Authors:	Katia Bodard MA (UG-KUL-VUB), LLM (VUB), Lic R (VUB) Institute for European Studies, Vrije Universiteit Brussel
	Bruno de Vuyst Lic R (UIA), LLM (Columbia) Associate Professor, Vesalius College, Vrije Universiteit Brussel and Metropolitan College, Boston University
Subjects:	HTML (Document markup language) Intellectual property (Other articles) Internet computer network law and legislation (Other articles)
Issue:	Volume 9, Number 2 (June 2002)
Category:	Current Developments

Contents:

• The Internet, Information Growth, Search Engines And Meta Tags
• The Emerging Case Law On Meta Tags: Trademark And Unfair Competition Law Protection
• United States cases on meta-tags
• Case law from other jurisdictions
• Provisional conclusions derived from case law
• The Emerging Case Law On Meta Tags: Allowed Use
• Some Policy Conclusions
• Notes

The Internet, Information Growth, Search Engines And Meta Tags

1. Since the Internet has become available outside the United States military in the early 1990s, growth of information on the net has been exponential. Hal VARIAN estimated that new information posted on the net grows yearly by about one to two new, unique exabytes per year, which is roughly 250 megabytes for every man, woman and child on earth. An exabyte is a billion gigabytes, or 10 to the 18th bytes. Printed documents of all kinds compose only 0.003 of the total new information, while materials posted on the net form the large majority of new information.[1] For information to be meaningfully retrieved, the use of search engines is therefore key. These search engines use programs called 'spiders'.[2] Each of these programs has its own method of determining the relevance of a site hit to a search, but most roam the internet, collecting vast numbers of websites and creating indices based on some combination of visible text and meta tags to assign a number of hits and relevant importance to each web site. The sites with the most hits are displayed at the top of a search result's list. These top sites are usually viewed first by the searching user of the net. That searching entails big business is illustrated by the following table, showing February 2002 United States search engine usage in terms of number of public attained and usage length:

    

   Table: February 2002 United States search engine usage, visitors/millions per month and average minutes used/month

   February 2002 Data                                        Visitors/millions                  Avg. Minutes used/month

                   MSN Search                                                      40.5                                         6.0

                   Yahoo Search (incl. Google.Yahoo.com)           36.2                                         10.7

                   Google sites                                                         29.0                                         24.1

                   AOL.com Search                                                 25.2                                         6.9

                   Ask Jeeves                                                      17.7                                         16.2

                   LookSmart                                                            9.8                                           7.6

                   InfoSpace Search                                                   9.0                                           6.9

                   Netscape Search                                                      8.8                                           7.3

                   Overture.com                                                         7.4                                           3.4

                   AltaVista: Search                                                    6.4                                           18.4

   Source: Jupiter Media Metrix

    

2. Meta tags are part of a hypertext markup language (html) source code, which contains keywords that are used to describe the contents of a web page. When meta tags were first introduced in HTML 2.0, these keyword and description functions were not mentioned. Instead, the meta tag appeared to be merely envisioned as a device allowing reading machines to determine the correct way to use a file.[3]

3. Meta tags can contain, however, three kinds of meta tags: descriptive meta tags, keywords and robot tags. Descriptive meta tags describe the content of the html document. Keywords offer search engines 'relevant' words to retrieve a page quickly. Robot tags in a meta tag indicate to search engines that a specific website or certain website pages is/are not to be indexed by the search engine and should therefore not be retrievable. Most meta tags combine only the first two kinds of meta tags. Meta tags, which thus provide keywords and description and which help web directories to properly index, categorize and list web pages, are invisible to the Internet user, but are detectable by search engine spiders and are used by the latter to index the sites which they detect. A web designer may entice a search engine to rate a site higher than another site, or even the site intended to be reached if the spider's count is higher for another site.

4. Thus, a clever web designer can unscrupulously divert business to its site using meta tags, which refer to names or trade- or service marks of competitors. This may lead to 'pagejacking', a practice by which the page's contents, or the meta tags and title information of another's web page, are copied and resubmitted to a search engine.[4]

5. In another ploy, spamdexers do not falsify the descriptive meta tags but seek placement under keyword search terms that are unrelated to the content of their pages, but their descriptive meta tags identify their sites correctly. Spamdexers use thus the most popular search terms in their keyword meta tags, such as 'Disney', 'Princess Diana', or 'Britney Spears' to attract maximum traffic to their site.[5]

   The Emerging Case Law On Meta Tags: Trademark And Unfair Competition Law Protection

   United States cases on meta-tags

6. The first cases on meta tag use and abuse date from 1997.[6] In Oppedahl & Larson v. Advanced Concepts,[7] a law firm dealing with domain name disputes was confronted with the use of the meta tags 'oppedahl' and 'larson' as meta tags in the defendant's web site, without any clear reason but with the impression that the defendant hoped to capture traffic that would gain them domain name registration fees or web site hosting clients. The court banned the defendant from using the names without authorization in application of the Lanham Act, using unfair competition as a basis for its permanent injunction.

7. In Insituform Technologies Inc. v. National Environtech Group, L.L.C. [8] the court enjoined National Environtech for using the claimant's name and trademarked products called Insituform and Insitupipe in its meta tags. Clearer reliance on trademark law was provided in a number of Playboy (tm) related cases. In Playboy Enterprises Inc. v. Calvin Designer Label [9] the defendants spammed their web pages with the trademarks Playboy and Playmate hundreds of times in their web sites "playboyxxx.com" and "playmatelive.com". The court found for a trademark law violation under the classical trademark principles of the Lanham Act, even if the violation was invisible to the public, on the basis of the dilution of trademark theory. In Playboy Enterprises Inc. v. AsiaFocus International and Internet Promotional [10] the 'clickthrough' sites, guiding people towards porn sites, were found liable for a three million USD award, the court stating that the trademarks violated 'have acquired such goodwill and secondary meaning through national and international advertising and promotion that the public has come to associate the "Playmate" and "Playboy" trademarks exclusively with [Playboy Enterprises]".

8. In Brookfield Communications Inc. v. West Coast Entertainment Corp [11] the first appellate decision on meta tags and one in the important West Coast circuit, the court noted that, while the likelihood of confusion is the same in meta tag and in domain name analysis, the context was "quite different". Indeed, the court stated, confusion was less likely when a competitor's web site pops up in a search engine list which would - presumably - also include the trademark owner's own web site. Nevertheless, the court found actionable confusion because of what it termed, in a novel way of phrasing, "initial interest confusion". Such confusion would arise when a searching consumer would be possibly diverted to the defendant's site, even though the consumer discovers, upon visiting this site that this site is not affiliated with the plaintiff. The court found that cases outside of the Internet context, such as Dr. Seuss Enters., L.P. v. Penguin Books U.S.A. [12] and Mobil Oil Corp. v. Pegasus Petroleum Corp. [13] supported its determination that an initial confusion works a sufficient trademark injury to be actionable under the Lanham Act. The court referred back to the Playboy cases but also to Niton Corp. v. Radiation Monitoring Devices, Inc.[14] This case, an early example of pagejacking, is noteworthy because it relied on several findings of the defendant's bad faith. For example, the meta tag descriptions of the defendant turned out to be identical to the ones used by Niton on its web site. Many of the meta tags used by the defendant referred to products that were not sold or marketed by it. Also, and possibly most damning, among the meta tag descriptions used by the defendant was "the home page of [Niton]".

9. The court in Paccar, Inc. v. Telescan Technologies, L.L.C., followed Brookfield.[15] In this case, the defendant operated a number of web sites that provided truck locator services, and some of these sites provided information concerning new and used PETERBILT (tm) and KENWORTH (tm) trucks, which Paccar manufactures. The defendant had provided a disclaimer stating that "This website provides a listing service for name brand products and has no affiliations with any manufacturer whose branded products are listed herein". That disclaimer was, however, not sufficient to avoid initial interest confusion. The court held that the defendant sought to "derive a benefit from the reputation of [plaintiff's] Peterbilt and Kenworth marks".

10. In Ely Lilly & Co. v. Natural Answers Inc. [16] the appellate court deemed the use of the Eli Lilly trademark in the meta tags of the source code of its website probative of wrongful intent, which justified weighing that factor heavily in an analysis of the likelihood of confusion.

    Case law from other jurisdictions

11. The emerging case law emanating from other than U.S. jurisdictions appears to bear out that an aggrieved trademark owner may rely on a dilution of trademark theory, in the case of a famous trademark, or unfair competition law.

12. In Germany, two cases were found reported.[17] The Landesgericht Mannheim found for the owner of the service mark ARWIS who, upon a search for its own mark, found that it was directed to the website of a company called WMC. The court stressed that it acted on a finding of WMC attempting to misdirect customers, however it founded its decision on both trademark and on unfair competition law.

13. In the second case, the car rental company Sixt had put as meta tags the names of its competitors, such as Hertz, Avis, and EuropCar. It was finally Budget that put a stop to this unfair competition through diversionary tactics by a complaint to the Landesgericht Muenchen.

14. In Italy, a Rome Tribunal prohibited the use by Crowe Italia of the word 'Genertel', identifying Trieste Venezia As Genertel S.p.A., a competitor, as attempting to divert customers.[18]

15. In Belgium three cases were reported condemning the use of a competitor's name in meta tags. A first ruling dates from October 15, 1999 and involved the Belgian (former monopolist) telecommunication operator Belgacom v. Intouch.[19] The trademark name Belgacom and a comparative sounding name Belgacon were integrated in the meta tags of Intouch' website. Belgacom alleged misuse of its trademark and unfair commercial practices, which was upheld by the court. An interesting aspect of this case is the argument of the defendant that the meta tags were chosen by the web designer without knowledge of the defendant. The court dismissed, however, this argument holding that meta tags form a part of the content delivered to the web designer, that defendant was the best placed party to decide on the keywords to be used in the meta tags and that usually a contract is made between the web designer and the content provider whereby the former disclaims itself from any liability and provides for a clause in the contract stating that the content provided for is not illegal or does not infringe a third party's rights.

16. A second case, BVBA Aerodata International Surveys v. NV Eurosense Belfotop[20] went as far as the court of appeals of Antwerp. In this appellate case it was decided that it was prohibited to integrate the name of a competitor in the meta tags since this could be considered to be a contravention of fair commercial practices. There was also no reason, according to the court, to withhold the argument of comparative advertising as defined in the Belgian Law on Commercial Practices. Furthermore, the appeals court stated that the injunction issued by the lower court was accurately formulated, and that there was no reason to deduce from this formulation that the plaintiff in the case (defendant in first instance) had to answer for the adjustment of the search results conducted by search engines like AltaVista and Infoseek under de search term 'Eurosense'.

17. However in another Belgian case, Resiplast c. Resin[21] which refers explicitly to the former two cases, the court decided that it was the responsibility of the defendant alone for inserting meta tags in the programming languages of its website, and therefore not the responsibility of the administrator of the search engine. The defendant had argued that the search engines had neglected to update their files, but this argument was clearly not retained by the court, which condemned the use of the trade name of the defendant's competitor in the programming language of its own website.

    Provisional conclusions derived from case law

18. From this, still mainly U.S., case law, certain action strategies for a trademark owner that perceives a use of its mark by another party appear to emerge.[22] First, the owner of a famous, well-known mark may allege trademark dilution. Second, any trademark owner may allege trademark infringement, alleging that there is a likelihood of confusion as to a source of goods or services. There are, however, some issues with the use of the trademark law's confusion theory. First of all, the meta tag is invisible; hence the searching potential customer has not been confused at the time of her or his search. It is also open to question whether a potential customer, upon landing on a competitor's site listed in the search results, believes that such a competitor or the products or services connected therewith are connected with the trademark owner.

19. Brookfield provides a way out with its 'initial interest confusion' theory to answer these remarks. Whether such initial interest confusion in itself is sufficient to warrant injunctive relief, i.e. the prohibition of using a trademark that one does not own as a meta tag attached to one's website, may be questioned. In each case, there appears to be commercial use of the meta tag, always in bad faith. Eli Lilly appears to bear out that such commercial use in bad faith is required to find for injunctive relief. Whether in such situation one may base oneself on any confusion theory under trademark law is to be questioned.

20. Indeed, Brookfield and Eli Lilly appear to point to allegations on the basis of the (common or other) law on unfair competition as the most effective potentially actionable way of proceeding. The Nettis Environmental Ltd. V. IWI, Inc. et al case appears to bear this out.[23] Both companies are competitors in the resolution of in-plant ventilation services. IWI registered its website with over 500 search engines with the tags Nettis and Nettis Environmental as search engine key words. The result was predictable but also the outcome of the case.

21. Absolute proof of the requisite intent on diverting potential customers appears - see Eli Lilly and the Belgacom case - not always required when operating in a commercial and competitive situation, although it appears to help the plaintiff such as in cases like Nettis or Niton; and proof of possible or actual harm, by diversion of customers, appears not required either.

    The Emerging Case Law On Meta Tags: Allowed Use

22. Not all use of another's meta tags stands prohibited.[24] In Playboy Enterprises, Inc. v. Welles [25] the court permitted Miss Welles, a Playboy's 1981 Playmate of the Year, to use terms such as Playmate and Playboy on her web pages and within her meta tags. The court deemed that Welles had a legitimate reason to use them in order to describe herself and, importantly, to catalog her web site properly with search engines. The court noted emphatically that Welles was not using the terms in an attempt to mislead visitors into believing that they were at a Playboy site. The U.S. Ninth Circuit Court of Appeals rejected playboy's appeal of the ruling on October 27, 1998. The plaintiff amended its initial complaint and added new complaints, however, its second attempt was dismissed in September 2000. The appeal of its second attempt was dismissed on February 1, 2002. A counterclaim by Terri Welles versus Playboy, the first of its kind against a trademark owner, over the latter's attempts to restrict her use of Playboy's trademarked items on her web site, seeking damages and a ruling against future suits being filed by Playboy, was dismissed by the U.S. District Court, San Diego, on January 4, 1999.

23. In Marianne Bihari and Bihari Interiors, Inc. v. Craig Ross and Yolanda Truglio [26] the defendants had hired Bihari, an interior designer, and her company and later sued in state court for breach of contract and fraud. In addition the defendants established a number of 'consumer' web sites on which they posted rather uncomplimentary statements about the plaintiffs and their performance on the job for the defendants. All of these websites used the plaintiffs' service mark, Bihari Interiors, in their meta tags. Plaintiffs claimed application of the Anticybersquatting Consumer Protection Act (ACPA), the Lanham Act and various state laws on libel seeking, among others, a preliminary injunction using the name Bihari or the service mark Bihari Interiors in the meta tags of their web sites.

24. The court found for defendants, as they were not likely to prevail under either ACPA or the Lanham Act as regards trademark infringement. To establish such a claim it was necessary for the plaintiffs to prove that the defendants had used the mark in commerce, in such a way that "it is likely to cause confusion ... as to the affiliation, connection, or association of such person with another person, or as to the origin, sponsorship, or approval of the [defendants'] goods, services or commercial activities by another person". The court held that the usage was not to warrant the application of the Lanham Act "because defendant's use of the 'Bihari Interiors' mark in the meta tag is not likely to cause confusion and is protected as a fair use." The court referred to the consideration that consumers who found the highly critical website were not likely to be confused as to the source of the site in question. The court also attempted a definition of fair use: "Fair use is established when the challenged term is a use, other than as a mark, ... of a term or device which is descriptive of and used fairly and in good faith only to describe the goods or services of such party..". One might note also that the court noted that using a website appeared the only way that the defendants could get their message to the public and that their website ran a disclaimer: "Keep in mind that this site reflects only the viewpoints and experiences of one Manhattan couple" which appeared as an indicator of good faith use of the service mark.

25. In Brookfield also, the appellate court recognized that fair use might give the right to use another one's trademark as a meta tag even in a commercial atmosphere. Here, the court drew the distinction between the use of 'movie buff' as two words, and 'moviebuff' as a single word. When used as two words, so the court remarked, the term is a descriptive phrase for people who are fans of movies, and the two-word phrase could be legitimately used in a meta tag. However, the single word would be nothing than the trademark itself, and would not be fairly usable as it referred to the trademark owner, and not to the products of the meta tag site.

26. Predating Brookfield, a Danish case, Melitta ctr. Coffilter Internationalt [27] provided similar guidance for fair use in a commercial context. Coffilter operated a website that made it possible to order paper coffee filters that fit the filter holders of Melitta coffee brewers. Coffilter used Melitta (tm) and Pyramiden (tm), both owned by Melitta as meta tags. After a challenge by Melitta, Coffilter removed these meta tags but instead of these single words used comparative statements such as "Attractively priced and packed Kvadrat Coffee Filters fitting the range of Melitta Pyramid Coffee Filter Brewer series.." This text was put on the first lines of the entrance page, ensuring that it would show up on a search engine's short presentation of the page.

27. The Danish court ruled, like it was suggested in Brookfield later, that the use of single keyword trademarks in meta tags constituted trademark infringement, however, that the comparative statements, which generated hits with search engines, were allowable use.

28. Significant is also that, beyond fair use, first amendment rights are at stake. In Bally Total Fitness v. Faber [28] the plaintiff attacked Faber, who ran a commercial website called "www.compupix.com" on which was also maintained a sub-domain called 'Bally sucks'.[29] In discussing the 'degree of care exercised' factor of its test for likely confusion, the court addressed the issue of the meta tags as follows:

    'Bally argues that that individual users may mistakenly access Faber's site rather than the official Bally site. Bally argues that this may happen when users employ an Internet search engine to locate Bally's site. Bally argues that the search result may list Faber's site and Bally' site. ...

    Here, Faber uses the Bally mark in the context of consumer criticism. ...

    He communicates [on the web page] that the site is unauthorized and that it is not Bally's official site. Moreover, Faber's use of the Bally mark does not significantly add to the large volume of information that the average user will have to sift through in performing an average Internet search. Whether the average user has to sift through 799 or 800 "hits" to find the official Bally site will not cause the frustration indicated in Teletech and Panavision because Faber is not using Bally's marks in the domain name. ...

    Further, the average Internet user may want to receive all the information available on Bally. The user may want to access the official Internet site to see how Bally sells itself. Likewise, the user may also want to be apprised of the opinions of others about Bally. This individual will be unable to locate sites containing outside commentary unless those sites include Bally's marks in the machine readable code upon which search engines rely. Prohibiting Faber from using Bally's name in the machine readable code would effectively isolate him from all but the most savvy of Internet users.' [30]

29. The court awarded summary judgment to Faber, holding that there was no likelihood of confusion, hence any infringement. On the dilution claim, it held that Faber did not use the Bally mark 'in commerce', even though he sold - unrelated - pictures from another part of the site. The court added significantly that under dilution theory

    'courts have held that trademark owners may not quash unauthorized use of the mark by a person expressing a point of view. This is so even if the opinion may come in the form of a commercial setting... [t]he Constitution does not... permit the range of the anti-dilution statute to encompass the unauthorized use of a trademark in a noncommercial setting such as an editorial or artistic context'.[31]

    Some Policy Conclusions

30. The landscape of the Internet has been transformed from a vast library into a sprawling mall. The broader implications of this transformation have been noted but are not yet fully appraised.[32] This article is highlighting one aspect of this transformation and the issues that arise in the context of meta tags.

31. Free accessibility to and free use of the Internet are not yet recognized human rights. Intellectual property ownership is, at present, often seen as granting absolute and unfettered rights. That view is apparently not shared by the courts, if meta tag litigation is a proxy, in a qualitative sense, for indicating limits on free searching, in particular as regards restraints imposed by intellectual property owners on meta tagging of names that are protected by their intellectual property rights.

32. The United States Appeals Court for the Second Circuit once compared attempts to formulate legal rules with respect to the Internet to boarding a moving bus.[33] It need not be put, perhaps, as starkly as this. There are solid foundations for anchoring a legal framework for freedom of access to Internet information, fair use and free speech, both in Europe and in the United States.

33. In the particular case analysis at hand, it would be wise to recall the words of an Olympian democrat, Oliver Wendell Holmes:

    '[W]hen the mark is used in a way that does not deceive the public we see no such sanctity in the word as to prevent its being used to tell the truth'.[34]

34. From the analysis above, one may infer that fair use of a trademark as a descriptive meta tag is allowed, where the user of the meta tag does not own the trademark, if it occurs in a non-commercial way. In addition, critical, satirical, artistic or editorial use may be made of a description or keyword meta tag.

35. In a commercial atmosphere, descriptive or comparative use of another one's trademark, as a meta tag, appears, from a growing and worldwide case law, allowed. Pagejacking or misleading use of trademarks in description and keyword meta tags is not.

36. There remains the tricky issue of keyword meta tags and spamming, which has not yet been addressed by the courts. The following may be relevant considerations on spamming. Keyword meta tags involve communication between web sites and search engines. These search engines have a commercial incentive to avoid spamdexing, as it annoys the user of the search engine. However, a keyword meta tag is not necessarily an advertisement or product or service label, but may be, as is clear from Welles, a legitimate request made to a search engine.

37. Html is an expressive language, and the Internet is still a library, and information exchange forum as well as, increasingly, a shopping mall. Underscoring a basic right to Internet information and access to its content, free use of language in keyword meta tags is a guarantee to authors, and even businesses, to reach Internet users. Search engines should therefore not, under the guise of avoiding spamdexing, impede free search and the ability to reach websites when there is no bad faith involved. Also, in general, trademark owners should not stand in the way of reaching websites in comparative, descriptive, artistic or critical and editorial use: there is no sanctity in the mark they own to prevent it from being reached, and used to tell, freely, the truth as one sees it.