TERENCE T. EVANS, Circuit Judge.

When Internet start-up Natural Answers, Inc. began marketing a line of herbal products called Herbscriptions, we assume it hoped to avoid getting embroiled in a nasty court fight, especially one with a global pharmaceutical giant like the Eli Lilly company. But no such luck, for here it is (along with its founder Brian Feinstein), asking us to reverse a district court order, granted on Lilly's motion for a preliminary injunction, which stopped Natural Answers from marketing its herbal "mood elevator" HERBROZAC. Although Natural Answers does its best to paint this case as David versus Goliath, and hopes for a result similar to that achieved by David, it doesn't have much of a slingshot to carry into the battle.

The basis of Lilly's claim is that the name HERBROZAC comes unfairly close to Lilly's protected mark for PROZAC®, a prescription drug used to treat clinical depression. PROZAC® has received considerable media attention since its rollout in 1988, appearing twice (March 26, 1990, and February 7, 1994) on the cover of Newsweek. In 1999 Fortune magazine named it one of the top six "health and grooming" products of the 20th Century. PROZAC® has been prescribed over 240 million times for 17 million Americans, generating sales of more than $12 billion.

The other five were Gillette disposable safety razors, Band-Aids, Penicillin, Tampons, and birth control pills. Any ranking of this sort, of course, is both subjective and suspect, for Fortune went on to put Tupperware, the penny loafer, and the skateboard on its list of 20th Century bests in the Home, Clothing, and Travel categories.

Natural Answers develops, markets, and sells a line of herbal dietary supplements that it calls Herbscriptions. Herbscriptions are positioned as "drug alternatives" and include supplements called HerbenolPM, HERBALIUM, HERBASPIRIN, and HERBADRYL. Natural Answers' Feinstein testified that each name was chosen to call to mind the function of a famous drug from which its name is derived. Although Natural Answers is careful to characterize its products as alternatives to drugs, as opposed to substitutes for drugs, Feinstein testified that he hopes to persuade consumers not to have "a default reaction to immediately go to their physician and have a prescription for something they may not need."

Notably, the Herbscriptions line contains one supplement, ZONK OUT, whose name is not derived from any other product. ZONK OUT, as its name suggests, is designed to promote deep sleep. Feinstein testified that he chose the name ZONK OUT because he "did not believe that there was a sleeping preparation out there that was so widely known with a suffix that would call to mind a certain function so as to help the consumer know that [it] . . . would help them sleep."

Natural Answers markets its Herbscriptions line exclusively through its Internet Web site but plans to expand its channels of distribution to include retail outlets like health food and convenience stores. Natural Answers' Web site contains a column labeled "Think Herbs — not drugs!", under which it reads "Don't get your prescriptions filled with drugs . . . Get your Herbscriptions filled with Nature!" The Web site also features a table labeled "Herbs v. Drugs" which contrasts Natural Answers' products with a generic "Drug Alternative." In addition, until December 1999, Natural Answers' Web site contained a source code which included the term "Prozac" as a metatag, and described HERBROZAC as "a powerful, and effective all-natural and herbal formula alternative to the prescription drug Prozac." Internet search engines read source codes, which are not immediately visible to an Internet user, in response to search queries. Natural Answers' use of "Prozac" as a metatag was an attempt to guide Internet users searching for information on PROZAC® to Natural Answers' Web site. This effort apparently was unsuccessful, however, because searchers entering the keyword "Prozac" were swamped with Web sites containing a greater number of references to PROZAC®. Natural Answers removed the term "Prozac" from its source codes in response to this lawsuit.

The Ninth Circuit recently explained the function of metatags in Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036, 1045 (9th Cir. 1999): Metatags are HTML codes intended to describe the contents of the web site. There are different types of metatags, but those of principal concern to us are the "description" and "keyword" metatags. The description metatags are intended to describe the web site; the keyword metatags, at least in theory, contain keywords relating to the contents of the web site. The more often a term appears in the metatags and in the text of the web page, the more likely it is that the web page will be "hit" in a search for that keyword and the higher on the list of "hits" the web page will appear.

HERBROZAC is positioned as a "mood elevator." In order to maintain its status as a "food," and thus remain beyond the scope of the Food and Drug Administration's drug approval process, HERBROZAC's label warns that "[t]his product is not intended to diagnose, treat or cure any disease." No prescription is necessary to purchase HERBROZAC. The tablets themselves are larger than average, dark brown with black specs, and have a distinctly herbal odor. In contrast, PROZAC® is available only with a doctor's prescription, which can be filled only by a licensed pharmacist. PROZAC® tablets are green and off-white, and smaller than average.

Before Natural Answers could get HERBROZAC off the ground, Lilly sued to enjoin use of that name, claiming infringement under the Lanham Trademark Act, 15 U.S.C. § 1125(a), dilution under the Federal Trademark Dilution Act, 15 U.S.C. § 1125(c), and violation of Indiana's unfair competition law. Addressing the federal claims only, the district court held that Lilly would likely succeed in proving at trial that the HERBROZAC name is likely to confuse consumers under the Lanham Act. In addition, after holding that Lilly was not required to demonstrate actual dilution in order to obtain relief under the Trademark Dilution Act, the court found that use of the HERBROZAC name created a likelihood of dilution of the PROZAC® mark. Accordingly, the district court enjoined Natural Answers from further use of the HERBROZAC name and ordered that the references to PROZAC® be removed from the source files of Natural Answers' Web site. At that time, Natural Answers had sold less than $2,000 worth of the supplement. This appeal followed.

The standard for deciding a motion for a preliminary injunction is well-established and need not be restated at length. Suffice it to say that the moving party is required to demonstrate a likelihood of success on the merits, that it has no adequate remedy at law, and that it will suffer irreparable harm if preliminary relief is denied. Abbott Labs. v. Mead Johnson Co., 971 F.2d 6, 11 (7th Cir. 1992). If the moving party can pass this threshold, the court will then consider any irreparable harm a preliminary injunction would cause to the nonmoving party, as well as the consequences to nonparties of granting or denying the requested relief. Id. at 11-12. Then, sitting as would a court of equity, the court weighs all of these factors on a sliding scale; the more likely that the plaintiff will succeed on the merits, the less the balance of harms need favor him. Diginet, Inc. v. Western Union ATS, Inc., 958 F.2d 1388, 1393 (7th Cir. 1992).

The district court (Judge David F. Hamilton presiding) painstakingly followed the prescribed formula, although it devoted the bulk of its analysis to Lilly's likelihood of success on the merits. Because the other factors are easily resolved, we'll follow the district court's lead. We review the district court's decision to grant a preliminary injunction under the abuse of discretion standard. Abbott Labs., 971 F.2d at 12. In analyzing the relevant factors, a district court abuses its discretion only when it makes a clear error of fact or an error of law. Id. at 13.

We turn first to the Lanham Act, enacted in 1946. In the Act, Congress explicitly stated it intended to protect registered marks from interference by state legislation, prevent unfair competition, and protect against fraud "by the use of reproductions, copies, counterfeits, or colorable imitations of registered marks. . . ." 15 U.S.C. § 1127. Against this background suggesting a broad legislative purpose, courts have come to a consensus that a Lanham Act plaintiff need only establish that its mark is protectable and that the junior mark is likely to cause confusion among consumers. Munters Corp. v. Matsui Am., Inc., 909 F.2d 250, 252 (7th Cir. 1990); International Kennel Club of Chicago, Inc. v. Mighty Star, Inc., 846 F.2d 1079, 1084 (7th Cir. 1988). Because Lilly registered the PROZAC® mark in 1985 and has used it continuously for more than 5 years, it is incontestable. See 15 U.S.C. § 1065. As a result, Natural Answers cannot — and does not — argue that the PROZAC® mark is outside the protection of the Lanham Act. See Park `N Fly, Inc. v. Dollar Park and Fly, Inc., 469 U.S. 189, 205, 105 S.Ct. 658, 83 L.Ed.2d 582 (1985).

Natural Answers does contend, however, that consumers are unlikely to be confused by the HERBROZAC name. In assessing the likelihood of consumer confusion, we generally consider seven factors: (1) the similarity between the marks in appearance and suggestion, (2) the similarity of the products, (3) the area and manner of concurrent use of the products, (4) the degree of care likely to be exercised by consumers, (5) the strength of the complainant's mark, (6) any evidence of actual confusion, and (7) the defendant's intent (or lack thereof) to palm off its product as that of another. Smith Fiberglass Prods., Inc. v. Ameron, Inc., 7 F.3d 1327, 1329 (7th Cir. 1993). As the district court recognized, these factors are not a mechanical checklist, and "[t]he proper weight given to each . . . will vary from case to case." Dorr-Oliver, Inc. v. Fluid-Quip, Inc., 94 F.3d 376, 381 (7th Cir. 1996). At the same time, although no one factor is decisive, the similarity of the marks, the intent of the defendant, and evidence of actual confusion are the most important considerations. G. Heileman Brewing Co., Inc. v. Anheuser-Busch, Inc., 873 F.2d 985, 999 (7th Cir. 1989).

After an exhaustive analysis of each of the relevant factors, Judge Hamilton concluded that use of the HERBROZAC name was likely to confuse consumers. Specifically, he found that the similarity between the two marks and products, the strength of the PROZAC® mark, consumers' degree of care in choosing between PROZAC® and HERBROZAC, alleged evidence of actual confusion, and Natural Answers' intent to palm off all created a likelihood of confusion. The conclusion regarding likelihood of confusion is a finding of fact which will not be reversed unless it is clearly erroneous. Smith Fiberglass Prods., 7 F.3d at 1329; AHP Subsidiary Holding Co. v. Stuart Hale Co., 1 F.3d 611, 616 (7th Cir. 1993). We'll now take a closer look at each of the factors bearing on the likelihood of confusion.

The similarity between the names PROZAC® and HERBROZAC is obvious: HERBROZAC contains five of the six letters in PROZAC®, and the "B" in HERBROZAC sounds similar to the "P" in PROZAC®. Natural Answers argues that the two marks are dissimilar because HERBROZAC may be pronounced with an emphasis on the first syllable (HERBrozac), thus rendering the "B" and "P" sounds less similar than when the second syllable of HERBROZAC is emphasized (herBROzac). Even under Natural Answers' preferred pronunciation (which seems contrary to the natural pronunciation of a combination of the root words "herb" and "Prozac"), however, the two words are strikingly similar.

Natural Answers next argues that the HERBROZAC name merely "calls to mind" PROZAC®, while at the same time distinguishing itself. The mere fact that one mark brings another mark to mind is not sufficient to establish a likelihood of confusion as to the source of the product. Application of Ferrero, 479 F.2d 1395, 1397 (C.C.P.A. 1973). In addition, Natural Answers asserts that its adaptation of the PROZAC® name is a parody and implies that, as such, it is entitled to some heightened form of protection from trademark liability. In any case, the HERB prefix, according to Natural Answers, so significantly distinguishes HERBROZAC from PROZAC® that consumer confusion is unlikely.

As the district court found, however, "HERBROZAC does much more than merely `call to mind' the PROZAC® mark." Unlike the mark at issue in Ferrero, PROZAC® is a fanciful word that has no meaning independent of Lilly's mark. Such marks are entitled to the highest protection. See Seven-Up Co. v. Tropicana Prods., Inc., 53 C.C.P.A. 1209, 356 F.2d 567, 568 (1966) ("with coined words which are meaningless so far as the English language is concerned, slight variations in spelling or arrangement of letters are often insufficient to direct the buyer's attention to the distinction between marks"). In addition, the district court received evidence that other pharmaceutical companies have expanded their product lines to include dietary supplements based on "St. John's Wort" (the principal herbal component of HERBROZAC), increasing the likelihood that consumers would mistakenly believe that HERBROZAC is affiliated with or sponsored by Lilly. Moreover, HERBROZAC is not a parody of PROZAC®. A parody is a humorous or satirical imitation of a work of art that "creates a new art work that makes ridiculous the style and expression of the original." Rogers v. Koons, 960 F.2d 301, 309-10 (2d Cir. 1992). Natural Answers does not even attempt to point out any humor or satire in its imitation of the PROZAC® mark. More important, we previously cautioned that even if a junior mark meets the definition of a parody, it still runs afoul of the trademark laws if it is likely to confuse consumers. See Nike, Inc. v. "Just Did It" Enterprises, 6 F.3d 1225, 1228 (7th Cir. 1993) ("the parody has to be a takeoff, not a ripoff"). Given the strong similarity in the marks and the potential overlap in Lilly's and Natural Answers' product lines, the mere addition of the HERB prefix does not go far enough to distinguish Natural Answers' product from PROZAC®. Accordingly, the strong similarity of the marks weighs heavily in favor of a finding of likelihood of confusion.

Another factor is whether HERBROZAC and PROZAC® are similar products. On a purely technical level they are different: HERBROZAC is a food and PROZAC® is a drug. See 21 U.S.C. § 321(f)(g)(1) (distinguishing between dietary supplements and drugs for purposes of Federal Food, Drug and Cosmetic Act); Dietary Supplement Health and Education Act of 1994, S. Rep. No. 103-410, pt. VIII, sec. 2(2)(A)-(B) (clarifying that dietary supplements are not drugs and should not be regulated as drugs). This is a critical distinction, moreover, because HERBROZAC is not subject to the rigorous testing and approval process applied to drugs.

The fact that HERBROZAC and PROZAC® are not the same product does not end our inquiry, however, because we are concerned only with whether they are similar. Although its legal papers take a contrary position, Natural Answers would certainly have consumers believe that HERBROZAC and PROZAC® are similar products. Indeed, Natural Answers describes HERBROZAC as a "mood elevator," the mirror image of anti-depressant, and markets it as an alternative to PROZAC®. Natural Answers' dual-track conception of HERBROZAC — as a PROZAC® alternative for marketing purposes, but a dissimilar product for legal purposes — is disingenuous. Moreover, contrary to Natural Answers' suggestion, our inquiry in comparing the two products is not whether they are interchangeable, but whether "the parties' products are the kind the public might very well attribute to a single source (the plaintiff)." International Kennel Club, 846 F.2d at 1089 (affirming preliminary injunction in favor of dog show sponsor against manufacturer of toy dogs); see also Forum Corp. of N. Am. v. Forum, Ltd., 903 F.2d 434, 442 (7th Cir. 1990) (quoting International Kennel Club). Because dietary supplements and drugs serve somewhat similar functions, and because dietary supplements are an area of natural expansion for pharmaceutical companies, HERBROZAC's similarity to PROZAC® is indicative of a likelihood of confusion.

It is unlikely that HERBROZAC and PROZAC® would ever be used concurrently. As Lilly points out, there is some degree of danger that consumers suffering from self-diagnosed depression will view HERBROZAC as a PROZAC® substitute and never seek medical advice. However, a consumer must have a prescription from a doctor in order to obtain PROZAC®, and we have no doubt that doctors would experience little difficulty in distinguishing between the two products. Thus, any consumer with access to PROZAC® would necessarily have the advice of a doctor, who presumably would explain the merits and demerits of each product. Accordingly, it is unlikely that HERBROZAC and PROZAC® would be used concurrently. This factor, therefore, weighs in favor of Natural Answers.

Although we will indulge in the assumption that doctors exercise great care in selecting medications for their patients, the record contains scant evidence of the degree of care likely to be exercised by ordinary consumers. Natural Answers suggests that the legislative history of the Dietary Supplement Health and Education Act — which found that consumers generally believe that dietary supplements should not be regulated as drugs — is indicative of a high standard of care among consumers. See Dietary Supplement Health and Education Act of 1994, S. Rep. No. 103-410, pt. VIII, sec. 2(9)(B). In addition, Natural Answers asserts that its Web site, the only medium through which HERBROZAC is advertised, makes clear that the product is not associated with Lilly.

We are not at all sure that Natural Answers' customers exercise greater than average care. Natural Answers would have us consider only the care exercised by doctors, but the first-line consumers of HERBROZAC are ordinary consumers, who may or may not seek medical advice before buying HERBROZAC. Indeed, Natural Answers' strategy is to divert consumers away from doctors by "hav[ing] them go natural first and not go with drugs," so we must look at the degree of care exercised by all consumers. See Forum Corp. of N. Am., 903 F.2d at 442 ("the court properly should . . . consider the group of potential purchasers of both products"). Unfortunately for Natural Answers, there is just no evidence that consumers as a whole are extraordinarily careful when it comes to dietary supplements. The legislative history cited by Natural Answers demonstrates that consumers are not generally confused between dietary supplements and drugs, and maybe even that they are unlikely to think that HERBROZAC is actually PROZAC®, but it certainly does not address the potential for confusion concerning Lilly's sponsorship of, or affiliation with, HERBROZAC. This is the question we must answer, see Pebble Beach Co. v. Tour 18 I Ltd., 155 F.3d 526, 543 (5th Cir. 1998), and neither the cited legislative history nor the fact that Natural Answers' Web site does not visibly mention PROZAC® or Lilly helps our analysis. We will therefore discount the degree of consumer care factor.

"The term `strength' as applied to trademarks refers to the distinctiveness of the mark, or more precisely, its tendency to identify the goods sold under the mark as emanating from a particular . . . source." Sands, Taylor Wood Co. v. Quaker Oats Co., 978 F.2d 947, 959 (7th Cir. 1992) (quoting McGregor-Doniger, Inc. v. Drizzle, Inc., 599 F.2d 1126, 1131 (2d Cir. 1979), superseded by rule on other grounds, as stated in Bristol-Myers Squibb Co. v. McNeil-P.P.C., Inc., 973 F.2d 1033, 1044 (2d Cir. 1992)). Natural Answers concedes, as it must, that PROZAC® is a strong mark. This factor weighs in favor of a finding of likelihood of confusion.

The record contains no evidence of actual consumer confusion, but Lilly cannot be expected to tender such evidence at this stage. When Lilly filed this lawsuit, Natural Answers had sold less than $2,000 worth of HERBROZAC, so it is not surprising that Lilly cannot identify consumers who were actually confused about the origin or sponsorship of HERBROZAC. Even a statistically reliable consumer survey would likely require a greater sampling than the total number of HERBROZAC customers.

Despite these limitations, Judge Hamilton found a risk of "initial interest confusion." Such confusion, which is actionable under the Lanham Act, occurs when a consumer is lured to a product by its similarity to a known mark, even though the consumer realizes the true identity and origin of the product before consummating a purchase. Dorr-Oliver, 94 F.3d at 382. The district court relied upon the phonetic similarity of HERBROZAC to PROZAC®, as well as Natural Answers' use of the "Prozac" metatag on its Web site, to find a risk of initial interest confusion.

But we are concerned here with evidence of actual confusion, not a mere risk of confusion, and we find no such evidence in the record. The similarity of the two marks is a separate consideration in our analysis, and although it does create a risk of confusion, it does not constitute evidence of actual confusion. The same can be said of Natural Answers' use of the "Prozac" metatag: it is evidence of Natural Answers' intent, see infra, and it creates a risk of confusion, but it is not evidence of actual confusion. Such evidence would most likely take the form of a consumer survey, which would not be limited to a sampling of HERBROZAC customers because "initial interest confusion" is complete prior to any actual purchase. Lilly failed to take such a survey, or to provide other evidence of actual confusion. So we must disagree with the district court that evidence of actual confusion indicates a likelihood of confusion. We have previously made clear, however, that evidence of actual confusion is not essential to a finding of a likelihood of confusion. E.g., Computer Care v. Service Sys. Enter., Inc., 982 F.2d 1063, 1070 (7th Cir. 1992).

The final factor for consideration is Natural Answers' intent. Natural Answers urges us to drop this factor from our analysis, asserting that its intent (or lack thereof) to confuse consumers is not probative of whether consumers are likely to actually become confused. We disagree. The fact that one actively pursues an objective greatly increases the chances that the objective will be achieved, and we have no doubt that this general principle applies here. For this reason, a defendant's intent is an "important factor," Computer Care, 982 F.2d at 1070; Processed Plastic Co. v. Warner Communications, Inc., 675 F.2d 852, 857 (7th Cir. 1982), and can even be weighed more heavily than other factors. See Schwinn Bicycle Co. v. Ross Bicycles, Inc., 870 F.2d 1176, 1184-85 (7th Cir. 1989).

Two facts point to the conclusion that Natural Answers intended to confuse consumers. First, Mr. Feinstein admitted that the name HERBROZAC was coined to remind consumers of PROZAC®, in accordance with Natural Answers' pattern of adapting famous marks for use in its Herbscriptions line. Although Natural Answers now claims that the intent of the name is only to describe the general function of the product and to distinguish it from PROZAC®, we have a difficult time believing this innocent explanation. The HERBROZAC name does nothing — other than through its reference to the PROZAC® mark — to describe the function of "mood elevation." If, at the conclusion of this case, Natural Answers is searching for a new name for its product that truly describes its function, it might come up with something like "Natural Answers' Herbal Mood Elevator," a name that would describe the product's function but not infringe upon Lilly's mark.

The second fact probative of Natural Answers' wrongful intent is its references to PROZAC® in the source codes of its Web site. The clear intent of this effort, whether or not it was successful, was to divert Internet users searching for information on PROZAC® to Natural Answers' Web site. As the Ninth Circuit explained in Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036, 1064 (9th Cir. 1999), "[u]sing another's trademark in one's metatags is much like posting a sign with another's trademark in front of one's store." As such, it is significant evidence of intent to confuse and mislead. See id. at 1062; New York State Soc. of Certified Public Accountants v. Eric Louis Assoc., Inc., 79 F.Supp.2d 331, 341 (S.D.N.Y. 1999) (defendant's bad faith in using plaintiff's trademark in metatag probative of likelihood of confusion). Because Natural Answers' wrongful intent is so obvious, we weigh it heavily.

Having considered each of the relevant factors in our likelihood of confusion analysis, we cannot say that the district court erred in finding such a likelihood. Although the district court improperly weighed the evidence of actual confusion, we believe the other factors — especially the similarity of the marks, the strength of the PROZAC® mark, and Natural Answers' intent to confuse — strongly support the district court's ultimate conclusion. Accordingly, we affirm the district court's preliminary injunction under the Lanham Act.

Lilly also brought a claim under the Trademark Dilution Act of 1995, which opens an alternative avenue of relief for the owners of famous marks. The Act provides:

The owner of a famous mark shall be entitled, subject to the principles of equity and upon such terms as the court deems reasonable, to 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, and to obtain such other relief as is provided in this subsection.

15 U.S.C. § 1125(c)(1). Courts recognize two principal forms of dilution: tarnishing and blurring. Dilution by tarnishing occurs when a junior mark's similarity to a famous mark causes consumers mistakenly to associate the famous mark with the defendant's inferior or offensive product. Panavision Int'l, L.P. v. Toeppen, 141 F.3d 1316, 1326 n. 7 (9th Cir. 1998). Dilution by blurring, the injury at issue here, occurs when consumers "`see the plaintiff's mark used on a plethora of different goods and services,' . . . `raising the possibility that the mark will lose its ability to serve as a unique identifier of the plaintiff's product.'" Hormel Foods Corp. v. Jim Henson Prod., Inc., 73 F.3d 497, 506 (2d Cir. 1996) (quoting 3 McCarthy on Trademarks and Unfair Competition § 24.13[1][a][i] (3d ed. 1995) (hereinafter " McCarthy"); Deere Co. v. MTD Prods., Inc., 41 F.3d 39, 43 (2d Cir. 1994)). The Trademark Dilution Act seeks to prevent both of these forms of dilution by "protect[ing] the trademark owner from the erosion of the distinctiveness and prestige of a trademark caused by . . . a proliferation of borrowings, that while not degrading the original seller's mark, are so numerous as to deprive the mark of its distinctiveness and hence impact." Illinois High School Ass'n v. GTE Vantage Inc., 99 F.3d 244, 247 (7th Cir. 1996). The strongest protection is reserved for fanciful marks that are purely the product of imagination and have no logical association with the product. Nabisco, Inc. v. PF Brands, Inc., 191 F.3d 208, 216 (2d Cir. 1999).

The classic hypothetical examples of dilution by blurring include DuPont shoes, Buick aspirin, Schlitz varnish, and Kodak pianos. "No one would confuse Kodak pianos with Kodak film, but the use of the name on the piano could dilute its effectiveness as a mark for the film." I.P. Lund Trading ApS v. Kohler Co., 163 F.3d 27, 49 (1st Cir. 1998).

There are four elements to Lilly's Trademark Dilution Act claim: (1) the PROZAC® mark is famous, (2) Natural Answers adopted its HERBROZAC name after the PROZAC® mark became famous, (3) use of the HERBROZAC name "causes dilution" of the PROZAC® mark, and (4) Natural Answers' use of the HERBROZAC name is commercial and in commerce. See 15 U.S.C. § 1125(c)(1); Syndicate Sales, Inc. v. Hampshire Paper Corp., 192 F.3d 633, 639 (7th Cir. 1999). The likelihood of confusion, and whether HERBROZAC competes with PROZAC®, are irrelevant. See 15 U.S.C. § 1127. Natural Answers concedes all but the third element — that HERBROZAC "causes dilution" of PROZAC® — so we will focus our attention solely on that issue.

Natural Answers argued before the district court that PROZAC® is not a famous mark but concedes the point on appeal. Even without a concession, however, we agree with the district court's thorough observations concerning the extraordinary fame PROZAC® has achieved.

As a threshold issue, the parties dispute whether the "causes dilution" element can be satisfied by evidence of a mere likelihood of dilution, or whether proof of actual dilution is required. The legislative history of the Act does not specifically address this question, and the circuit courts to have considered it are divided. Compare Westchester Media v. PRL USA Holdings, Inc., 214 F.3d 658, 671 (requiring proof of actual harm) and Ringling Bros.-Barnum Bailey Combined Shows, Inc. v. Utah Div. of Travel Dev., 170 F.3d 449, 461 (4th Cir.) (same), cert. denied, 528 U.S. 923, 120 S.Ct. 286, 145 L.Ed.2d 239 (1999) with Nabisco, 191 F.3d at 224-25 (likelihood of dilution sufficient). See also Times Mirror Magazines, Inc. v. Las Vegas Sports News, L.L.C., 212 F.3d 157, 179 n. 11 (3d Cir. 2000) (dissenting judge agreeing with majority's implicit holding "that the District Court did not err in finding that irreparable injury may be shown even in the absence of actual economic harm"). With guidance from these courts, we'll take our own stab at deciding what "causes dilution" means, reviewing the district court's statutory interpretation of that phrase de novo. Komorowski v. Townline Mini-Mart and Restaurant, 162 F.3d 962, 965 (7th Cir. 1998).

In Ringling Brothers, the Fourth Circuit set out a cogent argument for a restrictive reading of the Trademark Dilution Act that requires the plaintiff to submit proof of actual economic harm. The court observed that state antidilution statutes — upon which the Trademark Dilution Act is based — typically prohibit the use of a mark that creates a "likelihood of . . . dilution of the distinctive quality of a [senior] mark." 170 F.3d at 456 (quoting Model State Trademark Act § 12 (1964), reprinted in 3 McCarthy § 22:8 (4th ed. 1998)); see also, e.g., N.Y. Gen. Bus. Law § 360-1 ("Likelihood of . . . dilution of the distinctive quality of a mark . . . shall be ground for injunctive relief. . . ."); Cal. Bus. Prof. Code § 14330 (same); Fla. Stat. ch. 495.151 (injunction available if "there exists a likelihood . . . of dilution of the distinctive quality of the [senior] mark. . . ."). The federal Act, on the other hand, does not incorporate the word "likelihood," which implies an application to future acts that cause dilution, but instead employs the present tense "causes dilution." In addition, whereas state antidilution statutes generally provide only injunctive relief, indicating that their purpose is to prevent future harm, the federal statute permits injured plaintiffs to obtain damages in limited circumstances. Ringling Brothers, 170 F.3d at 461. And the conduct proscribed by the federal statute is "another person's . . . use," not merely threatened use, of a protected mark. Id. Finally, Ringling Brothers found that the Act's definition of dilution, "the lessening of the capacity of a famous mark to identify and distinguish goods or services," "makes plain what the state statutes may not: that the end harm at which [the Act] is aimed is a mark's selling power, not its `distinctiveness' as such." Id. at 458. Accordingly, the Fourth Circuit held that the use of a name that erodes the distinctiveness of a senior mark but does not affect its holder's bottom line is not actionable under the Act. Id. at 461. The Fifth Circuit essentially adopted these conclusions in Westchester Media, 214 F.3d at 670-71.

The Second Circuit rejected Ringling Brothers in Nabisco, observing that the Fourth Circuit's reading of the statute would subject senior mark holders to uncompensable injury. After all, the primary remedy of the Act is an injunction against future use of an offending mark; damages are permitted only when the defendant acts wilfully. 15 U.S.C. § 1125(c)(2). Thus, if the "causes dilution" element requires proof of actual economic harm, senior mark holders will be restrained from bringing suit prior to suffering an injury for which the Act will not compensate them in many circumstances. Nabisco, 191 F.3d at 224. By the time plaintiffs were permitted to file suit, moreover, junior mark holders might have the defense that the senior mark had lost its distinctiveness due to the numerous other marks that have copied it. Senior mark holders would also be open to the argument that they had failed actively to protect their marks. Finally, if adjudication of a Trademark Dilution Act claim were not possible prior to economic injury, upstart companies would be unable to seek a declaratory judgment that their mark is sufficiently different from a senior mark prior to expending the funds necessary to launch their product. Id. Congress could not have intended these unjust and inefficient results.

In addition, the Fourth Circuit's interpretation of the Act holds plaintiffs to an impossible level of proof. In the case of an immensely successful product such as PROZAC®, it is possible that the distinctiveness of its mark could be diluted even as its sales are increasing, albeit not increasing as much as they would in the absence of the offending mark. See Times Mirror Magazines, 212 F.3d at 179 n. 11 (Barry, J., dissenting on other grounds; citing S. Rep. No. 100-515, at 108 (noting that distinctive quality of a mark "could be materially reduced during a period of rising sales")). Even if diminished revenue could be shown, moreover, it would be immensely difficult to prove that the loss occurred as a result of the dilution of the senior mark. Nabisco, 191 F.3d at 224. Customer surveys, the only other sufficient means of proof suggested in Ringling Brothers, are expensive, time consuming, and subject to manipulation. Nabisco, 191 F.3d at 224. In addition, we doubt that dilution of the distinctiveness of a mark is something that can be measured on an empirical basis by even the most carefully constructed survey. It is hard to believe that Congress would create a right of action but at the same time render proof of the plaintiff's case all but impossible. We therefore side with the Second Circuit and hold that proof of a mere "likelihood of dilution" is sufficient to satisfy the "causes dilution" element of Lilly's case.

Ringling Brothers states that "relevant contextual factors," such as the extent of the junior mark's exposure, the similarity of the marks, and the strength of the senior mark, are also relevant. 170 F.3d at 465. The court then suggests, however, that none of these factors would alone be sufficient, but would only be relevant "as indirect evidence that might complement other proof." Id.

How then, should we decide whether Lilly has demonstrated a likelihood of dilution? The Second Circuit adopted a number of factors from the much criticized concurrence in Mead Data Central, Inc. v. Toyota Motor Sales, U.S.A., Inc., 875 F.2d 1026, 1035 (2d Cir. 1989), added several new factors, and strongly urged courts to think for themselves about what considerations they believe are relevant to dilution. Nabisco, 191 F.3d at 227-28. We'll accept the Second Circuit's invitation and discard a number of the Mead Data factors that we consider irrelevant.

In suggesting a test for dilution under New York's antidilution statute, the concurring judge in Mead Data identified the following pertinent considerations: (1) similarity of the marks, (2) similarity of the products covered by the marks, (3) sophistication of consumers, (4) predatory intent, (5) renown of the senior mark, and (6) renown of the junior mark. Mead Data, 875 F.2d at 1035. As a number of courts and commentators have observed, however, several of these factors are more suited to an inquiry into the likelihood of confusion. E.g., 4 McCarthy § 24:94.2 (4th ed. 2000). For example, the second factor, the similarity of the products in question, is completely irrelevant because the Trademark Dilution Act explicitly states that dilution can occur "regardless of the presence or absence of . . . competition between the owner of the famous mark and other parties. . . ." 15 U.S.C. § 1127. Indeed, the primary application of the Act is to cases involving widely different goods (i.e., Kodak pianos and Kodak film). I.P. Lund Trading ApS v. Kohler Co., 163 F.3d 27, 49 (1st Cir. 1998). Similarly, the sophistication of consumers, intent of the junior mark holder, and renown of the junior mark are all probative of the likelihood of confusion but not particularly relevant to the "capacity of [the senior] mark to identify and distinguish" itself. Id. at 49-50. Accordingly, we will consider, as the district court did, only two of the six factors — the similarity between the HERBROZAC and PROZAC® marks, and the renown of the PROZAC® mark.

The same can be said of several additional factors considered in Nabisco: "actual confusion and the likelihood of confusion, shared consumers and geographic isolation, the adjectival quality of the junior use, and the interrelated factors of duration of the junior use, harm to the junior user, and delay by the senior in bringing the action." 191 F.3d at 228. Each of these factors is either relevant only to likelihood of confusion, or already is incorporated into our injunction analysis.

We already have addressed the similarity between the two marks at issue and concluded that they are highly similar. As for our second factor, PROZAC® has achieved substantial renown. As we noted earlier, PROZAC® has received an immense amount of attention from the news media. In addition to what we have already described, PROZAC® has been the subject of two national bestsellers (Peter D. Kramer's "Listening to Prozac" (1993) and Elizabeth Wurtzel's "Prozac Nation: Young and Depressed in America" (1994)) and segments on 60 Minutes, Good Morning America, and the Oprah Winfrey Show. According to a Baltimore Sun columnist, PROZAC® is "a designer label, a buzzword, a brand name familiar to . . . Americans who have taken it, but also [to] those who wonder if they, too might find a cure for whatever ails them in the little green-and-off-white capsule." Given this evidence, we cannot disagree with the district court's conclusion that PROZAC® "has achieved extraordinary fame in American culture." We therefore affirm the district court's holding that Lilly has shown a likelihood of success in proving a likelihood of dilution.

We also affirm the district court's assessment of the remaining preliminary injunction issues. Irreparable harm is generally presumed in cases of trademark infringement and dilution. See Abbott Labs., 971 F.2d at 16 (recognizing the "well-established presumption that injuries arising from Lanham Act violations are irreparable, even absent a showing of business loss"); American Dairy Queen Corp. v. New Line Prod., Inc., 35 F.Supp.2d 727, 729 (D.Minn. 1998) (presuming irreparable harm by dilution). The brief sales history of HERBROZAC minimizes any potential harm to Natural Answers posed by the district court's injunction, especially given the structure of the injunction, which permits Natural Answers to continue to market its product under an alternative name. Finally, the public interest is served by the injunction because enforcement of the trademark laws prevents consumer confusion. Abbott Labs., 971 F.2d at 19; International Kennel Club, 846 F.2d at 1092 n. 8.

AFFIRMED.