| Cite as: 2 Nw. J. Tech. & Intell. Prop. 5 at http://www.law.northwestern.edu/journals/njtip/v2/n2/5 | NJTIP Home > Volume 2 > Issue 2 (Spring 2004) |
¶ 1 It is said that, in the realms of high fashion and fine wine, America follows where France leads. The most important import from France this season, however, may come not from the runways of Paris or from the vineyards of Bordeaux, but from a humble courtroom in the French town of Nanterre.
¶ 2 Nanterre is a bustling, conservative suburb west of Paris, not known for breaking new ground in fashion, wine or the law. On October 13, 2003, a three-judge superior court in Nanterre awarded approximately US$81,000 to two French travel companies on their claims that Google France had infringed their trademarks by selling "sponsored links" to the plaintiffs' competitors.1 Google claims that "more than 200 million times a day people use Google and its partner sites to find what they're looking for."2 Under Google's "Adwords" and "Premium Sponsorship" programs (which appear to be essentially the same in the United States as in France), companies can buy prominent "sponsored links" to their own websites in the search results obtained when Internet users type in specified "keywords." In this case, the keywords were rather descriptive marks3 registered by the plaintiffs: Societe Viaticum and Societe Luteciel.4
¶ 3 The court expressly rejected Google France's argument that it did not commit any acts of infringement but was merely a passive intermediary acting in good faith.5 After noting that Google France has some control over its clients' choice of keywords in that it requires that selected keywords be related to a client's business, the court placed on Google the responsibility of screening the keywords to be sure they are not a competitor's registered marks.6 Most, importantly, the court distinguished the search engine's role from that of an innocent publisher and pointed out that the sale of keyword advertising is a lucrative business for Google.7
¶ 4 This author believes that the French court got it right and that U.S. trademark law should heed the French precedent. That would further the two widely acknowledged goals of U.S. trademark law: protecting consumers from confusion as to the source or sponsorship of goods and services while also protecting trademark owners from loss of the good will built up in their marks.8 Moreover, adopting the French view would merely extend to Internet commerce the same rules of play that have long governed commercial conduct in the physical world. As shown below, U.S. trademark law is already evolving in the same direction as its French counterpart, and the imposition of liability on Internet search engines is a logical next step in that evolution.
¶ 5 Since enactment of the first federal trademark statute in 1870, U.S. trademark law has been viewed as a branch of the law of unfair competition, the basic underlying principle of which is that "no one has any right to represent his goods as the goods of another."9 The modern federal trademark law, enacted in 1946 and often referred to as the "Lanham Act" in recognition of its Congressional sponsor, Representative Fritz Lanham (D-TX), provides protection against infringement of both registered and common law marks, as well as against unfair competition in the form of: (1) deception as to the source or sponsorship of goods or services, (2) a possible affiliation with the trademark owner, and (3) dilution of a famous mark.10
¶ 6 The most fundamental protection afforded by the Lanham Act prevents the unauthorized use of another's trademark for the same or related goods or services, since consumer confusion is inevitable in most cases where two sources put out related goods or services under the same marks.11 This concept is hundreds of years old—developed during the sixteenth century by European guilds—and serves the essential goal of enabling consumers to purchase from sources of established quality and reputation while permitting the purveyor to develop and retain the good will that flows from providing desirable goods and services.12
¶ 7 Although the Internet is of recent vintage, the traditional rules of trademark law and unfair competition are sufficiently elastic to apply in this new setting, and there is no reason to abandon them and thereby lose the value of centuries of legal reasoning and evolution.13 As shown below, U.S. courts are increasingly cognizant that commerce over the Internet is akin to old wine in new bottles, and they are turning to pre-Internet precedent to resolve disputes arising from trademark use in online advertising and sales.
¶ 8 One of the first ways courts used earlier precedent to decide claims related to the Internet was to apply traditional notions of trademark law to disputes over domain names.14 This required overcoming a significant difference between trademarks and domain names: the inability of domain names to be registered by multiple parties for unrelated goods and services.15 Nevertheless, by applying traditional trademark notions such as prior use, registration, good faith adoption, and the distinctiveness of the name or mark, U.S. courts as well as international arbitral bodies have quickly brought relative order and predictability to this area of disputes.16
¶ 9 Perhaps the one trademark doctrine that has proved most useful in combating unauthorized use of another's trademark in a domain name or on a website is that of "initial interest confusion." Briefly stated, initial interest confusion occurs where someone intentionally uses another's mark to capture initial consumer attention, even though subsequent inspection by the consumer dispels confusion as to the source of the goods or services before a purchase is made. As stated by the U.S. Court of Appeals for the Ninth Circuit in a seminal case, Brookfield Communications, Inc. v. West Coast Entertainment Corp.,17 "even where people realize, immediately upon accessing 'moviebuff.com', that they have reached a site operated by West Coast and wholly unrelated to Brookfield, West Coast will still have gained a customer by appropriating the good will that Brookfield has developed in its 'MovieBuff' mark."18 This rationale has been applied by numerous U.S. courts to prohibit the use of another's trademark in a domain name,19 and this prohibition has been codified as part of U.S. trademark law.20 A notable exception to this rule permits a reseller of the goods or services of the trademark owner to use the trademark as part of its site, if the domain name accurately describes the goods or services offered and is not likely to engender consumer confusion.21
¶ 10 In addition to the doctrine of initial interest confusion, courts have used the doctrine of trademark dilution to award relief to trademark owners whose marks are used by others on the Internet. Dilution is broader than infringement in that it involves unauthorized use that tends to diminish the value of a mark, even if the two parties are not selling related goods or services, but it is narrower than infringement in that it is applicable only to "famous" marks.22 For example, the domain name panavision.com was found to dilute the famous PANAVISION mark, even though the domain name was not used in connection with motion picture technology.23 The applicability of dilution was sharply curtailed by the U.S. Supreme Court in 2003, however, when it ruled that a dilution plaintiff must show actual dilution and not just its likelihood.24
¶ 11 Building on this experience with domain names, U.S. courts began using traditional principles of trademark law—especially initial interest confusion—as the rationale for deciding disputes over use of another's trademark as a "metatag". Metatags are terms on an Internet website that are normally invisible to the user but can be read by search engines such as Google, Yahoo, and others. When a consumer enters company A's name or trademark as a search term on an Internet search engine in order to find that company or its products, the consumer can be diverted to the site of competitor B if the competitor has loaded its site with company A's name or marks, whether visible on the site or in the form of metatags.25 Studies have shown that consumer inertia will often cause them to shop at company B rather than take the time to try again to find company A.26
¶ 12 Because metatags are not seen by most Internet users, the unauthorized use of another's trademark as a metatag is particularly pernicious. The consumer who is misled to the site is unlikely to understand how or why she arrived at that site and may assume it is legitimately related to the company, product, or service whose name or trademark she entered in a search engine. Most U.S. courts that have addressed this issue have found the doctrine of initial interest confusion to be a valid basis for prohibiting use of another's trademark in a metatag.27 As the Ninth Circuit opined in Brookfield, "[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."28 The Seventh Circuit Court of Appeals embraced and expanded this analogy, concluding: "[c]ustomers believing they are entering the first store rather than the second are still likely to hang around before they leave. The same theory is true for websites."29
¶ 13 When a website owner is not satisfied with its placement in search engine results, despite possibly having larded the site with his leading competitor's names and marks in metatags, he currently has yet another option: pay the search engines to feature his website whenever an Internet user inputs the competitor's name or mark. Search engine companies refer to the competitor's names and marks simply as "keywords", and they are very happy to sell them to the highest bidder.30 As with metatags, the Internet user may not know the connection—if any—between the search term she used and the goods or services offered on the paid advertisement; Consumers Union national survey found that sixty percent of U.S. Internet users did not know that Internet search engines are paid fees to give prominent placement to certain websites.31 When the consumer-protection group Commercial Alert complained to the Federal Trade Commission about this practice by AOL, Alta Vista, Microsoft, and others, the lack of clear legal precedent limited the FTC to warning Internet search engines that they must use "clear and conspicuous disclosures" to alert consumers to the effect that paid sponsorship has on Internet search results.32 That is why some search engines now provide two sets of search results: sites with a high incidence of the search term, and those whose owners have paid to have their site appear when the keyword is entered.
¶ 14 It is difficult to see how the purchase of a competitor's name or mark as a keyword is any less infringing than use of that name or mark as a metatag. As noted in dictum in one of the first decisions to address this issue, Nissan Motor Co. v. Nissan Computer Corp.,33 "[t]here appears to be no good cause for not extending these protections and limitations to cases where one infringes or dilutes another's mark by purchasing a search term—as opposed to using another's mark in one's metatags."34 Both practices harm the trademark owner and have the potential to mislead consumers.
¶ 15 Although it did not mention the French case, in the most recent U.S. case on point the Ninth Circuit Court of Appeals embraced the same logic as its Gallic counterpart. Giving new life to Playboy's complaint against two Internet search engine companies, Netscape Communications and Excite, the Ninth Circuit reversed the lower court's grant of summary judgment against Playboy.35 Defendants sold a list of pre-selected terms—or keywords—to advertisers in various industries so that the advertisers' websites would be featured whenever an Internet user entered one of the keywords in an Internet search in that search engine. The lists of proffered keywords related to sex and adult-oriented entertainment included plaintiff's registered marks PLAYBOY and PLAYMATE. Playboy brought suit for infringement and dilution of its marks. In reaching its conclusion that, at least, a genuine issue of material fact regarding the likelihood of initial interest confusion precludes summary judgment against Playboy, the majority opinion followed both its own Brookfield precedent and a survey that purported to show that a statistically significant number of Internet users searching for the terms "playboy" and "playmate" would think that Playboy, or an affiliate, sponsored the paid third-party banner ads containing adult content that appear on the search results page.
¶ 16 One might argue that it is the adult content website owners who purchase Playboy's marks as keywords who should be the target of an infringement suit rather than the search engine companies. In truth, both advertisers and search engines are culpable, but it is far more efficient to stop this misconduct by felling the tree at its base rather than snipping off individual leaves each time they bud. As there are no more than twenty search engine companies in the U.S.,36 policing their sale of trademarked terms is far simpler than going after every potential advertiser, especially since not all advertisers are located in the U.S. In addition, as noted above, search engines stand to benefit from the sale of keywords as they derive substantial revenue from such sales. Also critical to the court's reversal was the finding that "defendants do not require that advertisers identify themselves on their banner ads,"37 which permits the initial interest confusion to take place.
¶ 17 In their seminal analysis of the economics of trademarks, Trademark Law: An Economic Perspective,38 Professor William Landes39 and Judge Richard Posner40 explain that the underlying rationale of trademark protection is the promotion of economic efficiency.41 They observe that "intellectual property tends to be particularly costly to protect"42 and refer to unauthorized use of another's trademarks as "free-riding". They warn in terms that have prescient application to the sale of keywords on the Internet that:
The free-riding competitor will, at little cost, capture some of the profits associated with a strong trademark because some consumers will assume (at least in the short run) that the free rider's and the original trademark holder's brands are identical. If the law does not prevent it, free riding will eventually destroy the information capital embodied in a trademark . . . .43
¶ 18 Although the article by Landes and Posner was written before the advent of today's Internet, this author believes they would embrace the rationale of the French court in placing liability on Google for its sale of trademarked keywords. As the Ninth Circuit noted in its Playboy ruling, "defendants do nothing to prevent click-throughs [to third-party sites] that result from confusion. Moreover, they profit from such click-throughs."44 The economics of the Internet point to the search engine companies as the parties most-efficiently positioned to protect consumers and trademark owners from misuse of marks as keywords. The availability on the Internet of a free, fast, and fairly reliable database of marks both registered and pending registration with the U.S. Patent and Trademark Office gives search companies little excuse for accepting—much less affirmatively suggesting—the purchase of third-party trademarks as keywords. Protecting only those marks already on the federal register is like the proverbial "half a loaf," since, as Landes and Posner rightly point out: "the owner of a registered mark bears a lighter burden of proving likelihood of confusion... Registration warns off potential infringers in a way that mere use does not."45
¶ 19 For all of these reasons, the French finding that Google was liable for selling third-party registered trademarks as keywords is fully consonant with U.S. trademark law, in both its traditional application and as it has been adapted to the Internet. If trademarks are to continue to serve as reliable indicators of source, unauthorized use must be prohibited on the Internet as in other media. This is as true for keywords as it is for domain names and metatags, and the most efficient way of preventing unauthorized use is to require search engine companies to verify that the keywords they sell are not registered to a third party. Although U.S. courts have not yet gone so far as the court in Nanterre, that is the next logical step in the evolution of U.S. trademark law. Search engine companies in the U.S. are well-advised to protect themselves against future liability by ceasing their trade in property they do not own.
| © Copyright 2004 by Northwestern University School of Law, Northwestern Journal of Technology and Intellectual Property | Volume 2 Issue 2 (Spring 2004) |