeal, however, and weighing admissible evidence at this stage is improper. 29 Defendants’ arguments prove the point that a genuine issue of material fact exists regarding actual confusion. The presence of Dr. Ford’s criticized (but uncontradicted) report, with its strong conclusions that a high likelihood of initial interest confusion exists among consumers, thus generates a genuine issue of material fact on the actual confusion issue.

Because actual confusion is at the heart of the likelihood of confusion analysis, 30 Dr. Ford’s report alone probably precludes summary judgment. In the interest of being thorough, however, we will examine the other seven Sleekcraft factors. On balance, they also support PEI.

b. Factor One: Strength of the Mark.

PEI has established that strong secondary meanings for its descriptive marks exist, and that a genuine issue of material fact exists as to whether it created the secondary meanings. 31 Thus, the first Sleekcraft factor favors PEI.

At this point, defendants concede that they use the marks for their secondary meanings. 32 Thus, they concede that the marks have secondary meanings. They offer only a weak argument regarding the strength of the meanings. 33 Given that defendants themselves use the terms precisely because they believe that Internet *1028 searchers associate the terms with their secondary meanings, disputing the strength of the secondary meanings is somewhat farfetched. The only meaningful dispute is whether PEI created the strong secondary meanings associated with the mark.

PEI offered evidence, in the form of expert reports, tending to show that PEI did create the secondary meanings of “playboy” and “playmate.” PEI’s expert evidence countered the defendants’ expert evidence to the contrary, and suffices to generate a genuine issue of material fact on this issue.

c.Factor Two: Proximity of the Goods.

From an Internet searcher’s perspective, the relevant “goods” are the links to the websites being sought and the goods or services available at those sites. The proximity between PEI’s and its competitor’s goods provides the reason Netscape keys PEI’s marks to competitor’s banner advertisements in the first place. Accordingly, this factor favors PEI as well.

d. Factor Three: Similarity of the Marks.

No doubt exists regarding this factor. Aside from their lack of capitalization, their font, and the fact that defendants use the plural form of “playmate,” the terms defendants use are identical to PEI’s marks. Thus, they are certainly similar. 33

e. Factor Five: Marketing Channels Used.

This factor is equivocal. PEI and the advertisers use identical marketing chan-neis: the Internet. More specifically, each of their sites appears on defendants’ search results pages. Given the broad use of the Internet today, the same could be said for countless companies. Thus, this factor merits little weight.

f.Factor Six: Type of Goods and Degree of Consumer Care Expected.

This factor favors PEI. Consumer care for inexpensive products is expected to be quite low. 34 Low consumer care, in turn, increases the likelihood of confusion. 35

In addition to price, the content in question may affect consumer care as well. We presume that the average searcher seeking adult-oriented materials on the Internet is easily diverted from a specific product he or she is seeking if other options, particularly graphic ones, appear more quickly. Thus, the adult-oriented and graphic nature of the materials weighs in PEI’s favor as well.

g.Factor Seven: Defendants’ Intent in Selecting the Mark.

This factor favors PEI somewhat. A defendant’s intent to confuse constitutes probative evidence of likely confusion: 36 Courts assume that the defendant’s intentions were carried out successfully. In this case, the evidence does not definitively establish defendants’ intent. At a minimum, however, it does suggest that defendants do nothing to prevent click-throughs 37 that result from confusion. Moreover, they profit from such click-throughs.

*1029 Defendants monitor “click-through” rates on the advertisements they display. That is, they monitor the number of times consumers are diverted to their advertisers’ sites. They use the click-through rates as a way to gauge the success of the advertisements and to keep advertisers coming back to their services. Although some click-throughs may be the result of legitimate consumer interest, not confusion, some may be expected to result from confusion. Defendants will profit from both kinds of click-throughs. And they do nothing to ensure that only click-throughs based on legitimate interest, as opposed to confusion, occur.

PEI introduced evidence suggesting that labeling the advertisements would reduce click-through rates. It would also reduce confusion. However, although defendants control the content of advertisements in other contexts, defendants do not require that advertisers identify themselves on their banner ads. Moreover, they do not label the advertisements themselves. Perhaps even more telling, defendants refuse to remove the highly-rated terms “playboy” and “playmate” from their lists of keywords, even when advertisers request that they do so. 38

The above evidence suggests, at a minimum, that defendants do nothing to alleviate confusion, even when asked to do so by their advertisers, and that they profit from confusion. Although not definitive, this factor provides some evidence of an intent to confuse on the part of defendants. This factor thus favors PEI.

h. Factor Eight: Likelihood of Expansion of Product Lines.

Because the advertisers’ goods and PEI’s are already related, as discussed within factor two, this factor is irrelevant.

Having examined all of the Sleekcraft factors, we conclude that the majority favor PEI. Accordingly, we conclude that a genuine issue of material fact exists as to the substantial likelihood of confusion. We now proceed to the defenses advanced by defendants.

3. Defenses.

Defendants assert three defenses: fair use, nominative use, and functional use. Because we have found that a genuine issue of fact exists as to likelihood of confusion under Sleekcraft, we must deny summary judgment as to the fair use defense. A fair use may not be a confusing use. 39 Accordingly, we turn to defendants’ other asserted defenses.

Defendants assert that they make a nominative use of PEI’s marks. We apply a slightly different test for confusion in the nominative use, as opposed to the fair use, context. 40 To be considered a nominative use, the use of a mark must meet the following three-factor test:

Before we apply this test to the facts at hand, we would like to emphasize what facts are not at hand. We note that defendants’ use of PEI’s marks to trigger the listing of PEI sites, and other sites that legitimately use PEI’s marks, 42 is not at issue here. In addition, we note that we are not addressing a situation in which a banner advertisement clearly identifies its source with its sponsor’s name, 43 or in which a search engine clearly identifies a banner advertisement’s source. We are also not addressing a situation in which advertisers or defendants overtly compare PEI’s products to a competitor’s — saying, for example “if you are interested in Playboy, you may also be interested in the following message from[a different, named company].” Rather, we are evaluating a situation in which defendants display competitors’ unlabeled banner advertisements, with no label or overt comparison to PEI, after Internet users type in PEI’s trademarks.

The situation with which we are dealing runs afoul of the first requirement for nominative use. Accordingly, we do not consider the other prongs.

Defendants could use other words, besides PEI’s marks, to trigger adult-oriented banner advertisements. Indeed, they already do so. The list they sell to advertisers includes over 400 terms besides PEI’s marks. There is nothing indispensable, in this context, about PEI’s marks. 44 Defendants do not wish to identify PEI or its products when they key banner advertisements to PEI’s marks. 45 Rather, they wish to identify consumers who are interested in adult-oriented entertainment so they can draw them to competitors’ websites. Accordingly, them use is not nominative. Thus, we reverse the district court’s grant of summary judgment based on nominative use.

Defendants’ final asserted defense, functional use, also fails. Defendants appear not to have raised this defense before the district court. Even if they have not waived the defense, however, it fails. Under the functional use doctrine, parts of a design that have a functional . use may not receive trademark protection. 46 We do not have such a case here.

Nothing about the marks used to identify PEI’s products is a functional part of *1031 the design of those products. PEI could easily have called its magazine and its models entirely different things without losing any of their intended function. Thus, the marks are not functional and may be granted trademark protection. 47

The fact that the marks make defendants’ computer program more functional is irrelevant. Defendants designed their program to identify consumers interested in adult-oriented entertainment so that some percentage of those consumers might be attracted to competitors’ websites, thereby helping defendants generate advertising revenue. Thus, defendants might conceivably be unable to trademark some of the terms used in their program because those terms are functional within that program. Because we are not dealing with defendants’ wish to trademark their computer program, but with PEI’s ability to protect the trademarks it already uses to identify its products, the doctrine of functional use does not help defendants here.

We hold that genuine issues of material fact exist with respect to defendants’ keying practices. Thus, we conclude that summary judgment was inappropriate on the trademark infringement claim.

B. Trademark Dilution

We reverse the district court’s grant of summary judgment on PEI’s second cause of action, trademark dilution, 48 and remand for further proceedings. We conclude that PEI has established that genuine issues of material fact exist regarding two of the three elements that the parties dispute: the famousness of the marks and defendants’ commercial use of the mark. 49 We will address each of the three disputed elements in turn.

1. Famousness of the mark.

The federal dilution statute provides eight factors courts may use, along with other relevant factors, “[i]n determining whether a mark is distinctive and famous.” 50 Those eight factors are:

No grounds exist to contest factors (B), (C), (D), (E), and (H), all of which favor PEI. Defendants directly contest only factor (G): “the nature and extent of use of the same or similar marks by third parties.” 52 We conclude that PEI has established a genuine issue of material fact regarding factor (G). Accordingly, the first contested requirement of trademark dilution favors PEI for purposes of summary judgment.

Defendants introduced evidence that more than forty third-party trademark registrations exist for the terms “playboy” and “playmate,” as well as evidence that hundreds of companies use the terms within their company names. Plaintiffs countered, however, by showing that: (a) many of the companies cited by defendants are active infringers whom PEI is diligently pursuing; (b) others are merely companies who have applied for similar marks but who have not yet received them; and (c) still others are listed several times. The remainder, PEI asserts, are in different fields or in localized areas and should not be counted, at least not when considering whether PEI’s marks are famous within their market. 53

Thus, defendants introduced evidence of third-party use and PEI disputed the evidence with evidence of its own showing that defendants’ list was substantially over-inclusive. A dispute of material fact thus exists as to the only factor relevant to the famousness of the marks that defendants contest. Accordingly, the first contested requirement of dilution favors PEI on summary judgment.

2. Defendants’ commercial use of the mark.

Congress intended to limit only commercial speech, as opposed to political or other more closely protected speech, when it passed the dilution statute; thus, it included the requirement that the use be a commercial one. 54 A successful argument that defendants make no commercial use of the marks, then, would be an argument that the speech associated with their actions was political, not commercial. Defendants do not make such an argument, and it would be difficult to do so in light of the clear evidence of the commercial nature of then* enterprise. Accordingly, PEI has satisfied the second disputed requirement of dilution.

*1033 3. Dilution of distinctive quality of marks.

We conclude that the district court erred when it held, applying the standard then in force, 55 that PEI had shown no likelihood of dilution. However, because the Supreme Court recently clarified the standard for withstanding summary judgment on dilution claims, we vacate the district court’s decision on this point and remand with instructions to reopen discovery to allow the parties to introduce evidence that may satisfy, or undermine, the new standard. 56

Under the old standard, PEI established a genuine issue of material fact regarding likelihood of dilution. With respect to blurring, 57 PEI introduced evidence suggesting that a significant number of Internet users assume that advertisements are sponsored or somehow affiliated with PEI after a search using PEI’s trademarked terms. Defendants did not counter that evidence. With respect to tarnishment, 58 plaintiffs introduced evidence tending to show that consumers consider the materials in the banner ads to be inferior to the materials offered by PEI and that consumers are confused regarding sponsorship of the banner ads. Defendants also did not counter that evidence.

Defendants argue that dilution cannot be found because they do not label their own goods with PEI’s marks. However, when one considers things from the consumers’ perspective, defendants’ argument fails. According to PEI’s evidence, in the minds of consumers, defendants implicitly label the goods of PEI’s competitors with its marks.

Finding fault with the methods used to collect and evaluate PEI’s evidence regarding Internet searchers’ association between the keyed advertisements and PEI, defendants criticize PEI’s evidence regarding the likelihood of blurring. As with the evidence regarding the likelihood of confusion in the infringement claim, however, defendants’ critique of PEI’s evidence pointed to a genuine issue of fact on this issue, not to summary judgment.

Defendants did not counter PEI’s evi-dentiary showing in support of tarnishment. Accordingly, PEI showed a likelihood of tarnishment as well.

Because we conclude that the district court erred under the traditional theories of dilution, we need not reach the parties’ arguments regarding whether Panavision International, L.P. v. Toeppen 59 applies, whether it created a new species of dilution, or whether it remains valid after Congress enacted the Anticybersquatting Consumer Protection Act of 1999. 60 Ac *1034 cordingly, if the old standard applied, we would reverse and remand for post-summary judgment proceedings.

Because the old standard, requiring a showing of a mere likelihood of dilution, no longer applies, we vacate the district court’s decision as to the third element of the dilution claim and remand in order to allow the district court to apply the proper standard. 61 Under that standard, to withstand summary judgment, a party must show that actual dilution has occurred. 62 PEI’s current evidence does not establish actual dilution. Thus, we remand with instructions to re-open discovery and to allow motions directed at the new standard.

Genuine issues of material fact preclude summary judgment on PEI’s dilution claim. The fame of the marks and the likelihood of dilution are in dispute, thereby precluding summary judgment.

IV. CONCLUSION

Genuine issues of material fact exist as to PEI’s trademark infringement and dilution claims. Accordingly, we reverse the district court’s grant of summary judgment in favor of defendants and remand for further proceedings.

REVERSED AND REMANDED.

. Not all banner ads are keyed. Some advertisers buy space for their banner ads but only pay to have their ads displayed randomly. Such ads cost less because they are un-target-ed and are therefore considered less effective.

. The other terms are generally un-trade-marked words associated with adult entertainment, ranging from the expected (sex, parts of the human anatomy, etc.) to the disturbing (gangbangers).

. The search results page lists websites relevant to the search terms pursuant to the search engine's computer program. A user can click on any item in the list to link to the website of the organization listed. Defendants’ search results pages for the terms "playboy” and "playmate” include links to PEI’s websites.

. Playboy Enters., Inc. v. Netscape Communications Corp., 55 F.Supp.2d 1070 (C.D.Cal.), aff'd, 202 F.3d 278 (9th Cir.1999).

. Lopez v. Smith, 203 F.3d 1122, 1131 (9th Cir.2000).

. Id.

. T.W. Elec. Serv., Inc. v. Pac. Elec. Contractors Ass'n, 809 F.2d 626 , 630 (9th Cir.1987).

. Id. (quoting Fed.R.Civ.P. 56(e)).

. Abdul-Jabbar v. Gen. Motors Corp., 85 F.3d 407, 410 (9th Cir.1996).

. Brookfield Communications, Inc. v. West Coast Entm’t Corp., 174 F.3d 1036, 1053 (9th Cir.1999). Because California trademark law claims are “substantially congruent,” we do not examine them separately in this opinion, just as the district court did not. Denbicare U.S.A. Inc. v. Toys R Us, Inc., 84 F.3d 1143, 1152 (9th Cir.1996) (internal quotation marks omitted).

. Federal jurisdiction over trademark cases rests on the Commerce Clause, sweeps as broadly as possible, and clearly encompasses the circumstances of this case. 15 U.S.C. § 1127 (defining "commerce” for jurisdictional purposes as "all commerce which may lawfully be regulated by Congress”); see Steele v. Bulova Watch Co., 344 U.S. 280, 283-84 , 73 S.Ct. 252 , 97 L.Ed. 319 (1952). In addition to defining "commerce,” 15 U.S.C. § 1127 also defines "use in commerce.” 15 U.S.C. § 1127 . That latter definition applies to the required use a plaintiff must make in order to have rights in a mark, as defined by 15 U.S.C. § 1051 . See Planetaty Motion, Inc. v. Techsplosion, Inc., 261 F.3d 1188 , 1194—95 (11th Cir.2001). It does not enter into our jurisdictional analysis.

. 15 U.S.C. § 1114 (l)(a).

. Indeed, we find insufficient evidence to defeat summary judgment on any other theory-

. Brookfield, 174 F.3d at 1062-63 .

. Id. at 1057 .

. Note that if a banner advertisement clearly identified its source or, even better, overtly compared PEI products to the sponsor's own, no confusion would occur under PEI’s theory.

. Brookfield, 174 F.3d at 1057 .

. 174 F.3d 1036 (9th Cir.1999).

. Id. at 1042 .

. Id. at 1043 .

. Id. at 1057 .

. Id.

. Id. at 1062-65 .

. 599 F.2d 341, 348-49 (9th Cir.1979).

. Id.

. Brookfield, 174 F.3d at 1054 . In this case, we conclude that only the final factor — the likelihood of expansion of product lines — does not apply.

. Thane Int’l, Inc. v. Trek Bicycle Corp., 305 F.3d 894 , 902 (9th Cir.2002) ("Evidence of actual confusion constitutes persuasive proof that future confusion is likely.... If enough people have been actually confused, then a likelihood that people are confused is established.”) (internal quotation marks omitted).

. Surveys are commonly introduced as probative evidence of actual confusion. See Schering Corp. v. Pfizer Inc., 189 F.3d 218, 225 (2d Cir.1999).

. Abdul-Jabbar, 85 F.3d at 410 .

. Thane, 305 F.3d at 902.

. Sleekcraft, 599 F.2d at 349 n. 12 (noting that, once a party establishes that it has created secondary meaning, "the protection afforded should be commensurate with the degree of consumer association proven”).

. Indeed, to argue that they use the marks for their primary meaning, as defendants did below, is absurd. Defendants obviously do not use the term "playmate,” for example, for its dictionary definition: "a companion, especially of a child, in games and play.” WEBSTER’S NEW WORLD DICTIONARY, 3d coll. ed. (1988).

. Defendants cite third-party use of the mark as evidence that the secondary meanings of PEI's marks are weak. However, as discussed in the dilution context in section III.B, the degree of third-party use is in dispute in this case, and we do not find their evidence helpful here. Although evidence of extensive third-party use of a mark may be useful in evaluating the strength of the secondary meaning of a mark, we note that such evidence can cut both ways. On the one hand, extensive 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 markholder's. See, e.g., Amstar Corp. v. Domino's Pizza, Inc., 615 F.2d 252, 259-60 (5th Cir.1980). However, if consumers associate the mark with the markholder, nd the markholder’s secondary meaning, despite extensive third-party use, the third-party uses would tend to show the strength of the association created by the markholder. Finally, the markets in which the markholder and the third parties use the mark must be considered. See Nat’l Lead Co. v. Wolfe, 223 F.2d 195 , 204 (9th Cir.1955) (considering, and rejecting, evidence of third-party use because use within the relevant market, for paint, was de minimis). Evidence of third-party use in markets similar to the markholder's is more compelling than evidence of third-party use in unrelated markets. See id. Thus, even if relevant to our inquiry in the infringement context, the evidence would not be dispositive on summary judgment.

. See Sleekcraft, 599 F.2d at 350-52 .

. See Brookfield, 174 F.3d at 1060 (“[W]hen dealing with inexpensive products, customers are likely to exercise less care, thus making confusion more likely.").

. Id.

. See Sleekcraft, 599 F.2d at 348 n. 9.

. If users click on a banner advertisement, Netscape has designed its program to link them immediately to the advertiser’s website. Thus, the user has "clicked-through" the advertisements to the advertiser’s website.

. PEI introduced evidence that, even when advertisers objected to using PEI’s marks to key advertisements, defendants refused to remove the marks from the keying list. This places advertisers in a difficult situation, as described infra.

. Lindy Pen Co., Inc. v. Bic Pen Corp., 725 F.2d 1240, 1248 (9th Cir.1984) (declining to adopt an interpretation of fair use under which a use might be fair "even where likelihood of confusion has been shown,” but noting that liability may not be imposed for truthful comparative advertising, an example of a nominative use).

. See PEI v. Welles, 279 F.3d 796 , 801 (9th Cir.2002) (citing New Kids on the Block v. News America Publ'g, Inc., 971 F.2d 302 , 309 (9th Cir.1992)).

. New Kids on the Block, 971 F.2d at 308 (footnote omitted).

. See, e.g., PEI v. Welles, 279 F.3d at 803-04 (concluding that defendant's use of PEI's marks in the metatags of her website was a permissible, nominative use).

. Doing so might eliminate the likelihood of initial interest confusion that exists in this case.

. Compare Welles, 279 F.3d at 802 (explaining that, because Welles would have to use absurd and lengthy turns of phrase to describe her title as a Playboy Playmate of the Year without using the marks, her use of the marks satisfied the first requirement of nominative use).

. Id. at 801 (noting that, unlike a traditional fair use, a nominative use is a defendant’s use of a mark to identify "not its own product, but the plaintiff's”).

. Qualitex Co. v. Jacobson Prods. Co., 514 U.S. 159, 164-66 , 115 S.Ct. 1300 , 131 L.Ed.2d 248 (1995) (finding color of sponge to be non-functional and therefore granting it trademark protection); Fisher Stoves, Inc. v. All Nighter Stove Works, Inc., 626 F.2d 193, 195-96 (1st Cir.1980) (finding stove roof design functional and therefore denying it trademark protection); Compaq Computer Corp. v. Procom Tech., Inc., 908 F.Supp. 1409, 1423 (S.D.Tex.1995) (finding that words may, in rare instances, be functional and that use of the word "Compaq” as an identifier was functional in case at hand).

. Qualitex, 514 U.S. at 169-70 , 115 S.Ct. 1300 .

. PEI asserted claims under federal and state law. Analysis of the state law is substantially similar to analysis of the federal law. See Avery Dennison Corp. v. Sumpton, 189 F.3d 868, 874 (9th Cir.1999). Accordingly, we do not separately address state law claims.

. The dilution statute, 15 U.S.C. § 1125 (c), provides relief to the owners of famous marks by providing "an injunction against another person's commercial use in commerce of a mark or trade name, if such use begins after the mark has become famous and causes dilution of the distinctive quality of the mark....” The only portion of PEI’s claim not in dispute is the time of defendants’ use, which was clearly after PEI registered the mark. That leaves the famousness of the mark, defendants' commercial use of the mark, and dilution of the mark in dispute.

. 15 U.S.C. § 1125 (c).

. Id. at (c)(1).

. Id. We reject defendants' contention that PEI has waived the remaining factors. PEI argued that its mark was famous before the district court and both parties introduced evidence that was relevant to that issue, including evidence relevant to the factors listed by the statute. Thus, no waiver has occurred.

. PEI thus argues that the market for seeds to which “Playboy” sweet potatoes and yams are targeted and the market for children’s toys and games to which “Playmate” toys are targeted do not affect PEI’s fame within its entirely different niche. PEI makes a strong argument for niche fame which defendants do not adequately counter and which the district court appears not to have considered. This alone provides grounds for reversal. See Syndicate Sales, Inc. v. Hampshire Paper Corp., 192 F.3d 633, 640-41 (7th Cir.1999) (holding that the lower court erred in failing to consider fame within a niche market where defendant was directing activity towards the same market); Avery Dennison Corp. v. Sumpton, 189 F.3d 868 , 878 (9th Cir.1999).

. See J. McCARTHY, TRADEMARKS AND UNFAIR COMPETITION § 24:97:2 (4th ed.2001); House Rep. No. 104-374 (Nov. 30, 1995).

. See Moseley v. V Secret Catalogue, Inc., 537 U.S. 418 , 123 S.Ct. 1115, 1124 , 155 L.Ed.2d 1 (2003) (setting forth the current standard, which requires a showing of actual dilution to withstand summary judgment).

. See Horphag Research Ltd. v. Pellegrini, 337 F.3d 1036, 1041 (9th Cir.2003) (remanding to allow district court to re-consider motion to dismiss under new standard, where mark holder had satisfied old standard).

. Blurring occurs when another's use of a mark creates "the possibility that the mark will lose its ability to serve as a unique identifier of the plaintiff’s product." Panavision Int'l, L.P. v. Toeppen, 141 F.3d 1316 , 1326 n. 7 (9th Cir.1998).

. Tarnishment occurs "when a famous mark is improperly associated with an inferior or offensive product or service.” Id.

. 141 F.3d 1316, 1326 (9th Cir.1998) (approving a district court's finding of dilution where the defendant’s conduct — registering plaintiff Panavision's trademarks in domain names — diminished "the capacity of the Pana-vision marks to identify and distinguish Pana-vision’s goods and services on the Internet”) (internal quotation marks omitted).

. 15U.S.C. § 1125(d).

. See Horphag, 328 F.3d at 1113.

. Moseley, 123 S.Ct. at 1124 .