Trademarks and Consumer Search Costs on the Internet

Stacey L. Dogan* & Mark A. Lemley**+

 
In theory, trademarks serve as information tools, by conveying product information through convenient, identifiable symbols. In practice, however, trademarks have increasingly been used to obstruct the flow of information about competing products and services. In the online context, in particular, some courts have recently allowed trademark holders to block uses of their marks that would never have raised an eyebrow in a brick-and-mortar setting – uses that increase, rather than diminish, the flow of truthful, relevant information to consumers. These courts have stretched trademark doctrine on more than one dimension, both by expanding the concept of actionable “confusion” and by broadening the classes of people who can face legal responsibility for that confusion. And they have based their decisions not on the normative goals of trademark law, but on unexplored instincts and tenuous presumptions about consumer expectations and practices on the Internet. We argue that this expansionist trend in Internet trademark cases threatens to undermine a central goal of the Lanham Act – to promote fair and robust competition through reducing consumer search costs.
 

Introduction

Trademark law, in theory, fosters the flow of information in markets. By protecting against deceptive uses of trade symbols in commerce, the law enables sellers to create their own reliable shorthand to identify their goods, and reduces search costs for consumers.[1] Trademarks thus have the potential to lead to better-informed customers and more competitive markets. In practice, however, overreaching trademark law can have exactly the opposite effect. If trademark holders were allowed, say, to prevent the use of their marks to critique the trademark holders’ products or compare them to others, trademarks would become tools for suppressing information that is critical to a functioning market. To avoid this result, courts and Congress have historically insisted on clear substantive rules in trademark cases. Many trademark rules – the requirement that a defendant use a mark in connection with goods or services to infringe,[2] the insistence on likelihood of confusion,[3] and the exemption for truthful comparative advertising[4] – arose in part from attempts to make trademark law serve, rather than impede, the free flow of non-deceptive information.[5]

A recent series of Internet-related trademark cases has seriously eroded these substantive requirements, to the point at which trademark law’s anticompetitive tendencies threaten to overwhelm its information-facilitating function in some contexts. The expansions have occurred on two axes. First, a number of courts have broadened the class of people who can face legal responsibility for trademark infringement. Historically, to face liability under trademark law, a party had to “use” a mark as a brand in connection with the offering of goods and services, typically in competition with the trademark holder.[6] A party could face contributory liability for another’s infringing use of a mark only in the rare case when it “intentionally induce[d] another to infringe a trademark,” or “continue[d] to supply its product to one whom it kn[ew] or ha[d] reason to know [wa]s engaging in trademark infringement….”[7] But a handful of courts have lately dispensed with these limiting rules and endorsed liability against parties that neither used the trademark as a brand for their own products, nor satisfied the rigorous standard for contributory liability. In particular, two recent decisions held that Internet intermediaries could face liability for using trademarks to help advertisers identify consumers interested in the trademark holder’s products. These decisions – Playboy Enterprises, Inc. v. Netscape Communications Corp.[8] and 1-800 Contacts, Inc. v. WhenU.com[9] – either created a new, broader species of contributory trademark infringement, or applied the doctrine of direct infringement by broadening the meaning of trademark “use” to include parties that did not even arguably offer their own products or services under the mark. Either way, these courts broke new and troubling ground.

Compounding the effects of this first expansion, courts have increasingly shifted the focus of infringement analysis away from consumer confusion and toward a more generalized inquiry into whether a challenged use diverts attention away from the trademark holder. “Initial interest confusion,” which historically referred to a form of consumer confusion that occurred before the point of sale, has morphed into a standalone doctrine whose criteria bear little relationship to a traditional likelihood of confusion claim. Some courts have used the initial interest confusion doctrine to justify claims against virtually any use that temporarily diverts customers to a website not authorized by the trademark holder, without regard to whether the diversion resulted from confusion or harmed consumer interests in any way.[10] The courts’ adoption of the “initial interest confusion” doctrine as a replacement for proof of likelihood of confusion, along with a related conclusion that consumer frustration can cause dilution of a famous mark,[11] has allowed trademark holders to enjoin competitors from using marks in metatags, domain names, or other contexts that might conceivably result in a consumer mistakenly reaching a website not authorized by the trademark holder.[12] Whether one agrees with the outcome of these decisions or not, they undoubtedly chart new territory in trademark law and expand the kinds of trademark usage that can subject a defendant to an infringement claim. And while the initial interest confusion doctrine at first reached only competitors who were clearly attempting to confuse consumers into reaching the wrong sites,[13] over time, the doctrine has been invoked to prevent non-competitive uses as well as uses that appeared more likely to enlighten than to confuse.[14]

We argue that, cumulatively, these two axes of trademark expansion pose a grave danger to the law’s information-facilitating goals. The extension of trademark law to search engines, directories, and other parties that use marks as classification tools poses a grave threat to speech and to the dissemination of truthful information. By suggesting that Internet intermediaries have a responsibility to police the content of their advertisers and cannot use marks as keywords to identify logically related products, Playboy v. Netscape and its ilk threaten to chill a vast sector of informative speech, particularly (but not solely) in the Internet context. This chilling effect occurs both directly – by restricting the dissemination of truthful comparative advertising – and indirectly – by eliminating the primary source of revenue that has supported the development of sophisticated and accurate search engine technology.

The chilling effect grows particularly acute as confusion and deception become less central to trademark infringement claims. Even if intermediaries faced liability only for blatantly deceptive uses of trademarks by their advertisers, the costs of policing all of their advertisers’ copy for such content would be prohibitive and unprecedented. But as trademark law increasingly condemns behavior that diverts attention but does not necessarily deceive, search engines have no practicable means of distinguishing between legitimate and unlawful uses of marks. They can avoid liability in such a system only by requiring trademark holders’ authorization for every single use of their marks – in branding, as classification, or otherwise. This expansive approach to trademark law – which arguably follows from recent interpretations of the initial interest confusion doctrine – conflicts with the longstanding rule in favor of comparative advertising and the normative goal of producing better-informed consumers. Ironically, under this restrictive legal approach the challenge of lowering barriers to entry in markets dominated by a single brand could well prove harder in the age of the Internet than in an era in which vendors could simply place like products beside one another on store shelves.

Each of the decisions that have distended trademark law in these ways rested at least partly on an instinct that the law should prevent parties from profiting from the use of another party’s mark.[15] In analyzing Netscape’s good faith in the Playboy case, for example, the Ninth Circuit emphasized that Netscape had profited when its advertisers used Playboy’s trademarks to call attention to their competing products.[16] The court’s overall analysis implied that there was something unsavory about benefiting financially from the goodwill of an established trademark. But “unfair competition” is not redundant.[17] This unjust enrichment instinct runs counter to the core values of trademark law, which make the value of fair competition paramount.[18] If carried to its logical extreme, this instinct would argue in favor of a property right in gross for trademark holders – something the courts have vigorously resisted in the past.[19]

Admittedly, the courts have no easy task in deciding whether certain behavior on the Internet is likely to confuse consumers. Search technologies, as well as consumer practices and expectations, are constantly evolving in a way that makes it impossible to assess the existence and the costs of consumer confusion.[20] But courts in these cases must keep in mind that they are more than enforcers of existing norms – they are norms creators, in the sense that the rules they develop will determine practices on the Web and whether the Internet realizes its potential as a vast clearinghouse of information and content.[21] If the courts create norms in which search engines must avoid “using” trademarks in any way that might bring financial benefit to the trademark holder’s competitor, to some non-competing business, or to themselves, then they will have disserved the objectives of the Lanham Act by turning trademarks into vehicles for suppressing information.

Part I describes the relationship between trademarks and information. It begins with the economic defense of trademarks as tools for reducing consumer search costs, and describes a number of doctrines in trademark law designed to promote, rather than impede, the free flow of product information. While the view of trademark enforcement as reducing consumer search costs is well-known in the economic literature, we extend that work by showing how many of the limitations and defenses in trademark doctrine also serve that goal.

Part II explores the recent distortions of trademark doctrine along the axes described above. It concludes that these extensions of trademark law collide with both longstanding trademark doctrine and the normative goals of trademark law, and discusses the implications of this result for the availability of information in the marketplace. Beginning with the trend toward liability against parties who have not themselves “used” trademarks in the traditional branding sense, we contend that these cases will inevitably chill the behavior of publishers, search engines, and other intermediaries whose information-facilitating services implicate the use of all kinds of words, including trademarks. This chilling effect is only compounded by the indeterminacy of trademark analysis in an “initial interest confusion” era. The absence of normative guidelines in initial interest confusion cases, together with the courts’ increased reliance on an unjust enrichment rationale, at best make it tricky to predict whether a court would condone or condemn a particular use of a trademark on the Internet. At worst, these developments suggest that parties now risk liability by using trademarks in truthful, information-enhancing ways. The combined effect of these developments could well be to make consumer search more difficult on the Internet than in the offline context.

We offer two modest suggestions to reverse these trends and reinstate trademark law as an information facilitator. First, liability under the trademark and unfair competition laws should apply only against parties who use trademarks or misleading advertisements to sell their own products, or those acting closely in concert with them. To extend liability beyond this tight circle would defeat the Lanham Act’s objective of facilitating truthful information to ensure efficient, competitive markets. Second, courts should resist the temptation to turn “initial interest confusion” into an alternative theory of trademark liability, rather than an occasionally useful lens for assessing traditional infringement and unfair competition claims. The difference is not merely semantic; by reorienting trademark analysis toward the capturing of “initial interest” rather than on the likelihood that consumers will actually be confused, some courts have ventured away from the core factual inquiries that determine whether a particular use of a trademark will distort the information marketplace. While initial interest confusion can certainly distort the marketplace and increase consumer search costs, evaluation of both the existence and the costs of such confusion requires the rigorous, fact-specific analysis that trademark infringement and unfair competition law have historically involved.

Part III closes with some general observations about the role of normative goals in trademark rulemaking. None of the changes that we advocate in this article require new law, but simply a reinstatement of longstanding legal standards in trademark cases. Indeed, while we criticize some courts for ignoring these guiding principles, other courts are getting it right, applying the doctrine of trademark use in the online environment. Because the courts in the cases we criticize appear willing to stretch or ignore doctrine because of their equitable concerns about unjust enrichment, the solution lies, in part, in reminding them of the equities on the other side – the reasons we have limiting doctrines in trademark law and the importance of applying them uniformly.

I. Trademarks and Information

A. Economic Theory – Trademarks and Search Costs

Most people think of trademark law in terms of what it forbids: the use of another party’s trademark, or something resembling it, in a way that will cause confusion among consumers in the marketplace. Courts commonly describe the goal of trademark law as avoiding consumer confusion, which has the corollary effect of preventing the appropriation of a producer’s good will. Both consumers and producers, these courts point out, benefit from having access to truthful information about the source of products and services.

In economic terms, trademarks contribute to economic efficiency by reducing consumer search costs.[22] Rather than having to inquire into the provenance and qualities of every potential purchase, consumers can look to trademarks as shorthand indicators. With information less expensive, consumers will demand more of it, and will arguably become better informed, resulting in a more competitive market.[23] This system works, of course, only if consumers can trust the accuracy of trademarks, and this is where the law comes in.[24] By protecting established trademarks against confusing imitation, the law ensures a reliable vocabulary for communications between producers and consumers. Both sellers and buyers benefit from the ability to trust this vocabulary to mean what it says it means. Sellers benefit because they can invest in goodwill with the knowledge that others will not appropriate it.[25] Consumers benefit because they don’t have to do exhaustive research or even to spend extra time looking at labels before making a purchase; they can know, based on a brand name, that a product has the features that they are seeking.[26] Trademark law, in other words, aims to promote rigorous, truthful competition in the marketplace by preserving the clarity of the language of trade.[27]

Overly restrictive trademark law has the potential to stifle competition, rather than facilitating it. Particularly when trademark holders have economic power, giving them absolute control over uses of their marks could erect significant barriers to entry by competitors seeking to describe their own products.[28] Even in less differentiated markets, strong trademark rights come at a cost because they have the potential to remove words from our language and product features from competition.[29] One task of trademark law, then, is to preserve the informative role of trademarks while minimizing these downside risks.

B. Some Limiting Rules of Trademark Law, and Their Search Costs Rationale

The pro-information, pro-competition goal of trademark law has several important implications for the scope of trademark protection, particularly in comparison to other areas of intellectual property law. First and most generally, trademarks are not in gross property rights, but limited entitlements to protect against uses that diminish the informative value of marks.[30] Trademark law thus historically limited itself to preventing uses of marks that “defraud[ed] the public”[31] by confusing people into believing that an infringer’s goods were produced or sponsored by the trademark holder.[32] Likelihood of confusion does not necessarily follow every time a party adopts another’s trademark; it turns on a complex analysis that considers competitive proximity, consumer sophistication, and other factors that explore whether consumers will truly presume a false association and thus taint the information marketplace.[33] And while Congress recently added a federal cause of action based on the “dilution” of famous trademarks,[34] the statute is expressly designed to focus on cases in which consumer search costs are increased by “blurring” the significance of a unique mark[35] or giving it a negative association, and to permit uses such as commentary and comparative advertising that actually facilitate consumer search.[36] The Supreme Court has further limited the dilution law by interpreting it to require actual harm to a mark’s selling power – i.e., actual injury to the source-identifying function of a famous trademark.[37] Like the more traditional likelihood of confusion analysis, therefore, dilution – at least as properly understood[38] – thus turns on injury to the informative value of a mark.[39]

Second, trademark law rewards – and provides incentives for – investment in goodwill, but does not provide rights to all of the economic value that derives from that goodwill. Our competitive economy is based on the premise that competitors can generally appropriate ideas for products and services, as long as they are doing so in a non-deceptive way and not infringing some other exclusive right such as copyright or patent.[40] The patent and copyright systems represent a response to the potential market failure that can result from copying of public goods.[41] By contrast, trademark law is avowedly not designed to resolve any perceived failure in the market for quality products and services.[42] Thus, trademark law is reluctant to provide protection for product configurations, because doing so may give the trademark owner control not just over search characteristics but the intrinsic value of the product itself.[43] Only where the product configuration has an established meaning in the minds of consumers is it entitled to protection.[44]

Trademark law’s pro-competitive objectives sometimes require limitations on trademark holders’ rights, even when competitors might appear to receive a windfall as a result.[45] The Supreme Court recently emphasized, for example, that even if the public associates a particular feature with its first producer, the Lanham Act does not prevent others from copying the feature if it is part of what makes the product work.[46] The law may require some remedy, including clear source identification, to protect against consumer confusion and the resulting increase in search costs, but it does not allow exclusive rights to the functional product feature. In other words, when market access and competition run in tension with the trademark holder’s interests in protecting its product-associated goodwill, the competitive interests trump.[47]

Many venerable doctrines of trademark law reflect these general guiding principles. The genericness doctrine, for example, prevents a party from claiming rights to a term “that refers, or has come to be understood as referring, to the genus of which the particular product is a species.”[48] Competitors, in other words, have the right to explain what they are selling, even when their use of the generic term clearly piggy-backs on the efforts of the party that first introduced the product.[49] The genericness doctrine arises out of a concern for consumer search costs. Consumers will be misled if what they believe is a generic term is in fact a product sold by only one company.[50] And if competitors cannot use the generic term to describe their own products, consumers will incur unnecessary expense in trying to locate the competitors’ versions. At the same time, the genericness doctrine can impose search costs on consumers, particularly when a once-famous mark such as “aspirin” or “thermos” becomes generic. Consumers who associate the famous mark with the company that uses it may well be confused when competitors begin using the mark as a generic term.[51] The law is willing to make that sacrifice once a critical mass of consumers treat the term as generic, because the harm to them outweighs the harm to the diminishing number of consumers who view the term only as a mark.[52] Even so, courts sometimes take steps to protect trademark owners in this situation, for example by establishing rules requiring competitors who adopt a generic term that was once a protectable trademark to take steps to minimize confusion with the former mark owner.[53]

The functionality doctrine likewise prevents parties from claiming trademark rights in a product feature that “’is essential to the use or purpose of the article or if it affects the cost or quality of the article.’”[54] Even when consumers have come to associate a particular feature with a single seller, it cannot serve as a trademark if exclusive use of it would put competitors at a disadvantage.[55] The connection between the functionality doctrine and a functioning market is even more fundamental than search costs – consumers cannot choose between competing products if one manufacturer can use a law designed to reduce consumer search costs to eliminate competing products altogether. Preventing trademark owners from protecting functional aspects of their products is therefore consistent with a search costs rationale. Because functional characteristics, aesthetic appeal, and source-identifying information may sometimes be lumped together in the same product, however – think of the Ferrari[56] – the functionality doctrine has the potential to increase consumer search costs in some cases.[57] When a functional product feature has achieved secondary meaning, for example, some consumers might assume that all products with that feature come from a single source. As with generic marks, however, courts have responded to this possibility not by prohibiting use of the feature, but by requiring competitors to “use reasonable care to inform the public of the source of [their] product[s].”[58] To the extent that the use may even then mislead some members of the public, the functionality doctrine presupposes that the harm to consumers in these cases is outweighed by the greater availability of competitive products in the first place.

In all of the examples described above, the law seeks to minimize consumer search costs. While trademark law generally does that by providing protection to trademark owners, sometimes achieving that goal requires limiting the scope of trademark rights. In these instances, the interests of robust competition and reducing search costs outweigh the interest of the trademark claimant in appropriating the full value of its goodwill. The law limits trademark rights in these cases despite the apparent windfall to competitors. The genericness and functionality doctrines present “hard cases” precisely because there are search cost rationales on both sides of the argument.

The rule that competition and information dissemination trump trademark holder economic interests where the two conflict applies with even more force when the use that a competitor wishes to make of a trademarked term is one that unambiguously reduces search costs. Resellers of both new and used products, for example, have a right to use trademarks to describe the items that they sell. “The result is, of course, that the second-hand dealer gets some advantage from the trade mark. But … that rule is wholly permissible so long as the manufacturer is not identified with the inferior qualities of the product resulting from wear and tear or the reconditioning by the dealer.”[59] Competitors, moreover, may use descriptive marks in their non-trademark sense, to describe the features or qualities of their own products.[60]

Perhaps most significantly in light of the recent trends in trademark law, competitors have an affirmative right to use others’ trademarks to capture public attention in an attempt to divert it to their own products. As long as they do not mislead people into presuming some kind of affiliation between themselves and the trademark holder, competitors may use the mark to explain that their product imitates or aspires to the qualities of the trademark holder’s goods. In Saxlehner v. Wagner,[61] the Supreme Court allowed a natural water producer to use its competitor’s mark to identify the product that it was copying. Justice Holmes explained that, as long as the defendants did not create confusion about the real source of their product, they were free to “tell the public what they are doing, and to get whatever share they can in the popularity of the water by advertising that they are trying to make the same article, and think that they succeed.”[62] The Court distinguished between deceptive appropriation of goodwill and legitimate comparative advertising, concluding that by flagging its product as an imitator of the original, “they are not trying to get the good will of the name, but the good will of the goods.”[63]

Similarly, in Smith v. Chanel, Inc.,[64] the court allowed a knock-off perfume manufacturer to advertise that its perfume smelled like Chanel No. 5.[65] The court dismissed Chanel’s argument “that protection should also be extended to the trademark’s commercially more important function of embodying consumer good will created through extensive, skillful, and costly advertising,” reasoning that “the courts … have generally confined legal protection to the trademark’s source identification function for reasons grounded in the public policy favoring a free, competitive economy.”[66] Landes & Posner explain that the result in Smith is entirely consistent with the search costs rationale: “It would have been very costly for consumers to acquire such information [about the smell of the original perfume and the copy] before purchasing the copier’s perfume because the perfume was sold through the mail.”[67]

The same rationale has led courts to allow generic manufacturers to imitate branded trade dress in a way that evokes but doesn’t confuse.[68] These cases, like the others involving comparative advertising, emphasize that the public benefits from having fuller information about the products available in the marketplace.[69] The consumer search cost rationale thus both justifies trademark protection in the first place, and helps to set the limits on the scope of that protection.

Two final limiting doctrines help to ensure that the trademark grant does not stifle informative speech by non-competitors. First, to infringe, a defendant must “use” a mark “in commerce,” “on or in connection with any goods or services….”[70] Courts historically insisted that trademark “use” required that the defendant market goods or services under the mark.[71] As the Eighth Circuit recently explained, “the mark holder is generally not entitled to relief unless the defendant advertises or otherwise promotes [the actual mark] thereby causing the public to see the protected mark and associate the infringer’s goods or services with those of the mark holder.”[72] Limiting trademark rights to a right to prevent confusing uses of the mark as a brand helps ensure that trademark rights remain tied to their search cost rationale – only those individuals or companies who are using the mark to advertise their own products or services have the motive and opportunity to interfere with the clarity of the mark’s meaning in conveying product information to consumers, and so only those uses ought to be of concern to trademark law.[73]

Second, and relatedly, defendants who do not themselves “use” a mark in commerce can face liability for another’s infringement only if they actively induce that infringement, or knowingly help to bring it about.[74] By limiting trademark claims to those who themselves use marks in a way that suggests some affiliation between themselves and the trademark holder (and to others intimately involved in their infringement) the law ensures that information facilitators, publishers, and others who bear only a tangential relationship to trademark infringement can go about their business without the responsibility of having to police all of the parties with whom they have commercial relations.

C. Alternatives to the Search Costs Rationale

There are other economic justifications sometimes offered for trademark law.[75] Some have spoken of trademark law as promoting a “signaling” function of advertising. Others talk about the growth of national brands as facilitating the expansion of companies into different territories and product markets. Still others reason that trademarks are a form of property, and their owners are therefore are entitled to a broad right to prevent others from using the marks.[76] These alternative justifications should not distract the reader from a proper focus on search costs, however.

The signaling theory suggests that firms advertise because the fact that they can afford to advertise signals that their goods are of higher quality, since the advertising pays off. But this theory is problematic, because it works only if no one knows about it. If consumers actually treat advertising as a signal of quality, makers of low-quality goods have an incentive to deceive those consumers by advertising their goods, leading the consumer to mistakenly believe they are of high quality. Signaling theory is therefore self-limiting.

The role of trademarks in franchising, brand differentiation, and national expansion is indeed an important one.[77] But that role is consistent with – and indeed subsidiary to – the search costs rationale we discuss in text. Trademarks can serve as the basis for franchising or expansion into related markets only to the extent that consumers connect the mark with the maker of a set of core products. If the mark is not serving that purpose, building a national reputation on the basis of it will prove impossible. Some have suggested that companies who leverage their brand into multiple products are diluting this connection, interfering with consumer search costs.[78] But companies have an incentive to maintain the value of their brand, and in a competitive market they are unlikely to expand the reach of a single trademark so far that consumers are confused rather than enlightened by the use of the brand name.[79]

Finally, it is increasingly common to simply assume that trademarks are a form of (intellectual) property, and therefore that their “owner” is entitled to control virtually any use of them. There are several problems with this rationale. First, it is not true that even the owners of real property have unlimited power over their land.[80] Second, intellectual property is fundamentally different from real property, and the rights granted to owners of patents and copyrights are substantially weaker than those granted to real property owners.[81] Third, and most important, trademarks are not like other forms of intellectual property. We grant patents and copyrights in order to encourage the creation of patented and copyrighted works. There is no similar need to encourage the creation of brands. While it is occasionally suggested that trademarks are protected in order to encourage the creation of new brands,[82] that is not in fact a generally accepted rationale for trademark protection, with good reason. As the Fourth Circuit observed decades ago, “a man of ordinary intelligence could easily devise a score of valid trade-marks in a short period of time.”[83] Without such a rationale, calling trademarks property reduces to an unsupported conclusion that we should grant a certain set of rights over language without an accompanying justification.[84]

II. The Dismantling of Trademark Standards in Internet Cases

Perhaps ironically, the Internet – with its promise as an informational mecca[85] – has triggered a rethinking of the rules of trademark law in a way that could well restrict the free flow of comparative information in the marketplace. Courts in Internet cases have resurrected the oft-rejected unjust enrichment rationale for trademark rights, and have used it as the basis for a steady expansion of trademark holders’ rights, in two distinct directions. First, a handful of courts have widened the net of trademark infringement to encompass search engines, advertising firms, and others who help competitors to reach their audience through nontraditional “use” of established marks. Second, courts have extended the definition of actionable confusion to include the mere possibility of customer diversion to competitors’ websites, even in some cases in which the consumer knows where she is headed at all times. Both of these expansions undermine the core goal of trademark law – to reduce consumer search costs and increase the availability of information in the marketplace. Their combination in cases like Playboy v. Netscape[86] and 1-800 Contacts, Inc. v. WhenU.com[87] is particularly problematic.

Playboy v. Netscape involved a suit against Netscape based on its use of trademarks as “keywords” to help target advertising to relevant audiences. Netscape’s search engine, like most such services, offered advertisements alongside the search results generated in response to user queries. Netscape had allowed – indeed, required – adult services that advertised through its search engine to include “Playboy” as one of the search terms that, if entered by users, would generate an ad on Netscape’s search results page. The Ninth Circuit held that this use of the “Playboy” mark could confuse consumers into believing that the advertisers were affiliated with Playboy,[88] and that Netscape could face legal responsibility for creating that confusion. In allowing Playboy’s claim against Netscape to proceed, the Ninth Circuit decided not to choose between a direct and contributory infringement theory, concluding “that defendants are either directly or contributorily liable,” so that it “need not decide” which.[89] While the court’s reasoning remains a mystery, the decision left no doubt that, in the panel’s view, the Lanham Act could reach parties that used trademarks, not as brands, but as classification tools for targeting advertisements whose content they did not control.

In the 1-800 Contacts case,[90] a trademark holder sued to prevent the use of its mark to generate pop-up advertisements. The plaintiff, which sold contact lenses under the mark “1-800 Contacts,” sought to prevent its competitors’ pop-up ads from being displayed when users entered <www.1800contacts.com> into their web browser or used the “1-800 Contacts” mark as a query in an Internet search. Plaintiff sued both the advertisers and the firm whose software generated the pop-up ads, claiming that both had “used” its marks in a way that created a likelihood of confusion among consumers. In allowing the claims to proceed, the court held that both categories of defendants had “used” the trademark as contemplated by the Lanham Act. First, the court held that by generating pop-up ads to appear “when users have specifically attempted to access Plaintiff's website – on which Plaintiff's trademark appears – Defendants are displaying Plaintiff's mark ‘in the ... advertising of’” the defendant advertiser’s competing services.[91] Second, by including the trademark in a directory of terms that trigger advertisements on computers that use defendant’s software, the software developer “used” the mark “to advertise and publicize companies that are in direct competition with Plaintiff.”[92] In doing so, the court put itself in direct conflict with the only two other cases to consider trademark liability for pop-up ads.[93]

A. Should Intermediaries Face Liability at All for Trademark Infringement?

1. Trademark Use

The Playboy and 1-800 Contacts cases demonstrate a disturbing trend away from the statutory requirement of trademark use, at least as traditionally interpreted by the courts. The Lanham Act prohibits only use of a trademark “in connection with” the offering of goods or services in commerce, and defines such a use as occurring in the context of services “when [the mark] is used or displayed in the sale or advertising of services and the services are rendered in commerce."[94] It is the use of the mark to brand or advertise the defendant’s services – so-called “trademark use” – that triggers trademark law.[95]

The trademark use requirement serves a gatekeeper function, limiting the reach of trademark law without regard to a factual inquiry into consumer confusion. The rationale for the doctrine stems from the practical reality that it would be both unwise and impossible to permit trademark owners to control every use of their mark. People and businesses use trademarks every day, in conversation, in news reporting, in songs and in books. Trademark law has never given trademark owners exclusive control over any use of their marks. Rather, the law is designed to prevent consumer confusion by those who brand their own goods or services with a mark sufficiently similar to the plaintiff’s mark that consumers may be deceived into believing there is some connection between the two. Individuals and companies may make reference to or use of a trademark without fear of liability unless they are making a trademark use.

The courts that have found Internet service companies liable for initial interest confusion err by holding that using a trademarked term as a “trigger” for supplying an advertisement is a trademark use of that term. Both 1-800 Contacts and Playboy make this error.[96] The courts appear to have been misled by the similarity in terminology of two very different elements of a trademark cause of action: the “use in commerce” requirement and the “trademark use” requirement. It is true that in both these cases the trademarks are used in interstate commerce in the minimal way necessary to invoke the jurisdiction of the federal courts.[97] Unfortunately, an increasing number of courts apparently believe that satisfying this minimal requirement is the same as proving that the defendants had made trademark use of the plaintiff’s brand.[98] As a result, they do not analyze the requirement of trademark use directly. This is a crucial error.[99] Courts outside the United States also exhibit confusion about the proper scope of trademark use, though they have no jurisdictional “use in commerce” requirement to apply and so are led astray in other ways.[100]

Selling advertising space based on an Internet keyw