METATAGS TO ADWORDS: A POOR ANALOGY

Metatags to Adwords: A Poor Analogy?

Christopher St. Pierre

I. Introduction

The web search industry is a driving force behind the explosive growth of the Internet and online businesses. Internet search engines, like Google, connect end users with the products, both the information and the commercial products, which the end users are seeking quickly and easily. Were it simply the charitable connection of searchers with what they seek, the costs of running a major Internet search engine would quickly put them out of business. Advertising is the revenue stream that keeps the search industry in business. Search engines allow advertisers to directly target the advertising to the desires of consumers by virtue of the fact that consumers are telling the search engine what they are looking for. Acting in this role, connecting consumers with a product’s source or its seller, search engines have important implications in the context of trademark infringement.

A. Search Engine Advertising and AdWords

Each of the various search engines presents advertising content on the search results page based upon terms used in a given search. Advertisers purchase keywords and depending on the correspondence of advertiser-purchased keywords and the user-selected search terms, appropriate ads are selected and displayed in the search results. The high level operations of all the various keyword-sponsored advertising systems in use in the search industry are roughly equivalent. Given this fact, and the fact that Google is one of the leaders in the search industry, this paper will be limited to an analysis of trademark infringement liability as related to Google’s sponsored-keyword advertising system, AdWords.

Advertisers that wish to advertise on Google have a great degree of flexibility in building their ad campaign[1]. An advertiser chooses keywords they believe are best suited to trigger the display of their content and the text that they wish to display in the advertisement. Additionally, the advertiser selects the maximum price they are willing to pay for each user who clicks on the hyperlink in their advertisement. Finally, the advertiser selects their maximum daily budget. Google’s AdWords system, with some knowledge of the statistical click-through rate for given search terms and keywords, will present enough ad impressions to satisfy the advertiser’s daily budget. If there is a conflict of advertiser-sponsored-keywords, as there often is, the advertiser paying the most per click and having highest ad budget will inevitably get the most advertising impressions.[2]

B. Search Engines and MetaTags

The specific mechanisms that make Internet searching work are a closely held secret of the search engine industry. At a high level, search engines go to each website that they are aware of and download the content of the website to analyze and index that content. The index that the search engine builds from this “spidering” enables web searchers to find a given website. One tool that website developers can use to inform search engines what the purpose or content of the website entails is metatags. Metatags are bits of textual data, hidden from display on the web page, that include keywords. The content provider chooses those keywords based upon the search terms she thinks should trigger the display of the website in search results. The intent of metatags is to enable the same type of association between search terms and content that AdWords sponsored-keyword advertising provides. The key difference is that metatags are free, in contrast with the pay per click used in sponsored-keyword advertising. Additionally, the content provider maintains sole control over metatags where in sponsored-keyword advertising, that control is split between the advertising system operator and the content provider.

C. Extending the MetaTags Precedents to AdWords

An open question in the sponsored-keyword advertising context is whether it is trademark infringement when an advertiser using the Google AdWords system selects the trademark of another to use either as a sponsored keyword or within the text of their ad. Beyond the question of infringement is the question of who is liable: the advertiser, or Google, or both. Different circuits have ruled differently on both of these questions. At first blush, the issue of whether an advertiser is liable for trademark infringement for choosing a competitor’s mark as their advertising keyword seems analogous to when a company uses a competitor’s mark as a metatag on their website. However, the consequences for the Internet search industry of such a finding are broad reaching and unjustified. If the analogy of sponsored-keyword advertising to metatag infringement becomes the norm, it is not a far stretch to find search engine liability for the infringement of advertisers.

II. Trademark Background and Theories of Infringement

Trademark law has medieval roots. [3]Craftsmen would mark their products so that consumers would be able to identify the maker of a given commodity. At common law, the principal user of a mark could bring an action for deceit when someone else used the same mark to sell goods not manufactured by principal user of the mark. A distinct action, sounding in the tort of fraud, known as “passing off”, developed over time.[4] Congress enacted the first trademark statute in 1870, and the current trademark statute, the Lanham Act, was enacted in 1946[5].

A. Direct Infringement

A finding of trademark infringement requires that the infringer used a confusingly similar mark in commerce[6]. For the purposes of infringement, the mark’s use in commerce depends upon whether the accused infringer is using the mark in the course of selling products or services[7]. Additionally, the use of the mark must be a “trademark use”, that is the alleged infringer must be using the mark in the manner that trademarks are intended to be used: as a source identifier[8]. Finally, courts have required that there be some likelihood of consumer confusion[9]. Because of the difficulty in proving actual consumer confusion, the various circuit Courts have adopted a likelihood of confusion test to determine whether the accused infringing mark is confusingly similar to trademark holder’s mark. The likelihood of confusion test varies from circuit to circuit but typically includes elements similar to the elements first set forth by the Second Circuit in Polaroid Corporation v. Polarad Electronics Corporation[10]. These factors include

the strength of [a prior owner's] mark, the degree of similarity between the two marks, the proximity of the products, the likelihood that the prior owner will bridge the gap, actual confusion, and the reciprocal of defendant's good faith in adopting its own mark, the quality of defendant's product, and the sophistication of the buyers[11].

If, on the balance of the evidence and considering the factors, the Court thinks that the accused mark is confusingly similar, it will find that the defendant infringed the plaintiff’s mark[12].

The general policies underlying trademark law pertain primarily to the interest of the public in not being deceived by a misleading or confusing use of a mark in commerce[13]. Secondarily, trademark law seeks to prevent others from unjustly enriching themselves by using another’s mark. Finally, businesses expend large sums of money developing the mark and goodwill surrounding the mark; trademark law recognizes and protects this investment[14].

B. Contributory Infringement

The Lanham Act has provisions protecting a mark holder against the direct infringment of a competitor[15]. However, no section of the Lanham Act addresses the liability of an entity contributing to the direct infringement of a third party. Additionally, there is some difficulty applying the traditional likelihood of confusion factors to a party who enables the trademark infringement of a third party[16]. The United States Supreme Court recognized this deficiency and established the test used to determine the contributory liability of a party for the infringement of a third party in Inwood Laboratories, Inc. v. Ives Laboratories, Inc[17]. The test established by the Inwood Court states that a defendant may be adjudicated as a contributory infringer when “a manufacturer or distributor intentionally induces another to infringe a trademark, or if it continues to supply its product to one whom it knows or has reason to know is engaging in trademark infringement, the manufacturer or distributor is contributorially responsible for any harm done as a result of the deceit.”[18]

Some Courts have extended the rule established in Inwood beyond the domain of manufacturers and distributors[19]. The basic requirement of Courts finding contributory liability outside of the manufacturer/distributor context is that the alleged contributory infringer either must know or should have reason to know of the infringing activities of the direct infringer[20]. The duty to police the infringement of third parties has been limited to only circumstances where it is reasonable. While it may be reasonable to impose a duty to police a flea market for infringers upon the operator of a flea market, it is not necessarily reasonable (given the scope of the Internet) to impose a duty to police domain name registration upon a domain name registrar[21].

C. Initial Interest Confusion

In the Internet context, an additional theory of liability has been adopted, or at least cited approvingly, by several circuits. This theory of liability is known as the initial interest confusion doctrine[22]. The initial interest confusion doctrine developed from a decision of the Ninth Circuit, Brookfield Communications, Inc. v. West Coast Entertainment Corp[23]. Traditionally, the requirement was that in order for the use of a mark to be an infringing use, the use had to be likely to confuse the consumer. The initial interest confusion doctrine finds the use of a mark an infringing use, even absent any likelihood of confusion, if the use of the mark captures the initial interest of a consumer[24]. The analogy the Court used to explicate the doctrine is as follows:

Suppose West Coast's competitor (let's call it "Blockbuster") puts up a billboard on a highway reading - "West Coast Video: 2 miles ahead at Exit 7" - where West Coast is really located at Exit 8 but Blockbuster is located at Exit 7. Customers looking for West Coast's store will pull off at Exit 7 and drive around looking for it. Unable to locate West Coast, but seeing the Blockbuster store right by the highway entrance, they may simply rent there. Even consumers who prefer West Coast may find it not worth the trouble to continue searching for West Coast since there is a Blockbuster right there. Customers are not confused in the narrow sense: they are fully aware that they are purchasing from Blockbuster and they have no reason to believe that Blockbuster is related to, or in any way sponsored by, West Coast. Nevertheless, the fact that there is only initial consumer confusion does not alter the fact that Blockbuster would be misappropriating West Coast's acquired goodwill.[25]

The doctrine evolved out the perceived difficulty to apply the standard likelihood of confusion factors to Internet cases. Some courts use initial interest confusion as a substitute for a full likelihood of confusion analysis, but others uses it in place of actual confusion in the likelihood of confusion elements. Consequently, the use of the trademark of a competitor to capture the “initial interest” of the consumer has been found to be an infringing use because it captures the goodwill of the mark holder.[26]

D. Defenses

There are several defenses to a claim of trademark infringement, but the most apropos, for the purposes of this paper, is nominative fair use. The defense of nominative fair use was first recognized by the Ninth Circuit in New Kids on the Block v. New America Publishing, Inc.[27] This defense is applicable when a defendant uses the plaintiff’s mark in commerce, not for the purposes of identifying the defendant’s goods but rather for identifying the plaintiff’s goods[28]. The test the Ninth Circuit laid out for determining whether the use of a trademark was a nominative fair use includes the following:

First, the [plaintiff’s] product or service in question must be one not readily identifiable without use of the trademark; second, only so much of the mark or marks may be used as is reasonably necessary to identify the plaintiff's product or service; and third, the user must do nothing that would, in conjunction with the mark, suggest sponsorship or endorsement by the trademark holder.[29]

Nominative fair use comes into play in several contexts, not the least of which is comparative advertising. It would be very difficult for an company to compare their product to their competitors’ products without the ability to identify their competitor’s product using the mark by which it is known.[30]

III. Trademark Infringement in the Metatags Context

The factual situation presented in most metatags trademark infringement cases involves a defendant who uses the trademark of a competitor company as a metatag on the defendant’s website. Usually, the mark used by the defendant is the exact mark of the plaintiff, not merely a similar mark. The trademark infringement claim is based upon the fact that the mark is being used in commerce to attract users to the website of a competitor. Plaintiff’s either assert that there is a likelihood of confusion because consumers are confused as to source of the goods they find on the defendant’s website, or they assert that their trademark has been infringed through the defendant’s capturing of the initial interest of consumers. For the most part, current metatags cases involve a trademark holder suing a direct infringer, a defendant who directly used the plaintiff’s mark in commerce.

Plaintiffs relying on likelihood of confusion to prove trademark infringement have had difficulties proving that any consumer confusion that may have arisen from searching for the plaintiff’s mark and subsequently being directed to a competitor’s website is not dispelled by the time of consummating any commercial transaction.[31] Consequently, a majority of the metatags cases are prosecuted based upon an initial interest confusion theory.

A. Direct Infringement

Often, the elements of the likelihood of confusion test that are of import to a metatags trademark infringement action are “the similarity of the marks, the defendant's intent, and evidence of actual confusion.”[32] In Promatek Indus., LTD v. Equitrac Corp., both Promatek and Equitrac were competitors in the cost-recovery business. Promatek was the holder of a trademark, Copitrak, which Equitrac used as a metatag on its website. The Court found that several factors of the likelihood of confusion test adopted by the Seventh Circuit weighed against Equitrac. Because the marks were similar, the use of the mark by Equitrac referred to Promatek’s registered trademark, and because of the direct competitor relationship of the two parties, the Court held there was a “strong likelihood of consumer confusion.”[33] The Court goes beyond a finding that there was sufficient evidence to allow a preliminary injunction under a likelihood of confusion theory. In addition, the Court found that Equitrac could be held liable under an initial interest confusion theory. The Court reasoned that the degree of care exercised by consumers, in conjunction with the strong likelihood of confusion, could lead to initial interest confusion.[34] One consideration that the Court gives short shrift is the fact that Equitrak repairs the equipment referred to by the trademark in question, Copitrak.[35] The Court allows that Equitrak may advertise that it services Promatek’s Copitrak equipment, but Equitrac may not succeed in this case because “it used that trademark in a way calculated to deceive consumers into thinking that Equitrac was Promatek.”[36]

In Tdata Inc. v. Aircraft Tech. Publrs., the District Court for the Southern District of Ohio used a more traditional likelihood of confusion analysis in conjunction with an initial interest confusion analysis to find that Tdata did indeed infringe Aircraft Technical Publishers’ (ATP’s) trademarks by using them as metatags.[37] The Sixth Circuit has adopted the initial interest confusion doctrine, but unlike other circuits, the Sixth Circuit has applied the initial interest confusion doctrine outside of the Internet context.[38] Outsider of the Internet context, it uses the initial interest confusion as a substitute for actual confusion in the traditional likelihood of confusion test.[39] Despite the fact that Tdata is a traditional Internet case, the District Court applies initial interest of confusion in place of actual confusion in its likelihood of confusion analysis. On the question of whether there was any initial interest confusion, the Court finds that

Tdata's use of ATP's mark--use resulting from the affirmative act of including the mark as a metatag--can only serve to bring to Tdata's website potential customers, some of whom might never have gone there but for use of ATP's mark. The Court thus agrees with ATP that use of the company's mark in metatags constitutes infringing use of the mark to pull consumers to Tdata's website and the products it features, even if the consumers later realize the confusion.[40]

The remainder of the Court’s analysis focuses on the Sixth Circuit’s more traditional likelihood of confusion test. According to the Court, the weight of the evidence on each of these factors weighs against Tdata.[41] Tdata claims that regardless of the likelihood of confusion, its use of the ATP’s marks as metatags was a nominative fair use of those marks. Because the mark was used as a metatag rather than as another type of information on Tdata’s website, the Court found that “the hidden-from-public-eye use of the mark lends itself to an inference disfavoring Tdata.”[42]

B. Contributory Liability – Metatags to Keyword Advertising

There are no cases directly on point where a defendant was found contributorially liable for the metatags used by a third party. This is largely because the third party is in control of the metatag content and the burden would be too great on search engine operators to police the entire Internet for metatag trademark infringement. However, using the metatag trademark infringement paradigm, and focusing on initial interest confusion, courts have found that operators of keyword advertising systems operated in conjunction with Internet search engines are contributorially liable for trademark infringement.

In Playboy Enters. v. Netscape Communs Corp., the Court analogized Netscape’s advertising system to the use of metatags in Brookfield.[43] In Playboy, the Ninth Circuit found that Netscape could be liable under either a direct infringement or a contributory infringement theory.[44] Netscape operated a keyword linked banner advertising service in conjunction with its search engine. Banner ads were displayed in the search results that corresponded to the terms a search engine user entered as their search terms. Rather than give advertisers unrestricted freedom to choose whichever keyword they may desire, Netscape maintained lists of keywords corresponding to the type of advertiser. Adult oriented advertisers were given a choice of some 400 keywords, and they were required to select their keywords from that list. Among this list of keywords are two keywords for which Playboy (PEI) has trademarks: “playmate” and “playboy.” The consequence of this is that if an adult advertiser selected either of these trademarked terms as their keywords, when a searcher used either “playmate” or “playboy” in their search terms a banner ad would be displayed and if the searcher clicked on that banner ad they would be directed to the advertiser’s website.[45]

The Ninth Circuit found that such use of the trademarked keywords was likely to create an initial interest confusion, but to be certain that the weight of the evidence was sufficient to overturn the summary judgment for Netscape in the lower Court, the Court goes through the eight factor likelihood of confusion test used by the Ninth Circuit.[46] However, because the Court weighs each of the factors in light of the prejudged initial interest confusion it is unsurprising that they find the weight of the evidence in Playboy’s favor. Netscape claimed several defenses, including traditional fair use and nominative fair use.[47] The Court disposes of the nominative fair use defense because the Court finds that there is no need to use Playboy’s marks in the manner in which Netscape is using them. “Defendants do not wish to identify PEI or its products when they key banner advertisements to PEI's marks. Rather, they wish to identify consumers who are interested in adult-oriented entertainment so they can draw them to competitors' websites.”[48]

On the question of whether the defendant is directly liable or contributorially liable, the Court does not come to a specific conclusion[49]. The Court finds that the defendant could be liable under either theory, and consequently the Court had no need to decide the issue. However, this fails to address the issue of “use in commerce” versus “trademark use.” Playboy and Netscape are not even ancillarily in the same trade or business. Therefore, therefore there is little likelihood of consumer confusion as to the source of a product as between Netscape and Playboy. Additionally, Netscape was not making a “trademark use” of the keywords. Consequently, the case against Netscape is much stronger under a contributory liability theory. There are direct predicate offenders, those adult content providers who chose “playboy” and “playmate” as their keywords, and because Netscape provided the list of available keywords and maintained that list in its sole discretion, Netscape clearly knew or should have known of the infringing activity of those advertisers choosing those keywords. Unfortunately, the Court punts on this particular liability question early on and the resulting analysis completely ignores the fact that Netscape and Playboy are not competitors and the case for a use in commerce and trademark use, as required by the Lanham Act, is weak.

IV. Trademark Infringement in the AdWords Context

To some degree, cases involving trademark infringement in the AdWords context have followed the metatags precedents. Choosing a keyword for use in the AdWords system has the same consequences as choosing a keyword as a metatag. When a search engine user uses a keyword in a search they will be presented with an advertisement on the search results page corresponding to the keyword. This advertisement is analogous to a link in the search results that is returned due to a correspondence between search terms and metatags. A key distinction is that in the AdWords context both advertiser who choses the term and Google have a monetary interest. The advertiser makes money if a searcher clicks on the advertisement link and subsequently purchases a product, and Google makes money when the searcher clicks on the link. Courts have split on the liability of Google and the liability of an advertiser who chooses a competitors trademark as its keyword for use in the AdWords system.

A. Direct Infringement

While there is some case law that has allowed several cases brought by the holders of registered marks whose marks where purchased as AdWords keywords to proceed, there is no case finding actual liability for such a use. It seems likely that under the initial interest confusion doctrine elucidated by the Ninth Circuit, and adopted in various other circuits, an advertiser may potentially be found liable for trademark infringement through capturing the initial interest of a consumer.

Some Courts have established that purchasing a competitor’s trademark as a keyword is sufficient use in commerce to sustain a Lanham Act trademark infringement claim.[50] The defendant in Buying for the Home, LLC v. Humble Abode, LLC allegedly purchased the phrase “total bedroom” as its AdWords keyword. Additionally, it was alleged that the keyword would trigger commercial advertising including a link to the defendant’s website. Therefore, because the plaintiff’s mark was connected to the promotion of the defendant’s goods and the mark was used to connect a user to the defendant’s website to make purchases of the defendant’s goods, this use was found sufficient to satisfy the Lanham Act’s use requirement.[51]

In a recent case out of the Seventh Circuit, Int'l Profit Assocs. v. Paisola, the Court found sufficient evidence to proceed under a claim of trademark infringement for the use of trademark as a sponsored keyword in the AdWords system.[52] In this case, the defendant purchased the sponsored keyword the phrase “International Profit Associates,” a registered trademark of the plaintiff. The defendant was competitor and the websites linked to the plaintiff’s trademark through AdWords contained material that was disparaging in nature. The Court relied on the Buying for the Home decision to establish the defendant’s use as sufficient to sustain a claim for trademark infringement under the Lanham Act.[53] Additionally, relying on the affidavits of the plaintiff, the Court found sufficient evidence of a likelihood of confusion due the similarity of the domain names chosen by the defendant in conjunction with the purchase of the trademarked term to issue a temporary restraining order against the defendant on the trademark infringement claim.[54]

Some plaintiffs have tried to proceed against Google on a direct trademark infringement claim based upon Google’s sale of their registered trademark as an AdWords keyword.[55] Rescuecom, a computer repair and recovery business, holds a registered trademark for the term “Rescuecom.” Rescuecom’s competitors have used Rescuecom’s trademark as their sponsored-keyword, and Rescuecom contends that Google is at fault because its keyword suggestion tool recommends Rescuecom as a keyword to Rescuecom’s competitors.[56] Google moved to dismiss the case because its use of the term Rescuecom was not a sufficient use for trademark infringement liability under the Lanham Act. Because the Second Circuit considers trademark use, likelihood of confusion, and use in commerce as three distinct elements of an infringement claim, it declined to follow the precedents of other Courts which tied the use requirement to allegations of a likelihood of confusion or use in commerce. The plaintiff argues that Google’s use of its trademark is use within the requirements of the Lanham Act on several grounds. The Court finds that even if the plaintiff can prove that the defendant was capitalizing on the goodwill of the plaintiff, or that consumers believed that the defendant was authorized to sell the plaintiff’s mark, none of these factors can show trademark use as required by the Lanham Act. While they may be sufficient evidence to make a showing of use in commerce or likelihood of confusion, they are insufficient to prove the trademark use requirement of the Lanham Act. The Court concludes that

in the absence of allegations that defendant placed plaintiff's trademark on any goods, displays, containers, or advertisements, or used plaintiff's trademark in any way that indicates source or origin, plaintiff can prove no facts in support of its claim which would demonstrate trademark use.[57]

Consequently, the defendant’s motion to dismiss was granted because there was no use of the defendant’s trademark as a trademark.

B. Contributory Infringement

The question of whether Google may be contributorially liable for the direct infringement of the advertisers using its AdWords system remains an unanswered question. Key concerns in this form of liability are whether Google or its advertisers made a “trademark use” of the mark of a third party. A concern rarely brought into focus by the Courts that have heard these cases is the question of whether Google knows or should know about the infringement of its advertisers. Additionally, the extent of the duty to know about the infringement of the advertisers has not been addressed.

One of the requirements to establish contributory trademark infringement is the direct infringement of a predicate infringer. In Google Inc. v. Am. Blind & Wallpaper Factory, Inc., American Blind & Wallpaper Factory alleged that Google was liable contributorially for the infringement of advertisers using the AdWords system.[58] American Blinds uses several marks to sell its products, some registered, and some unregistered. The registered marks include the terms “American Blind & Wallpaper Factory,” “American Blind Factory,” and “DecorateToday.” In addition, American Blinds has used the phrases “American Blind” and “American Blinds” in conjunction with its services.[59] American Blinds alleged that, through its AdWords program, Google has sold many of the terms that American Blinds used as marks and has continued to do so over American Blinds’ objections. American Blinds further alleged that Google actively suggested, through its keywords optimization system, that users who chose some combination of “American Blind” as their keywords also purchase such keywords as “american blinds” and “american blinds and wallpaper.” Finally, because Google had the ability to monitor and control the use trademarks not associated with the advertiser, and they had in the past done so, it was in a position to know about and control the infringement of its advertisers. In summary, American Blinds alleged that Google was selling the possibility or intercepting traffic intended for American Blinds to their competitors.[60] Google claimed that its use of the terms in question was not a “trademark use” in that those terms were not being used to identify goods or services sold by Google. Additionally, Google claimed that because the advertisers were only using the American Blind marks as triggers to display their ads and not as a source identifier, there is no “trademark use” of the marks, and consequently Google cannot be liable.[61] Because of the precedents in other circuits regarding what constitutes “trademark use” and the sufficiency of the allegations of direct infringement against American Blinds’ competitors, the Court found that American Blinds had established a sufficient facts to survive a motion to dismiss.[62]

In at least one case, Gov't Emples. Ins. Co. v. Google, Inc.. Google was found not liable for the use of trademarks as keywords in the AdWords system.[63] Initially, Google moved to dismiss based upon its claim that it had not made a “trademark use” of the keyword in question, “GEICO.”[64] Google argued that its use of the mark at issue was simply for internal coding and association of ads to search terms. GEICO argued that Google used the mark to sell advertising and as such, it was a “trademark use.” Relying on precedents from other circuits, the Court found GEICO’s argument more persuasive and ruled against Google on its motion to dismiss finding that Google’s use of the plaintiff’s trademark was a “trademark use.”[65] At trial on the merits, GEICO presented survey evidence pertaining to consumer confusion.[66] At the close of the plaintiff’s case, Google moved for a judgment as a matter of law based upon the evidence on the record. The Court relied heavily on survey data, and Google’s questioning of the validity of that survey data in coming to its decision. The Court ruled for Google on the question of whether the use of trademarks as keywords was trademark infringement because the survey data was flawed as related to ads that were triggered by keywords but did not contain GEICO’s mark in the text of the ad. The Court ruled against Google, finding sufficient evidence of confusion from the survey data in relation to ads that contained GEICO’s mark in the text of the ad. Subsequently Google and GEICO settled the remaining claims.

V. Failings of Traditional Infringement Liability as Applied to Search Engine Advertising

Any claim of direct infringement against a search engine operating a sponsored-keyword advertising system is a difficult claim to make. Firstly, the search engines use such sponsored-keywords “under the hood” to make an association between consumer search terms and advertisements. Search engines do not make a traditional “trademark use” of those terms because search engines do not use the marks as source identifiers for any product or service sold by the search engine. Additionally, the search engine did not create any confusion or even initial confusion about the source of given goods. Rather, the advertiser using the search engine’s keyword advertising system chose their keywords and consequently any confusion falls on their shoulders. Advertisers, on the other hand, seem to be exposed to the same level of liability as under prior metatags cases.

A more difficult case for search engines operating keyword advertising systems sounds in a theory of contributory infringement. The predicate infringement liability of advertisers, under recent precedents, especially given the increasing adoption of the initial interest confusion doctrine, seems possible to establish. Given established predicate liability, a plaintiff suing a search engine operating a keyword advertising system must show that the search engine operator knew or should have known of infringement by advertisers using their system. Additionally, the plaintiff must show that the search engine operators should reasonably be assigned the duty of policing the use of trademarks by their advertisers. Currently, there are no cases which give full weight to all of the factors required to find contributory liability in the AdWords context.

There are several problems with the current case law surrounding the liability of search engines. Many Courts have continued to conflate “trademark use” with “use in commerce.” In the contributory infringement cases, the Courts have given little weight to the reasonableness of search engines policing their system for the trademark infringement of their advertisers. Additionally, the metatags and AdWords cases have deprecated nominative fair use and have made it much more difficult to engage in comparative advertising due to the weight given to “initial interest confusion” in the infringement analysis.

A. Trademark Use

A minority of Courts have interpreted the use requirement of the Lanham Act to mean a “trademark use.”[67] In other words, for a trademark infringement claim to succeed, the use of the mark by the defendant must be a use in the way a trademark is intended to be used, as a source identifier of goods. Typically, these Courts recognize three distinct inquiries in the determination of trademark infringement. Was the given use a trademark use? Was there a likelihood of confusion? Was there a use in commerce? This breakdown best preserves the policy underlying trademark protection. By requiring that the use be a “trademark use,” the policy of protecting source identification is served. Additionally, by requiring the use to be “trademark use” the rights given to trademark holder are balanced against First Amendment interests. Only that use which we are seeking to protect, use as a source identifier, is removed from the public domain. Finally, by analyzing the “trademark use” distinctly from the “use in commerce,” non-commercial uses of a trademark as a trademark are protected.

The majority of Courts have not been careful about distinguishing between “trademark use” and “use in commerce.” (See generally American Blinds & GEICO). The consequence of the blurring of this distinction is that many uses of trademarks that are permissible considering the policies underlying trademark law are no longer available. Taken in consideration of a recent trend in the Courts to give greater weight to initial interest confusion in the Internet context, it is increasingly likely that trademark infringement will be found for the use of mark even when the use is not as a source identifier.

B. Reasonableness of Policing the Marketplace

Liability for contributory infringement is based upon whether the alleged contributory infringer knew or should have known of the activities of direct infringers whose infringement the contributory infringer is alleged to have contributed to. In the AdWords context, the allegation is that Google knew or should have known about the infringement of the advertisers using its system. No case has gone beyond finding that it is possible for Google to be contributorially liable for the infringement of its advertisers. GEICO was settled after the partial denial of Google’s motion for a judgment as a matter of law, and American Blinds simply found that there was sufficient evidence to support a claim of contributory infringement to survive a motion to dismiss. Therefore, the question of the extent of Google’s duty to police the AdWords marketplace for infringement has not been litigated. However, it seems that any “duty to know” about the infringement of advertisers must surely be a limited duty on Google’s part.[68] Google does maintain control of the keyword and ad content for the ads displayed in its AdWords system but how far must Google go to validate that a term or ad text purchased by an advertiser is either descriptive of the advertiser’s services or registered to the advertiser? Must Google do a trademark search and a detailed check into the business of its advertisers? It will be up to future Courts to decide the scope of this duty, but due to the burden that would be placed on such keyword advertising systems to fully police their advertisers for infringement, it seems the duty should be minimal.

C. Deprecating Nominative Fair Use

When the initial interest confusion doctrine is applied as it is in a majority of Courts, many nominal fair uses are seen as infringing uses. If Manufacturer A makes a Product A that is designed to compete directly with a Product B from Manufacturer B, it seems logical that Manufacturer A would include Product B as a metatag on their website. From Manufacturer A’s point of view, he is engaging in legal comparative advertising. Adding the competitors mark as a metatag is saying “Hey if you like their Product B, you’ll really like our Product A.” There is no way to tell the search engine Product A is like Product B short of including Product B as a metatag on the website for Product A. However, Courts that follow the Brookfield standard would find such a use infringing because when a consumer searches on a search engine for Product B mark he may be directed to the website for Product A, and such delivery captures some of the goodwill associated with Product B. How else is a Manufacturer A supposed to signal to the search engine that their Product A is like Product B? Because the metatags analysis found in Brookfield is being carried over into the AdWords context, the same problem occurs in that context. In that scenario, how can Manufacturer A advertise its Product A as a competitor to Product B without naming Product B, or using Product B as a keyword? These questions were given perfunctory consideration if they were considered at all by the Courts considering the issue.[69]

From the consumer’s perspective, there are additional problems with initial interest confusion. For example, a consumer searching for a new television may well choose an industry exemplar brand, e.g., Sony, for their search term even though they are not really looking for a Sony television. The purpose of choosing the exemplar brand as a search term is to find goods that are comparable to that exemplar. Is it unreasonable for a consumer to do this? When making all kinds of purchasing decisions, consumers choose what the wish they could buy and then compare the features of what they can afford to what they wish they could buy. Is an advertiser who enables this type of comparison shopping by holding their products out against the exemplar products infringing the trademarks of those products? Brookfield seems to support that conclusion.

Additionally, the analogy chosen by the Brookfield Court is flawed. Searching on the Internet is not analogous to driving down the highway and getting off on the wrong exit. The back button on any web browser is much easier to hit than getting back on the highway. In the internet context, except for in cases of truly deceptive advertising and website design, it is obvious when you get to a site you weren’t looking for. And the easy answer is to hit the back button, not stick around and do some shopping.

VI. Conclusions

The decline of a requirement of “trademark use” in favor or a more general use or commercial use requirement in conjunction with a corresponding increase in the application of the initial interest confusion doctrine has broadened the scope of the rights secured by trademark and increased the exposure to liability of advertisers who use the trademark of a competitor in Internet advertising. Additionally, without some clarification as to the duty of a sponsored-keyword advertising provider to police their marketplace, contributory liability may be found for such providers.