We must decide whether two remarkably similar logos used commercially on the World Wide Web are likely to confuse consumers under federal trademark law.

* The Walt Disney Company ("Disney") appeals the district court's grant of a preliminary injunction against it that was sought by GoTo.com ("GoTo"). The injunction prohibits Disney from using a logo confusingly similar to GoTo's mark. GoTo operates a web site that contains a pay-for-placement search engine, which allows consumers to locate items on the Web1 using a search algorithm weighted in favor of those advertisers who have paid to have their products given a priority by the engine. In December 1997, GoTo began using on its web site one of the two logos at issue in this appeal. The GoTo logo consists of the words "GO" and "TO" in a white font stacked vertically within a green circle. Although this green circle has been displayed against backgrounds of various colors, it is very often rendered against a square yellow background. To the right of the word "TO" are the characters ".com" in black, spilling out of the green circle onto the background color.

In preparing to launch a web site of its own, Disney commissioned a design firm, U.S. Web/CKS ("CKS"), to devise a logo for its Web portal, the Go Network, in April 1998. The Go Network is an interconnected collection of web sites, all belonging to Disney properties, designed to provide an easy starting point for consumers who use the Web. The Go Network integrates sites such as <disney.com>, <abc.com>, <abcnews.com>, <abcsports.com>, <espn.com>, <family.com>, and <infoseek.com>. CKS designed a logo that resembles a traffic light: it contains a green circle within a yellow square, with details and contouring that is suggestive of a traffic light with a single lens. Within the green circle, the word "GO" appears in a white font, and next to the traffic light, the word "Network" appears in a black font.

Michael Eisner, the chairman of Disney, approved the CKS logo at the end of August 1998. Then, in December 1998, Disney beta-launched2 the Go Network, displaying its logo prominently on all of the interconnected sites. On December 22, 1998, shortly after this beta launch and more than a fortnight before the formal launch, GoTo complained to Disney about its use of the logo on its Go Network web sites. Disney did not cease using the logo, and GoTo subsequently filed this lawsuit on February 18, 1999, alleging inter alia a violation of S 43(a)(1)(A) of the Lanham Act, 15 U.S.C.S 1125(a)(1)(A). On July 12, 1999, GoTo moved for a preliminary injunction.

On November 12, 1999, the district court granted GoTo's motion for a preliminary injunction. On November 15, Disney filed a notice of appeal and moved the district court to modify and to stay the preliminary injunction. GoTo responded by proposing an amendment to the preliminary injunction, which allowed Disney to phase out its use of the logo in many of its incarnations. On November 16, the district court amended its preliminary injunction order by adding language proposed by GoTo.

Disney again filed a timely notice of appeal on November 17. On November 18, 1999, this Court granted Disney's motion to stay the preliminary injunction pending this expedited appeal.

We review the district court's grant of a preliminary injunction for an abuse of discretion. See Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036, 1045 (9th Cir. 1999). The grant of a preliminary injunction will be reversed only when the district court has based its decision on an erroneous legal standard or on clearly erroneous findings of fact. See FTC v. Affordable Media, 179 F.3d 1228, 1233 (9th Cir. 1999). The legal issues underlying the injunction are reviewed de novo because a "district court would necessarily abuse its discretion if it based its ruling on an erroneous view of law." Brookfield, 174 F.3d at 1046. As to findings of fact, we may affirm the district court "as long as `the findings are sufficiently comprehensive and pertinent to the issues to provide a basis for the decision, or if there can be no genuine dispute about the omitted findings.' " Ocean Garden, Inc. v. Marktrade Co., 953 F.2d 500, 509 (9th Cir. 1991) (quoting Vance v. American Haw. Cruises, Inc., 789 F.2d 790, 792 (9th Cir. 1986). We review a legal and factual determination of likelihood of confusion under the trademark laws for clear error. See Brookfield, 174 F.3d at 1061.

A plaintiff is entitled to a preliminary injunction in a trademark3 casewhen it demonstrates either (1) a combination of "probable success on the merits" and "the possibility of irreparable injury"4 or (2) the existence of "serious questions going to the merits" and that "the balance of hardships tips sharply in his favor." Sardi's Restaurant Corp. v. Sardie, 755 F.2d 719, 723 (9th Cir. 1985). To prevail on a claim under the Lanham Act, GoTo must establish that Disney is using a mark confusingly similar to its own, which it began using a year earlier. See AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 348 (9th Cir. 1979). Or, as the court in Brookfield clarified: "[m]ore precisely, because we are at the preliminary injunction stage, [GoTo] must establish that it is likely to be able to show . . . a likelihood of confusion." 174 F.3d at 1052 n.15 (citing Sardi's Restaurant, 755 F.2d at 723).

The likelihood of confusion is the central element of trademark infringement, and the issue can be recast as the determination of whether "the similarity of the marks is likely to confuse customers about the source of the products."5 Official Airline Guides v. Goss, 6 F.3d 1385, 1391 (9th Cir. 1993) (quoting E. & J. Gallo Winery v. Gallo Cattle Co., 967 F.2d 1280, 1290 (9th Cir. 1992)). We have developed eight factors, the so-called Sleekcraft factors, to guide the determination of a likelihood of confusion. 599 F.2d at 348. Applied to this case, they are (1) the similarity of the marks; (2) the relatedness of the two companies' services; (3) the marketing channel used; (4) the strength of GoTo's mark; (5) Disney's intent in selecting its mark; (6) evidence of actual confusion; (7) the likelihood of expansion into other markets; and (8) the degree of care likely to be exercised by purchasers. See id. at 348-49.

In Brookfield, we noted that the eight-factor test is a "pliant" one, in which "some factors are much more important than others." 174 F.3d at 1054. In the context of the Web in particular, the three most important Sleekcraft factors are (1) the similarity of the marks, (2) the relatedness of the goods or services, and (3) the "simultaneous use of the Web as a marketing channel." Id. at 1034 n.16 (citing Comp Examiner Agency, Inc. v. Juris, Inc., No. 96-0213, 1996 WL 376600, at *1 (C.D. Cal. Apr. 26, 1996)).

We now examine whether GoTo can show that the public is likely to be confused about the source or sponsorship of Disney's logo. See 15 U.S.C. S 1125(a); Brookfield, 174 F.3d 1053. Although we review the district court's finding for clear error, its summary analysis merits an expanded discussion of the basis for that finding.

* [3] The first Sleekcraft factor -the similarity of the marks -has always been considered a critical question in the likelihood-of-confusion analysis. See Brookfield , 174 F.3d at 1054. Together with the relatedness of the services and the use of a common marketing channel, this first factor constitutes part of the controlling troika in the Sleekcraft analysis.

Because the similarity of the marks is such an important question, we must begin our analysis by comparing the allegedly infringing Disney logo to the GoTo mark. Obviously, the greater the similarity between the two marks at issue, the greater the likelihood of confusion. We have developed certain detailed axioms to guide this comparison: first, the marks must be considered in their entirety and as they appear in the marketplace, see Filipino Yellow Pages, Inc. v. Asian Journal Publications, Inc., 198 F.3d 1143, 1147-50 (9th Cir. Dec. 6, 1999); second, similarity is adjudged in terms of appearance, sound, and meaning, see, e.g., Dreamwerks Prod. Group v. SKG Studio, 142 F.3d 1127, 1131 (9th Cir. 1998); and third, similarities are weighed more heavily than differences, see Goss, 6 F.3d at 1392.

With a single glance at the two images, one is immediately struck by their similarity. Both logos consist of white capital letters in an almost identical sans serif font rendered on a green circle. The circle in turn is matted by a square yellow background.

Quibbles over trivial distinctions between these two logos are unimpressive. The logos are glaringly similar. We are unmoved by Disney's arguments to the contrary. The fact that the Patent and Trademark Office has not found the two confusingly similar is not surprising, given that Disney's application to that agency was not lined for color. It is precisely the identical colors that create the confusion: white script in a green circle on a yellow square. See Amsted Indus. Inc. v. West Coast Wire Rope & Rigging Inc., 2 U.S.P.Q.2d 1755, 1758-59 (T.T.A.B. 1992) (pointing out that the Trademark Trial and Appeal Board determines the likelihood of confusion based upon the mark as it is presented for registration, regardless of how the mark may be used in the marketplace).

While the record indicates that GoTo has occasionally displayed its logo on non-yellow backgrounds, and at times on no background at all, the vast majority of impressions of the logo have been of the logo in its prototypical form, i.e., on a yellow background. Viewings of the prototypical logo -with the yellow background -therefore account for almost ninety-eight percent of all impressions of the GoTo logo.

We have no difficulty concluding that the two marks are overwhelmingly similar. In Brookfield, we noted how only a subset of the Sleekcraft factors are needed to reach a conclusion as to whether there is a likelihood of confusion. 174 F.3d at 1054. We emphasize that observation here, and now turn to the remaining two factors in that subset.

The first of the other two controlling Sleekcraft considerations is that "[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. With respect to Internet services, even services that are not identical are capable of confusing the public. Although even Web tyros can distinguish between a web site that, for example, provides discounted travel tickets and one that provides free Webbased e-mail, a user would almost certainly assume a common sponsorship if the sites' trademarks were the same. The <Yahoo.com> web site is just one example of Web genies that coordinate a bevy of distinct services under a common banner. Indeed, Disney's own portal shows the potential for one company to provide a host of unrelated services. Whereas in the world of bricks and mortar, one may be able to distinguish easily between an expensive restaurant in New York and a mediocre one in Los Angeles, see, e.g., Sardi's Restaurant, 755 F.2d at 723-24, the Web is a very different world.

Our ever-growing dependence on the Web may force us eventually to evolve into increasingly sophisticated users of the medium, but, for now, we can safely conclude that the use of remarkably similar trademarks on different web sites creates a likelihood of confusion amongst Web users. The ever-growing number of tentacled conglomerates may force us to conclude that even one hundred and one products could all be sponsored by a single consortium.

In this case, the services offered by GoTo and Disney are very similar. Both entities operate search engines and are, therefore, direct competitors on this score. In Fleishmann Distilling Corp. v. Maier Brewing Co., 314 F.2d 149, 153-55 (9th Cir. 1963), we concluded that beer and whiskey were sufficiently similar products to create a likelihood of confusion regarding the source of origin when sold under the same trade name. Competing Internet search engines are even more similar services.6

Both GoTo and Disney use the Web as a substantial marketing and advertising channel, and we have given special consideration to that forum. In Brookfield, we stated that the use of the Web is a factor "that courts have consistently recognized as exacerbating the likelihood of confusion. " 174 F.3d at 1057 (citations omitted). We now reiterate that the Web, as a marketing 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.

In determining whether there is a likelihood of confusion, we rely heavily on the fact that the marks are similar, that Disney and GoTo offer very similar services, and that they both use the web as their marketing channel. This trinity constitutes the most crucial body of the Sleekcraft analysis, and, in this case, it suggests that confusion is indeed likely. We discuss the remaining Sleekcraft factors only because the parties raised them.

The more likely a mark is to be remembered and associated in the public mind with the mark's owner, the greater protection the mark is accorded by trademark laws. See Kenner Parker Toys Inc. v. Rose Art Indus., Inc., 963 F.2d 350, 353 (Fed. Cir. 1992). This "strength" of the trademark is evaluated in terms of its conceptual strength and commercial strength. See 2 J. Thomas McCarthy, Trademarks and Unfair Competition S 11:83, at 11-143 (rev. ed. 1999).

Marks can be conceptually classified along a spectrum of increasing inherent distinctiveness. See Brookfield, 174 F.3d at 1058. From weakest to strongest, marks are categorized as generic, descriptive, suggestive, and arbitrary or fanciful. See id. We are satisfied that GoTo's mark falls into the suggestive category.

GoTo's logo may appear to be weakened by the fact that the term "Go" and green "Go" circles are certainly common sights on the Internet, but it is the mark in its entirety that must be considered -not simply individual elements of that mark. See California Cooler Inc. v. Loretto Winery, Ltd., 774 F.2d 1451, 1455 (9th Cir. 1985). Further, many of these other "Go" circles might be considered forms of functional trade dress -in that they primarily serve as easily recognized command entry devices rather than marks like GoTo's logo -and are therefore in a separate category not protected by trademark law. See American Greetings Corp. v. Dan-Dee Imports, Inc., 807 F.2d 1136, 1143 (3d Cir. 1986) (holding that "tummy graphics" conveying emotive messages "to help children and adults express their feelings and share them with others" on stuffed toys were non-protectable functional elements). The record discloses thatGoTo's logo in its entirety has been displayed many billions of times, providing compelling evidence of the strength of GoTo's logo. Helpfully, Disney has enthusiastically underscored this point; in its attempt to show that GoTo has not been harmed by any possible infringement, Disney has cited the tremendous success of GoTo and its rise to the twenty-sixth most visited web site on the Internet. Such success only strengthens GoTo's mark.

We do not believe this factor to be of much importance in either the context of the Internet generally or in this case specifically, regardless of whether either logo had herculean strength. In Brookfield, we noted that in situations in which the appearance of the conflicting marks and the services provided are almost identical, "the strength of the mark is of diminished importance in the likelihood of confusion analysis." 174 F.3d at 1059 (citing McCarthy S 11:76). We underline that conclusion here.

Another of the Sleekcraft factors that does not carry much weight in this setting -despite the vigor with which the parties have contested it -is Disney's intent in devising its mark. Disney has produced thousands of pages of alternative logo designs in an attempt to persuade us that it devised its logo completely with its own creative talents, or rather, those of its designer, CKS. Those documents, however, have done little more than persuade us that Disney has many options on which to fall back should it need to find itself a new logo.

We have previously emphasized the minimal importance of the intent factor: "Importantly, an intent to confuse customers is not required for a finding of trademark infringement." Brookfield, 174 F.3d at 1059 (citing Dreamwerks, 142 at 1132 ("Absence of malice is no defense to trademark infringement.")). In fact, in Brookfield, we found that the defendant had no knowledge, actual or constructive, of the plaintiff's mark yet nevertheless ruled against the defendant. Id. Here, we choose not to rummage through the record in a quixotic attempt to determine Disney's intention. For even if we did and concluded that Disney was as innocent as a fawn with no intent to copy or appropriate GoTo's logo, it would prove nothing since no such intent is necessary to demonstrate a likelihood of confusion. We need inquire no further into Disney's intent.

Yet another Sleekcraft factor that sheds little light on the case before us is evidence of actual confusion. While "[e]vidence that the use of the two marks has already led to confusion is persuasive proof that future confusion is likely," the converse is not true. Sleekcraft, 599 F.2d at 352. Indeed, in Brookfield there was no possibility of actual confusion since the plaintiff filed suit before the defendant began using the allegedly infringing mark; nevertheless, the court still found for the plaintiff. 174 F.3d at 1060.

Here, the two sides have conducted studies that purport to show or to refute actual confusion. We decline to evaluate these dueling studies, for even if Disney could show GoTo's study was pure fantasy and that no one was actually confused, it would by no means refute a likelihood of confusion.

We did find it interesting, however, that the Securities and Exchange Commission ("SEC") required a disclaimer on the cover of GoTo's prospectus for its initial public offering, disavowing any connection between Disney and GoTo. The SEC did not require GoTo to disclaim a connection with any of the companies that share similar sounding names, such as Go2Net and Go-2, which suggests to us that the Commission's concern was with the similarity of these two logos rather than the similarity of the more common words "GoTo" and "Go Network."

Because Disney and GoTo compete with one another by providing similar Internet search engines, we decline to evaluate the issue of whether there is a likelihood of expansion of their product lines. See Brookfield, 174 F.3d at 1060 ("The likelihood of expansion in product lines factor is relatively unimportant where two companies already compete to a significant extent.").

In its analysis of the degree of care that users of the Internet take, at least one federal court has ascribed a certain sophistication to Web denizens. See Alta Vista Corp. v. Digital Equip. Corp., 44 F. Supp. 2d 72, 77 (D. Mass. 1998). Although the use of computers may once have been the exclusive domain of an elite intelligentsia, even modern-day Luddites are now capable of navigating cyberspace. Furthermore, the question in this analysis is not how sophisticated web surfers are but, rather, how high the cost is of choosing one service -that is, one web site -over another on the Web. We agree with our previous conclusion that this cost is negligible: it is simply a single click of a mouse.

In the Internet context, in particular, entering a web site takes little effort -usually one click from a linked site or a search engine's list; thus, Web sur fers are more likely to be confused as to the owner ship of a web site than traditional patrons of a brick and-mortar store would be of a store's ownership.

Brookfield, 174 F.3d at 1057.

Navigating amongst web sites involves practically no effort whatsoever, and arguments that Web users exercise a great deal of care before clicking on hyperlinks are unconvincing. Our conclusion is further supported by the Third Circuit's rule that "the standard of care to be exercised by the reasonably prudent purchaser will be equal to that of the least sophisticated consumer." Ford Motor Co. v. Summit Motor Prods., Inc., 930 F.2d 277, 293 (3d Cir. 1991).

From our analysis of the Sleekcraft factors, we conclude that GoTo has demonstrated a likelihood of success on its claim that Disney's use of its logo violates the Lanham Act. From this showing of likelihood of success on the merits in this trademark infringement claim, we may presume irreparable injury. See Brookfield, 174 F.3d at 1066; Metro Pub., 987 F.2d at 640. GoTo has therefore demonstrated the combination of success on the merits and the possibility of irreparable injury necessary to entitle it to a preliminary injunction in a trademark case. Because GoTo has prevailed under this avenue for obtaining injunctive relief, we need not decide whether there exist serious questions on the merits or whether the balance of hardships tips sharply in favor of GoTo.

Disney raises two equitable defenses in its attempt to stymie GoTo's preliminary injunction: laches and unclean hands. As to the first, we have certainly allowed laches to bar trademark infringement cases in the past, but "we have done so only where the trademark holder knowingly allowed the infringing mark to be used without objection for a lengthy period of time." Brookfield, 174 F.3d at 1061 (citing time periods of eight years and two years). Here, GoTo objected to Disney's use of its logo within a fortnight of Disney's beta launch. Although several months did pass between GoTo's filing suit in February 1999 and moving for a preliminary injunction in July 1999, the parties had entered into a tolling agreement under which Disney agreed not to raise this issue. Even without this tolling agreement, however, this delay of only a few months would not be sufficient to bar GoTo's recovery.

As to the question of unclean hands, again we disagree with Disney. We conclude that the record supports the district court's ruling, sub silentio, in favor of GoTo. We note that the district court'ssilence in this regard is not an abuse of discretion. See K-Mart Corp. v. Oriental Plaza, Inc. , 875 F.2d 907, 912 (1st Cir. 1989) ("Overruling the `unclean hands' defense, even without comment, cannot be deemed an abuse of discretion on this record."). While the record does not necessarily demonstrate that GoTo's hands were clean as snow, it does provide evidence to support the district court's decision to ignore Disney's allegations. First, evidence exists to support GoTo's assertion that it did not alter its logo to resemble Disney's; from March 1998 until December 1998, there were approximately 240 million impressions of the GoTo logo with yellow boxes. In fact, the version of the GoTo logo that most resembles the Disney logo was published in Time magazine well before Disney's beta launch. Patrick E. Cole, Capitalist Tool, Time, Mar. 8, 1998, at 20. As for the contention that GoTo manipulated Disney's logo shown in a press release, we recognize that the version of the Disney logo in the press release does, indeed, appear different from the usual image proffered by Disney. Nevertheless, GoTo correctly points to many available variations of the Disney logo that look less like the prototypical Disney lamp than the image used in the press release. Finally, as to GoTo's order to its public relations firm to destroy old drafts of the press release, evidence suggests that this was simply an attempt to avoid a leak of the press release to the media or to Disney. Although the district court did not catalog its findings on this issue, it did not abuse its discretion by failing to do so, because evidence in the record supports the denial of Disney's unclean hands defense.

Furthermore, we disregard the contention that this preliminary injunction alters the status quo ante litem. Disney's arguments to the contrary reveal its confusion as to this term of law. The status quo ante litem refers not simply to any situation before the filing of a lawsuit, but instead to "the last uncontested status which preceded the pending controversy," Tanner Motor Livery, Ltd. v. Avis, Inc., 316 F.2d 804, 809 (9th Cir. 1963) (quoting Westinghouse Elec. Corp. v. Free Sewing Mach. Co., 256 F.2d 806, 808 (7th Cir. 1958)). In this case, the status quo ante litem existed before Disney began using its allegedly infringing logo. The interpretation of this concept that Disney advocates would lead to absurd situations, in which plaintiffs could never bring suit once infringing conduct had begun. Disney severely mischaracterizes this concept, and we conclude that its argument is without merit.

The district court's ruling in this case, even as amended, is disappointing. In a case with the large record and multitude of issues that we have here, we expect the district court to provide findings of fact on specific issues. Nevertheless, a laconic ruling is not, in itself, grounds for reversal; we may still affirm the district court based on evidence in the record. We have, in the past, upheld equally terse district court rulings:

Nonetheless, while it is a close case, we do not believe we must remand for more detailed findings, for despite the factual shortcomings, the basis for the court's decision is clear. The record gives substantial and unequivocal support for the ultimate conclusion . . . .

Unit v. Aerospace Corp., 765 F.2d 1440, 1444 (9th Cir. 1985). In the case before us, we are equally capable of finding support for the district court's ruling in the record. So long as there "can be no genuine dispute about omitted findings" we may affirm the district court, and we do so now. Ocean Garden, 953 F.2d at 509 (quoting Vance, 789 F.2d at 792 (9th Cir. 1986)).

Disney contends that the district court abused its discretion by crafting the preliminary injunction too vaguely and broadly. We do not agree. See SEC v.Interlink Data Network, 77 F.3d 1201, 1204 (9th Cir. 1996) (holding that the scope of injunctive relief is reviewed for an abuse of discretion or application of erroneous legal principles). Although the preliminary injunction certainly does not catalog the entire universe of possible uses of Disney's logo, it is nevertheless sufficiently clear to protect GoTo's interests and to provide Disney with adequate notice. See Fed. R. Civ. P. 65(d); Granny Goose Foods, Inc. v. Brotherhood of Teamsters & Auto Truck Drivers, 415 U.S. 423, 444 (1974) (stating that the party enjoined should "receive fair and precisely drawn notice of what the injunction actually prohibits").

When the infringing use is for a similar service, a broad injunction is "especially appropriate." Century 21 Real Estate Corp. v. Sandlin, 846 F.2d 1175, 1181 (9th Cir. 1988). Here, we have concluded that the services are practically identical. Furthermore, we reiterate the rule in this circuit that the plaintiff should not be held to answer the infringer's subjective assertion that it cannot understand how best to comply with an injunction. See Triad Sys. Corp. v. Southeastern Express Co., 64 F.3d 1330, 1337 (9th Cir. 1995) (holding that placing the burden of determining how to comply with an injunction on the defendant was "appropriate" because it was the defendant who "is the infringer").

In addition, we take comfort in the knowledge that Disney has already demonstrated that it is capable of obeying this injunction, as it did when the clock struck midnight and the preliminary injunction was first entered on November 12, 1999.

Finally, as to whether the bond was adequate, we conclude that the district court did not abuse its discretion in requiring a bond of only $25,000. See Walczak v. EPL Prolong, Inc., 198 F.3d 725 (9th Cir. Dec. 3, 1999). Disney would have us increase that bond 800 times to at least $20,000,000. We decline to do so, and look to Rule 65(c), which places within the discretion of the district court the amount of the bond. Fed. R. Civ. P. 65(c). We have rejected similar requests to raise the bond dramatically. See, e.g., Brookfield, 174 F.3d at 1043-44 (rejecting the defendant's request to increase the bond from $25,000 to $400,000). Were we to grant Disney's request, we would risk denying GoTo access to judicial review since the preliminary injunction would not take effect until GoTo posted the bond. We decline to do so.

We conclude that the district court correctly found that two remarkably similar marks displayed commercially on the Web were likely to cause consumer confusion. We therefore confirm our order of January 27, 2000 vacating our stay of November 18, 1999 and reinstating the preliminary injunction as it was modified on November 16, 1999 by Judge Hatter.

AFFIRMED; preliminary injunction REINSTATED.