I. BACKGROUND

Between 1998 and 1999, Jacquelyn Tran (Tran), the presi- dent and owner of Perfumebay, decided to sell perfume on the internet. Tran developed several web sites, including per- fumebay.com. For her business, Tran utilized both conjoined and non-conjoined forms of “perfumebay,” including Per- fumeBay and Perfume Bay.

When Perfumebay applied for a trademark for its Perfume Bay mark, eBay filed an opposition with the United States Patent and Trademark Office. Perfumebay and eBay then entered into negotiations concerning a potential name change by Perfumebay. eBay and Perfumebay stipulated to a suspen- sion of the trademark opposition proceedings pending the result of the settlement negotiations.

When the parties failed to reach a settlement, Perfumebay filed a complaint in federal district court seeking declaratory judgment that its various marks did not infringe on eBay’s mark or otherwise violate the Lanham Act. A bench trial was held on the parties’ various claims.

During the bench trial, Gary Briggs (Briggs), an employee of PayPal, an eBay subsidiary, testified on eBay’s behalf. Briggs described eBay’s operations as “a marketplace.” eBay does not have warehouses or any inventory. Instead, it “bring[s] together sellers and buyers.” According to Briggs, “buyers come to the web site to then bid and/or buy.” There are two methods for “buying and selling on the eBay online marketplace,” including an “auction, in which the price changes over a time period,” and “fixed price” which eBay calls “Buy It Now.” Approximately one-third of eBay’s sales are made through the fixed price method. PERFUMEBAY.COM INC. v. EBAY INC. 14507 Between January, 2002, and September 1, 2004, eBay’s “total gross merchandise value of transactions in [its] fra- grance section” was approximately $6 million. According to Briggs, this indicates “that fragrance is an important category of items traded on the eBay web site.”

Briggs estimated that “about half of the people who come to the eBay web site will type in ‘eBay.com’ or ‘www.ebay.com. And the other half will come in through pages on the internet.” “The largest sites that they come from are Google and Yahoo in the United States.”1

Briggs explained that “text, as opposed to logos, is more important for internet marketing than with regard to non- internet marketing,” because consumers utilize text-based search engines like Google and Yahoo, as opposed to services driven by company logos. According to Briggs, eBay’s domi- nant method of advertising “is in paid search advertising, which are sponsored links, sponsored lines of text that are on a Google or a Yahoo.”

Briggs testified that eBay purchases keywords through two methods. eBay “contract[s] directly with companies like Yahoo and Google either through both short-term and long- term contracts. And then [it] also use[s] what’s called ‘affili- ates,’ which are third parties that make a living by — they buy keywords and market on [eBay’s] behalf.”2 1 “A domain name is the simplest way of locating a web site. If a com- puter user does not know a domain name, she can use an Internet search engine. To do this, the user types in a key word search, and the search will locate all of the web sites containing the key word. Such key word searches can yield hundreds of web sites. To make it easier to find their web sites, individuals and companies prefer to have a recognizable domain name.” Panavision Int’l, L.P. v. Toeppen, 141 F.3d 1316, 1319 (9th Cir. 1998) (internal quotation marks omitted). 2 Briggs explained that for keyword advertising, eBay purchases the right to have its ad displayed “when a single word is searched or a set of multiple words.” For example, if “eBay wanted . . . [its] ad to come up 14508 PERFUMEBAY.COM INC. v. EBAY INC. Briggs conducted a search of “perfume eBay” on the Google, Yahoo, and Ask Jeeves search engines. For the Google search, the natural search results produced links to eBay’s store, as well as “sponsored links for every-day per- fume specials . . .” The sponsored link was to Perfumebay’s web site, specifically its “fragrance museum,” which is “an area for hard-to-find fragrances on PerfumeBay.” According to Briggs, eBay also offered such “hard-to-find products.” Briggs testified that such results are a concern to eBay because “the fact that you have a search result, a paid search result, that’s ‘perfume’ and ‘eBay,’ and then you go to hard- to-find fragrances which indicate that you are not just buying the word ‘perfume,’ you’re buying ‘eBay’ because eBay is known for hard-to-find products.”

For the Ask Jeeves search, Briggs testified that the search for “perfume eBay” produced results for eBay.com and for PerfumeBay.

According to Briggs, the similarity in the names eBay and PerfumeBay concerns eBay because of potential consumer confusion. Briggs testified that he was “concerned that affili- ates are showing results that capitalize the B in PerfumeBay.” According to Briggs, “you have nine million unique users coming to eBay every day. And it doesn’t take, you know, much confusion at all to have someone, you know, think that something is eBay when it isn’t.” Briggs maintained that the problem would not be solved by placing a space between “Perfume” and “Bay” because of “the importance of ‘bay’ in eBay,” and the methods utilized by affiliates. According to Briggs, a space also would not be available, as the internet only utilizes “a connected string of text.”

when someone searched ‘Mercedes Benz,” [it] would purchase those two words in combination.” According to Briggs, “[a]n affiliate is another company, a third party that is in the business of generating new users or traffic for your web site. And they act on your behalf to drive customers to you.” PERFUMEBAY.COM INC. v. EBAY INC. 14509 During cross-examination, Briggs was presented with a search for the terms “perfume” and “bay” as two words. The demonstration produced a paid advertisement for PerfumeBay with a link to eBay. Briggs explained that eBay “commonly would buy the word ‘perfume.’ ” Briggs acknowledged that eBay does not sell PerfumeBay products on its web site. According to Briggs, a reasonable explanation for the adver- tisement was that affiliates would often purchase misspellings of a term. Briggs testified that its “affiliates are buying those links and not eBay.” Briggs also explained that the affiliates follow Google’s policy for the advertisements. Briggs acknowledged that eBay does permit “the use of someone else’s brand name in an advertisement on Google when it has no relationship with that brand name.”

Briggs also demonstrated searches that he conducted involving the terms “sf bay, Los Angeles freeway, Hollywood sign, [and] Gold Line . . .” The Google search for “Los Ange- les freeway” resulted in an ad for “great deals on Los Angeles freeway. Shop eBay and save.” Because it is not possible to purchase the Los Angeles freeway on eBay, Briggs opined that the advertisement was purchased by an eBay affiliate. Briggs presented similar Google search results for the Holly- wood sign; “sf bay”; and “Gold Line.”

Jay Monahan (Monahan), eBay’s Vice President of Litiga- tion, Intellectual Property, testified that he was “concerned about any use of bay which is likely to lead consumers into thinking that that web site is sponsored, affiliated, endorsed by us, or one where they’re basically frontally assaulting the brand by treating bay as a generic reference to online com- merce.” He was also concerned that the use of a generic word plus “bay” would lessen the distinctiveness of eBay’s trade- mark.

According to Monahan, eBay became aware of the Per- fumeBay web site in July, 2002. He was not aware of “any specific instances of confusion involving Perfumebay.” 14510 PERFUMEBAY.COM INC. v. EBAY INC. Tran testified that she developed the idea in “late 1998 or early 1999 . . . to sell perfume from [her] parents’ company on the internet . . .” Tran first sold perfume on eBay’s web site. She used “five or six different user names in order to trade on eBay,” two of which, Beautiful Perfumes and Classy Perfumes, were web sites owned by Tran for the sale of per- fume. She continued to sell perfume on eBay until October, 2004, when Tran decided that she no longer wanted to use eBay, because of eBay’s auction approach.

The “target customer for Perfumebay is anyone who has access to the internet and purchases fragrances and cosmetics . . .” Tran chose the name for the web site while she was still selling perfume on eBay. According to Tran, she “did not have any intention of copying eBay or trying to confuse cus- tomers to come to Perfumebay’s web site because of eBay.”

Tran selected the first part of the name, “perfume,” as a description of her products. She added the term “bay” because she “envisioned a bay filled with ships importing perfumes from all parts of the world and this bay would be the place where perfume lovers could go to locate the selection of fra- grances . . .” However, “there was never an image of a bay filled with ships on [her] web site.” There were not ships on the web site, and Tran “never explained on [her] web site at any time that the name Perfumebay was intended to suggest a bay into which products are brought by ship from abroad . . .” Her web site developers “never created any kind of graphical design for the Perfumebay web site that incorpo- rated a bay filled with ships . . .”

Tran acknowledged that “the Perfumebay web site has, from time to time, displayed the Perfumebay name as one word with a capital B . . .” Additionally, “Perfumebay some- times refers to itself [as] Perfume[B]ay, capitalizing the letters P and B in one word . . .” The sign-up documentation for affiliates “spells the Perfume[B]ay name in one word with the P and the B capitalized . . .” In its trademark application, “the PERFUMEBAY.COM INC. v. EBAY INC. 14511 header Perfume[B]ay.com is displayed as one word with a capital B . . .” According to Tran, the usage of “PerfumeBay” occurred because “it was a common thing [she] saw online where companies would capitalize the first letter of the first word if they had separate words.” However, during the litiga- tion, Tran stopped using “the spelling of Perfume[B]ay, one word, capital P, capital B.”

PerfumeBay also used a logo that included the phrase “Where Perfume Lovers Go.” The logo contains a starfish between “Perfume” and “Bay.” Tran testified that the starfish was used “[t]o convey the image of the bay with perfumes.”

According to Tran, Perfumebay.com operates five other web sites “called Beautiful Perfumes.com; Scentguru.com; Classy Perfumes.com; Beauty Encounter.com and . . . Pefume Shop.com.” “[A]ll of these sites offer the same products for sale as the Perfumebay.com web site . . .” These web sites “have the same categories of products . . . [a]nd they link to the same product descriptions . . .” The majority of key-word advertising is spent for Perfumebay. Tran acknowledged that she could “put more money into keyword advertising for Beautiful Perfume — Beautiful Perfumes.com if [she] wanted to . . .”

In 2004, Perfumebay.com’s sales totaled $6.6 million. The sales for all of Tran’s web sites totaled approximately $9 mil- lion.

Prior to the lawsuit, Tran learned through her customer ser- vice personnel that “people were calling Perfumebay asking if it was associated with eBay . . .” She directed her customer service representatives to inform these callers that Perfume- bay was not associated with eBay.

Philip Johnson (Johnson), chief executive officer of Leo Shapiro and Associates, a research and consulting firm, testi- fied for eBay that he developed a consumer survey “to mea- 14512 PERFUMEBAY.COM INC. v. EBAY INC. sure the extent to which the word ‘bay’ used in conjunction with a common name causes or is likely to cause confusion, . . . or dilution.” According to Johnson’s survey, “a substantial majority of consumers — the average would be 70 percent if one took California and the rest of the country and averaged the two together — believe that eBay is the company that comes to mind when they hear the word ‘bay’ used in connec- tion with a web site or is part of the name of a web site, when they are on the internet engaged in shopping behavior.” John- son concluded that initial interest confusion and actual confu- sion can result from usage of the term “bay.”

Another study designed by Johnson was conducted by tele- phone interviews. The survey reflected that fifty-two percent of the California sample and sixty-four percent of the nation- wide sample responded that eBay came to mind when they saw the term “bay” utilized by a web site. Based on this sur- vey, Johnson opined that “combining a generic term with the word ‘bay’ as part of an [sic] web site name or internet address . . . causes a substantial degree of confusion, and that that will lead to at least dilution if not initial-interest confu- sion.”

Johnson acknowledged that his survey was not designed to measure actual confusion or actual dilution. Johnson also con- firmed that he only tested “bay,” and not “PerfumeBay.”

Eli Seggev (Seggev) testified for Perfumebay and criticized Johnson’s methodology. Seggev opined that, because the Cal- ifornia and nationwide samples could not be combined prop- erly, the survey failed to “project to a population.” Seggev also pointed to Johnson’s failure to utilize a control group. In addition, according to Seggev, the survey neglected to prop- erly validate the interviews. Finally, Seggev faulted the sur- vey for using only telephone interviews, as “the entire process by which consumers are interacting with these companies is visual rather than oral.” PERFUMEBAY.COM INC. v. EBAY INC. 14513 The district court found that conjoined forms of “perfume- bay” created a likelihood of confusion. The district court per- manently enjoined Perfumebay from depicting “Perfume Bay” as a single word. However, the district court held that non-conjoined forms of “Perfume Bay” did not create a likeli- hood of confusion. The district court also held that the Per- fumebay marks did not produce a likelihood of dilution.

Additionally, the district court permanently enjoined eBay from displaying “Perfume Bay” in any advertising, except for “truthful, non-confusing comparative advertising,” based on an unclean hands defense raised by Perfumebay at the end of the bench trial. Perfumebay and eBay filed timely notices of appeal.

II. STANDARDS OF REVIEW

We “review the district court’s determination of likelihood of confusion for clear error.” Reno Air Racing Ass’n, Inc. v. McCord, 452 F.3d 1126, 1135 (9th Cir. 2006) (citation omit- ted). “If the district court’s account of the evidence is plausi- ble in light of the record viewed in its entirety, we may not reverse even though as the trier of fact, we would have weighed the evidence differently.” Id. (citation and alteration omitted).

“A district court’s interpretation of state law is reviewed de novo.” Jorgensen v. Cassiday, 320 F.3d 906, 914 (9th Cir. 2003) (citation omitted).

“The district court’s injunctive relief is reviewed for an abuse of discretion.” Interstellar Starship Servs., Ltd. v. Epix, Inc., 304 F.3d 936, 941 (9th Cir. 2002) (citation omitted). “The grant of a permanent injunction will be reversed only when the district court based its decision on an erroneous legal standard or on clearly erroneous findings of fact.” Id. (citation omitted). 14514 PERFUMEBAY.COM INC. v. EBAY INC. We “review the district court’s denial of attorney’s fees under 15 U.S.C. § 1117 for an abuse of discretion.” Rolex Watch, U.S.A., Inc. v. Michel Co., 179 F.3d 704, 711 (9th Cir. 1999) (citation omitted). “In making this assessment, we review de novo the district court’s legal analysis and interpre- tation of the Lanham Act.” Id. (citation omitted).

III. DISCUSSION

A. Perfumebay’s Appeal

Perfumebay contends that the district court erred in its like- lihood of confusion findings. Perfumebay believes that it should prevail because its marks are not similar to the eBay mark. However, the district court did not clearly err in finding a likelihood of confusion stemming from the conjoined forms of “Perfumebay.”

[1] “The core element of trademark infringement is whether customers are likely to be confused about the source or spon- sorship of the products.” McCord, 452 F.3d at 1135 (citation omitted). “An eight-factor test — the so-called Sleekcraft fac- tors — guides the assessment of whether a likelihood of con- fusion exists.” Id. (citation and internal quotation marks omitted). The Sleekcraft factors are:

(1) the strength of the mark; (2) proximity or related- ness 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.

McCord, 452 F.3d at 1136 n.9 (citation omitted). “The test is a fluid one and the plaintiff need not satisfy every factor, pro- vided that strong showings are made with respect to some of PERFUMEBAY.COM INC. v. EBAY INC. 14515 them.” Surfvivor Media, Inc. v. Survivor Prods., 406 F.3d 625, 631 (9th Cir. 2005) (citation omitted).

[2] In the internet context, “the three most important Sleek- craft factors in evaluating a likelihood of confusion are (1) the similarity of the marks, (2) the relatedness of the goods and services, and (3) the parties’ simultaneous use of the Web as a marketing channel.” Interstellar Starship Servs., 304 F.3d at 942 (citation and internal quotation marks omitted). “When this controlling troika or internet trinity suggests confusion is likely, the other factors must weigh strongly against a likeli- hood of confusion to avoid the finding of infringement.” Id. (citations, alteration, and internal quotation marks omitted). “If the internet trinity does not clearly indicate a likelihood of consumer confusion, a district court can conclude the infringement analysis only by balancing all the Sleekcraft fac- tors within the unique context of each case.” Id.

1. The Similarity of the Marks

“The similarity of the marks will always be an important factor. Where the two marks are entirely dissimilar, there is no likelihood of confusion.” Brookfield Commc’n, Inc. v. West Coast Entm’t Corp., 174 F.3d 1036, 1054 (9th Cir. 1999). “Nevertheless, the more similar the marks in terms of appearance, sound, and meaning, the greater the likelihood of confusion.” Id. (citation omitted). “However, similarity of the marks is but one factor in the Sleekcraft test, albeit an impor- tant one, and a court does not consider the similarity of the marks in the abstract, but rather in light of the way the marks are encountered in the marketplace and the circumstances sur- rounding the purchase.” McCord, 452 F.3d at 1137 (citation and internal quotation marks omitted).

“Because the result of the consideration of one factor can influence the consideration of another, if the trademark hold- er’s mark were strong, the fact that a consumer would likely notice the difference between two marks might not suffice for 14516 PERFUMEBAY.COM INC. v. EBAY INC. a finding that the marks are dissimilar . . . . [T]he fact that the similarity involves the use of a much stronger mark would make that similarity weigh more heavily in the analysis of this factor.” Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1145 n.9 (9th Cir. 2002).

[3] The record in this case reflects the requisite similarity between the two marks. “Perfumebay” incorporates the eBay trademark in its entirety, especially when Pefumebay utilizes the spelling as “PerfumeBay.” For example, in its publisher application, Pefumebay ostensibly utilizes the separate terms “Perfume” and “Bay” at the top. However, the application refers to “perfumebay.com” and “PerfumeBay,” both of which completely incorporate eBay’s strong mark. Although Perfumebay has utilized a logo on its web site separating “Perfume” and “Bay” with a starfish, the domain name itself is “perfumebay.com,” once again incorporating eBay’s mark. Additionally, the similarity is demonstrated in the search engine results, resulting in “eBay” and “perfumebay” links. Although differences exist between the two marks, the fact that eBay’s mark is a strong one weighs against Perfumebay. See id.

[4] The remaining two factors of the internet trilogy weigh equally against Perfumebay. See Interstellar Starship Servs., 304 F.3d at 942. Pefumebay and eBay sell similar products on the internet — perfume. Although they may differ slightly in that eBay offers the additional auction component of its web- site, “even services that are not identical are capable of con- fusing the public.” GoTo.com, Inc. v. Walt Disney Co., 202 F.3d 1199, 1206 (9th Cir. 2000). Thus, “[r]elated goods are generally more likely than unrelated goods to confuse the public as to the producers of the goods.” Brookfield, 174 F.3d at 1055 (citation omitted).

[5] Perfumebay and eBay both utilize the internet “as a marketing and advertising facility, a factor that courts have consistently recognized as exacerbating the likelihood of con- PERFUMEBAY.COM INC. v. EBAY INC. 14517 fusion.” Id. at 1057 (citations omitted). “[T]he Web, as a mar- keting channel, is particularly susceptible to a likelihood of confusion since, as it did in this case, it allows for competing marks to be encountered at the same time, on the same screen.” GoTo.com, 202 F.3d at 1207.

[6] Because the three primary factors for internet trade- marks weigh against Perfumebay, the remaining Sleekcraft factors, actual confusion, marketing channels utilized, degree of consumer care, Perfumebay’s intent, and likelihood of expansion into other markets, “must weigh strongly against a likelihood of confusion to avoid the finding of infringement.” Interstellar Starship Servs., 304 F.3d at 942 (citation and internal quotation marks omitted). The record does not reflect such a showing by Perfumebay. See McCord, 452 F.3d at 1136 n.9.

M2 Software, Inc. v. Madacy Entm’t, 421 F.3d 1073 (9th Cir. 2005), is not to the contrary. In M2 Software, we consid- ered whether the marks “M2 Entertainment” and “M2” were similar. Id. at 1076. We observed:

After considering (1) the addition of “Entertainment” by Madacy, (2) the use of the mark “interactive” in addition to M2 on M2’s Software CDs, (3) the use of other marks in addition to M2 Entertainment on Madacy’s CDs, and (4) the fanciful nature of M2 Software’s M2 mark, the district court ruled that the third Sleekcraft factor weighed only very slightly in M2’s favor. This ruling was not erroneous.

Id. at 1082.

[7] In contrast, Perfumebay failed to demonstrate any addi- tional distinguishing factors, particularly when its marks “per- fumebay” and “perfumeBay” are utilized as domain names, in internet search engine results, and in online advertising. M2 Software also is distinguishable in its sight, sound, and mean- 14518 PERFUMEBAY.COM INC. v. EBAY INC. ing analysis as it did not involve internet usage. See id. at 1084. In sum, the pivotal factor in this case is that Perfume- bay and eBay utilize the internet in selling and advertising similar products, a circumstance not present in M2 Software.

In Entrepreneur Media, we determined that “entrepreneur- pr.com” and “Entrepreneur” were not sufficiently similar. 279 F.3d at 1146-47. However, that case concerned only an inter- net user typing in the domain name. Id. We did not consider confusion resulting from internet search engine usage and online advertising. See id. Additionally, “Entrepreneur,” unlike “eBay,” was a weak mark, “which weighs heavily against finding infringement . . .” Id. at 1153.

Perfumebay is correct that “Perfumebay” and “eBay” are pronounced differently and have different meanings. How- ever, we must evaluate the marks as they are utilized in the marketplace. See McCord, 452 F.3d at 1137. This requires an analysis of the marks in their internet usage, not simply as the terms are pronounced or viewed in the abstract. Internet users type “perfumebay” as a domain name and as an internet search term, and click onto “perfumebay” links as internet search results. Internet users do not utilize verbal communica- tion as a basis for the services that they seek. The likelihood of confusion, therefore, does not arise in a vacuum, but rather from the manner in which “perfumebay” is used on the inter- net. The district court did not err when it considered the likeli- hood of confusion in the relevant context.

It is true that the term “Bay” may convey varying mean- ings. It may serve as a geographical reference to a body of water or describe a commercial storage compartment. How- ever, that the term may convey different impressions to differ- ent individuals is irrelevant to a likelihood of confusion analysis, as the district court’s findings only need be plausi- ble. See id. at 1135 (“If the district court’s account of the evi- dence is plausible in light of the record viewed in its entirety, we may not reverse even though as the trier of fact, we would PERFUMEBAY.COM INC. v. EBAY INC. 14519 have weighed the evidence differently.”) (citation and alter- ation omitted).

Perfumebay also emphasizes that eBay never demonstrated actual confusion. However, “the failure to prove instances of actual confusion is not dispositive against a trademark plain- tiff, because actual confusion is hard to prove; difficulties in gathering evidence of actual confusion make its absence gen- erally unnoteworthy.” Au-tomotive Gold, Inc. v. Volkswagen of America, Inc., 457 F.3d 1062, 1077 (9th Cir. 2006) (cita- tion and alteration omitted) (emphases in the original).

Finally, Perfumebay argues that the district court placed undue emphasis on one-word spellings in internet search engine results. However, the district court’s consideration of “perfumebay” in search engine results was consistent with the evidence presented. Briggs testified that a significant number of consumers reach the eBay web site through internet search engines. eBay’s dominant method of advertising is through sponsored links provided on Google and Yahoo. Thus, the district court did not clearly err in addressing the marks in the contexts which might produce the predominant level of con- sumer confusion.

[8] The district court did not err in holding that the marks were similar enough to create a likelihood of confusion.

2. The District Court’s Finding Of Initial Interest Confusion

Because the district court ruled that any initial interest con- fusion was not actionable, Perfumebay seeks to appeal a deci- sion upon which it prevailed. However, “courts review judgments, not statements in opinions.” Envtl. Prot. Info. Ctr. v. Pacific Lumber Co., 257 F.3d 1071, 1075 (9th Cir. 2001) (citation and internal quotation marks omitted). “A party who receives all that he has sought generally is not aggrieved by the judgment affording the relief and cannot appeal from it.” 14520 PERFUMEBAY.COM INC. v. EBAY INC. Id. (citations omitted). Perfumebay failed to demonstrate that it has the requisite standing to appeal the district court’s favorable ruling. See id.

In any event, the district court did not clearly err in its ini- tial interest confusion analysis. “[A]ctionable initial interest confusion on the Internet is determined, in large part, by the relatedness of the goods offered and the level of care exer- cised by the consumer.” Interstellar Starship Servs., 304 F.3d at 945 (citation omitted). “Of course, the remainder of the Sleekcraft factors complete the case-by-case inquiry necessary to evaluate initial interest confusion on the Internet.” Id.

[9] The district court had a valid basis for finding initial interest confusion.3 The in-court demonstration indicated that consumers might be confused by search results for “perfume” and “eBay” that provided links to Perfumebay. A consumer might assume that Perfumebay is part of eBay’s web site or one of eBay’s internet stores based on the search results. Additionally, initial interest confusion is analyzed pursuant to the Sleekcraft factors for likelihood of confusion. Interstellar Starship Servs., 304 F.3d at 945. As discussed, there is a like- lihood of confusion given the marks’ similarities, the market- ing channels utilized, and the similarity of the products.

3. The Permanent Injunction Against Using perfumebay.com and perfume-bay.com

[10] Perfumebay failed to establish that the district court abused its discretion. “15 U.S.C. § 1116(a) vests the district court with the power to grant injunctions according to princi- ples of equity and upon such terms as the court may deem rea- sonable, to prevent the violation of any right of the trademark 3 It is unclear why the district court relied on the Second Circuit’s approach in Savin Corp. v. Savin Group, 391 F.3d 439 (2d Cir. 2004), given our opinions in the internet context for initial interest confusion. See, e.g., Interstellar Starship Servs., 304 F.3d at 945-46. PERFUMEBAY.COM INC. v. EBAY INC. 14521 owner . . .” McCord, 452 F.3d at 1137 (citation and internal quotation marks omitted). “When the infringing use is for a similar service, a broad injunction is especially appropriate.” GoTo.com, 202 F.3d at 1211 (citation and internal quotation marks omitted). The district court did not clearly err in find- ing that the conjoined version of “perfumebay” created a like- lihood of confusion. eBay and Perfumebay’s services are similar, as they both sell perfume on internet web sites and utilize search engines extensively for attracting customers. A broad injunction addressing such similarities “is especially appropriate.” See id. eBay also presented evidence that the hyphenated term “perfume-bay” contributed to consumer con- fusion, as “e-bay” is associated with “ebay” by consumers.4

4. The District Court’s Asserted Failure To Balance The Harm To Perfumebay

[11] Perfumebay failed to demonstrate an abuse of discre- tion, as it has viable options other than using the infringing domain names. Perfumebay operates five other web sites which offer the same products as perfumebay.com. The web sites share the same product categories and provide links to the same product descriptions. Perfumebay acknowledged that it could shift advertising resources to these other web sites. Although perfumebay.com is the most profitable of these web sites, that fact does not mandate a conclusion that the district court failed to balance the harm to Perfumebay. See Idaho Watersheds Project v. Hahn, 307 F.3d 815, 833-34 (9th Cir. 2002) (approving the balance struck by the district court where the court “was mindful of the equities on all sides”). 4 Perfumebay similarly contends that the district court erred in enjoining Perfumebay from utilizing the mark solely as a referring Uniform Resource Locator (URL). However, Perfumebay failed to demonstrate that the district court abused its discretion. Perfumebay provides no legal sup- port for its assertion that it should be permitted to utilize the infringing mark as a referring URL. 14522 PERFUMEBAY.COM INC. v. EBAY INC. 5. The District Court’s Finding Regarding eBay’s Unclean Hands

“Unclean hands is a defense to a Lanham Act infringement suit.” Japan Telecom, Inc. v. Japan Telecom America Inc., 287 F.3d 866, 870 (9th Cir. 2002) (citation omitted). “Trade- mark law’s unclean hands defense springs from the rationale that it is essential that the plaintiff should not in his trade mark, or in his advertisements and business, be himself guilty of any false or misleading representation.” Id. (citation and internal quotation marks omitted). “To show that a trademark plaintiff’s conduct is inequitable, defendant must show that plaintiff used the trademark to deceive consumers.” Id. (cita- tion omitted). “Bad intent is the essence of the defense of unclean hands.” Id. (citations omitted).

At the conclusion of the bench trial, Perfumebay was allowed to amend its complaint to include the unclean hands defense. The primary testimony on the issue came from eBay’s witness, Briggs. Briggs testified that the advertise- ments for Perfumebay on eBay’s site may have resulted because affiliates would often purchase misspellings of a term. Briggs emphasized that “its affiliates were buying those links not eBay.” Briggs also demonstrated that affiliates had advertised the Los Angeles freeway, the Hollywood sign, sf bay, and Gold Line for sale on eBay. However, Briggs acknowledged that eBay does permit “the use of someone else’s brand name in an advertisement on Google when it has no relationship with that brand name[.]”

[12] The record does not affirmatively demonstrate that eBay used the advertisements to “deceive consumers.” See id. There also is limited evidence, if any, supporting the district court’s finding that “it is axiomatic that the principal is responsible for the acts of the agent acting within the scope of their authority, i.e. the affiliates.” Although eBay’s conduct may be questionable, the evidence does not adequately detail eBay’s relationship with or control over the affiliates. Thus, PERFUMEBAY.COM INC. v. EBAY INC. 14523 it is not clear from the record that eBay acted with the requi- site “bad intent” for an unclean hands finding. See id. As a result, we conclude that there was an insufficient foundation for the injunction against the advertising conducted by eBay’s affiliates. See Jarrow Formulas, Inc. v. Nutrition Now, Inc., 304 F.3d 829, 842 (9th Cir. 2002).

6. The District Court’s Denial of Attorneys’ Fees and Costs to Perfumebay

“Under 15 U.S.C. § 1117(a), a court may award reasonable attorneys’ fees to the prevailing party in exceptional circum- stances, which includes cases in which the act is fraudulent, deliberate, or willful.” Horphag Research Ltd. v. Garcia, 475 F.3d 1029, 1039 (9th Cir. 2007) (citation omitted). “Under both sections 1117(a) and 1117(b), awards are never auto- matic and may be limited by equitable considerations.” Rolex Watch, 179 F.3d at 711 (citation and internal quotation marks omitted).

[13] The district court did not abuse its discretion in deny- ing attorneys’ fees to Perfumebay. Perfumebay was not a pre- vailing party, as it was permanently enjoined from using the conjoined form of its mark. Perfumebay also failed to demon- strate that this was an “exceptional case” involving fraudu- lent, deliberate, or willful conduct. See Horphag Research, 475 F.3d at 1039. Perfumebay is not entitled to attorney’s fees.

B. eBAY’s Cross-Appeal

In its cross-appeal, eBay asserts that the district court erred on several grounds. eBay contends that the district court erro- neously denied its breach of contract claim based on Perfume- bay’s promise to change its name to one that did not utilize a “Bay” suffix. eBay also challenges the district court’s find- ings that there was not a likelihood of dilution, and that non- conjoined forms of “Perfume Bay” did not create a likelihood 14524 PERFUMEBAY.COM INC. v. EBAY INC. of consumer confusion. We conclude that the district court erred only in its finding that there was not a likelihood of dilu- tion.

1. The Alleged Contract Between eBay and PerfumeBay

[14] “The principles of contract formation are the same in both the settlement and the nonsettlement context.” Terry v. Conlan, 131 Cal.App.4th 1445, 1458 (2005) (citation omit- ted). “In order for acceptance of a proposal to result in the for- mation of a contract, the proposal must be sufficiently definite, or must call for such definite terms in the acceptance, that the performance promised is reasonably certain.” Wed- dington Prod., Inc. v. Flick, 60 Cal.App.4th 793, 811 (1998) (citation omitted). “In particular, a provision that some matter shall be settled by future agreement, has often caused a prom- ise to be too indefinite for enforcement.” Id. at 812 (citation and alteration omitted). “If an essential element is reserved for the future agreement of both parties, as a general rule the promise can give rise to no legal obligation until such future agreement. Since either party in such a case may, by the very terms of the promise, refuse to agree to anything to which the other party will agree, it is impossible for the law to affix any obligation to such a promise.” Id. (citation and alteration omitted).

[15] The parties’ communications reflect that only negotia- tions for a potential settlement agreement occurred. Perfume- bay and eBay conducted their negotiations through a series of e-mails between eBay’s counsel, and Perfumebay’s counsel. During these negotiations, Perfumebay acknowledged that it would enter into settlement negotiations if eBay would sus- pend its opposition to Perfumebay’s application for a trade- mark before the Patent and Trademark Office. Perfumebay and eBay filed with the Patent and Trademark Office a Request For Suspension providing that “[t]his request is being submitted in order to permit the parties to discuss possible set- PERFUMEBAY.COM INC. v. EBAY INC. 14525 tlement.” (emphasis added). The request for suspension was made “with the option of either party to resume proceedings at any time during the suspended period.” The parties’ negotiations became increasingly problematic. They failed to agree to the time period during which Perfume- bay would be permitted to use its mark; the acceptable name that Perfumebay would utilize; the duration of Perfumebay’s use of the terms “formerly known as PerfumeBay” and the use of “perfumebay” as a referring URL. At one point, eBay’s counsel observed that progress had been made on the “general contours of a mutually-agreeable settlement.” However, Per- fumebay ultimately rejected eBay’s counter-proposals. [16] The parties’ conduct reflected that there were essential material terms to be negotiated, and that they had only prelim- inary discussions on the goals of settlement. “[M]any settle- ments are reached by an initial agreement on the goals of the settlement. However, agreement to the goals alone may not result in a judicially enforceable settlement agreement.” Terry, 131 Cal.App.4th at 1458 n.3. “Here, the parties’ assent- ing to the goals of the settlement, without agreeing to the means that were material to the settlement, demonstrates that the parties never formed an enforceable contract.” Id. at 1459.5, 6 5 eBay’s reliance on Ersa Grae Corp. v. Fluor Corp., 1 Cal.App.4th 613 (1991), is misplaced. In Fluor, the court of appeal observed that “the con- tract will be enforced if it is possible to reach a fair and just result even if, in the process, the court is required to fill in some gaps.” Id. at 623 (citation omitted). However, the contract’s material terms were “suffi- ciently set forth.” Id. Cal. Civ. Code § 1657, providing that “[i]f no time is specified for the performance of an act required to be performed, a rea- sonable time is allowed,” also does not support eBay’s argument, as there were other disputed material terms. 6 eBay’s argument premised on Allergan Inc. v. Mira Life Group, Inc., 72 U.S.P.Q.2d 1756 (C.D. Cal. 2004), and Kelley Blue Book v. Car- Smarts, Inc., 802 F.Supp. 278 (C.D. Cal. 1992), that we should read a thirty-day time period into the alleged contract is also unavailing. Neither Allergan or Car-Smarts involved contractual interpretation. In Allergan, the district court ordered a thirty-day reporting requirement based on a default judgment for trademark infringement. 72 U.S.P.Q.2d at 1761. In Car-Smarts, the district court ordered a similar reporting requirement after finding trademark infringement. 802 F.Supp.2d at 294. Additionally, in order to read a fixed time period into the alleged contract, we would have to ignore the parties’ negotiations. 14526 PERFUMEBAY.COM INC. v. EBAY INC. 2. eBay’s State Trademark Dilution Counterclaim

[17] “Dilution is a cause of action invented and reserved for a select class of marks — those marks with such powerful consumer associations that even non-competing uses can impinge their value.” Thane Int’l, Inc. v. Trek Bicycle Corp., 305 F.3d 894, 907 (9th Cir. 2002) (citation omitted). “[I]njunctive relief is available under the Federal Trademark Dilution Act if a plaintiff can establish that (1) its mark is famous; (2) the defendant is making commercial use of the mark in commerce; (3) the defendant’s use began after the plaintiff’s mark became famous; and (4) the defendant’s use presents a likelihood of dilution of the distinctive value of the mark.” Avery Dennison Corp. v. Sumpton, 189 F.3d 868, 874 (9th Cir. 1999) (citation omitted). “California’s dilution cause of action is substantially similar, providing relief if the plain- tiff can demonstrate a likelihood of injury to business reputa- tion or of dilution of the distinctive quality of a mark notwithstanding the absence of competition between the par- ties or the absence of confusion as to the source of goods or services.” Id. (citation, alterations, and internal quotation marks omitted).7

[18] “The mark used by the alleged diluter must be identi- cal, or nearly identical, to the protected mark for a dilution claim to succeed.” Nissan Motor Co. v. Nissan Computer Corp., 378 F.3d 1002, 1011 (9th Cir. 2004) (citation, alter- ation, and internal quotation marks omitted). “For marks to be nearly identical to one another, they must be similar enough that a significant segment of the target group of customers sees the two marks as essentially the same.” Thane, 305 F.3d at 906 (citation and internal quotation marks omitted).8 7 The district court held that eBay’s mark was famous and that Perfume- bay used its mark in commerce. The parties do not challenge these find- ings on appeal. 8 eBay’s dilution claim is made pursuant to Cal. Bus. & Prof. Code § 14330. eBay’s “state law dilution claim is subject to the same analysis as its federal claim.” Panavision Int’l, 141 F.3d at 1324. PERFUMEBAY.COM INC. v. EBAY INC. 14527 eBay contends that the district court erred in applying Eli Lilly & Co. v. Natural Answers, Inc., 233 F.3d 456 (7th Cir. 2000), as opposed to the factors articulated in Nabisco, Inc. v. PF Brands, Inc., 191 F.3d 208 (2d Cir. 1999), abrogated by Moseley v. V Secret Catalogue, 537 U.S. 418 (2003). In Nabisco, the Second Circuit articulated several factors for demonstrating dilution, including distinctiveness; the marks’ similarities; “proximity of the products and likelihood of bridging the gap”; “shared consumers and geographic limita- tions”; consumer sophistication; actual confusion; “harm to the junior user and delay by the senior user”; and the “effect of senior’s prior laxity in protecting the mark.” 191 F.3d at 217-222.

The district court rejected the Nabisco test in favor of the test established in Eli Lilly, where the Seventh Circuit limited the likelihood of dilution factors to the marks’ similarity and the renown of the plaintiff’s mark. 233 F.3d at 469.

[19] In Thane, we acknowledged Nabisco’s incorporation of the standard that “the similarity requirement may be less stringent in circumstances in which the senior mark is highly distinctive and the junior mark is being used for a closely related product.” 305 F.3d at 907, n.7 (citing Nabisco). Thane focused on whether the marks were nearly identical and did not explicitly adopt the Nabisco factors. Thane, 305 F.3d at 906-07. Thus, under Thane the court must consider whether the plaintiff’s mark “is so highly distinctive that consumers are likely to view a junior mark that is a bit different as essen- tially the same as the senior one.” Id. at 907 n.7.9 9 Our emphasis on the extent to which the marks are nearly identical and the strength of the senior mark is bolstered by Congress’ passage of the Trademark Dilution Revision Act of 2006 (TDRA). The TDRA amended section 43 of the Trademark Act of 1946 to include a likelihood of dilution cause of action. Pub. L. No. 109-312, 120 Stat. 1730 (2006). The TDRA delineates the following factors in determining “whether a mark or trade name is likely to cause dilution . . .”: 14528 PERFUMEBAY.COM INC. v. EBAY INC. The district court erred in not fully considering the strength of eBay’s mark in making its dilution finding.10 “The stronger a mark — meaning the more likely it is to be remembered and associated in the public mind with the mark’s owner — the greater the protection that it is accorded by the trademark laws.” Brookfield, 174 F.3d at 1058 (citations omitted).

Even under Eli Lilly’s limited dilution test, the district court erred in its application. In Eli Lilly, the Seventh Circuit con- sidered whether there was a likelihood of dilution of the mark “PROZAC” by defendant’s usage of “HEBROZAC.” 233 F.3d at 459-61. The Seventh Circuit held that there was a like- lihood of dilution given the marks’ similarities and the renown of the PROZAC mark. Id. at 469.

A similar conclusion results when comparing the “eBay” and Perfumebay’s marks. The evidence reflected the strength of eBay’s mark, given the mark’s distinctiveness and fame. The evidence also demonstrated the strong recognition and association of eBay’s mark with its services. The Perfumebay marks contain either the entire eBay trademark or the domi- nant suffix “Bay.” Perfumebay utilizes its marks on the inter- net to sell products offered by eBay. With Perfumebay’s marks, consumers may no longer associate the usage of the “Bay” suffix with eBay’s unique services, specifically the sale

(i) The degree of similarity between the mark or trade name and the famous mark. (ii) The degree of inherent or acquired distinc- tiveness of the famous mark. (iii) The extent to which the owner of the famous mark is engaging in substantially exclusive use of the mark. (iv) The degree of recognition of the famous mark. (v) Whether the user of the mark or trade name intended to create an association with the famous mark. (vi) Any actual association between the mark or trade name and the famous mark. Id. at 1731. 10 The district court, without analysis, held that, pursuant to the Eli Lilly test, “the marks are not sufficiently similar to prove a likelihood of dilu- tion.” PERFUMEBAY.COM INC. v. EBAY INC. 14529 of products on an internet-based marketplace. The uniqueness of eBay’s mark is diluted in direct proportion to the extent consumers, particularly internet users, disassociate the eBay mark with eBay’s services.

[20] In sum, eBay possesses a famous and widely known mark, and has expended considerable resources in attaining this status. In applying Eli Lilly, it does not appear that the district court fully considered the highly distinctive qualities of eBay’s famous mark. Thus, the district court erred, as “the senior mark is so highly distinctive that consumers are likely to view a junior mark that is a bit different as essentially the same as the senior one.” Thane, 305 F.3d at 907 n.7 (internal quotation marks omitted).

3. The District Court’s Ruling That Non-Conjoined Variants of Perfumebay Did Not Infringe the eBay Mark

[21] We conclude that the district court did not clearly err in declining to enjoin uses of the non-conjoined versions of Perfume Bay. “[E]ven if a district court finds infringement, it retains the discretion to fashion any remedy which alleviates that confusion. Interstellar Starship Servs., 304 F.3d at 948 (citation omitted). “Certainly, it is not required to enjoin the infringer from all uses of the contested mark.” Id.

[22] The district court’s injunction maintained an equitable balance with respect to the disputed marks. The injunction reduces consumer confusion by eliminating the conjoined forms of “perfumebay.” The conjoined forms encompass all of eBay’s trademark, thus creating confusion when utilized in domain names, online advertising, and search engine results. Although eBay’s argument is persuasive, given the common suffix “Bay,” eBay failed to establish clear error. The sepa- rated forms of “perfume” and “bay” do not include eBay’s entire mark. Unlike the conjoined forms, the non-conjoined forms do not resemble one another in the same manner. The 14530 PERFUMEBAY.COM INC. v. EBAY INC. district court’s more limited injunction, therefore, “balanced the conflicting interests both parties have in the unimpaired continuation of their trademark use.” Interstellar Starship Servs., 304 F.3d at 948 (citation omitted). This is particularly true in view of the fact that non-conjoined forms, such as “Perfume Bay,” cannot be utilized as domain names.11

IV. CONCLUSION

The district court did not clearly err in finding that con- joined forms of “perfumebay” created a likelihood of con- sumer confusion. The district court, therefore, properly enjoined Perfumebay from utilizing such infringing marks. The district court also did not clearly err in finding that the non-conjoined forms of Perfumebay’s mark, such as Perfume Bay, did not create a likelihood of confusion.

However, the district court erred in holding that Perfume- bay’s marks did not produce a likelihood of dilution, as the marks are nearly identical to eBay’s mark. The district court also erred in finding that eBay acted with unclean hands in its advertising, as the record did not affirmatively demonstrate the requisite intent to deceive.

The district court did not abuse its discretion by declining to award attorneys’ fees to Perfumebay, and correctly rejected eBay’s breach of contract claim. 11 Although there was no clear error in the district court’s finding that the non-conjoined forms did not create a likelihood of confusion in this case, non-conjoined forms may nevertheless have a dilutive effect. “Dilu- tion . . . protects the distinctiveness of a particular mark whether or not the products compete or consumer confusion exists. Because dilution and like- lihood of confusion tests are directed at different actions, it does not make sense to import the relatively subjective similarity of the marks test from the likelihood of confusion context into the dilution context.” Thane, 305 F.3d at 906 (citations omitted). PERFUMEBAY.COM INC. v. EBAY INC. 14531 AFFIRMED in part and REVERSED in part.

Each party shall bear its costs on appeal.