Kevin Zeck, Referential Fair Use & Keyword Advertising: The Necessity of Product Placement to our Domestic System of Free-Market Enterprise, 44 Gonz. L. Rev. 519 (2009).[PDF] [Westlaw] [LexisNexis] %CODE1%
As this article was about to be published, a panel of the Second Circuit handed down its decision in Rescuecom Corp. v. Google Inc. The Rescuecom decision impacts parts of this article, and, at times, the impact is substantial. Due to time constraints, this preface was necessary to integrate the decision in Rescuecom with this article. The preface explains the history of Rescuecom, the substance of the panel’s decision, the panel’s interpretation of the definition of “use in commerce” found in 15 U.S.C. § 1127 (2008), and the impact of the decision on this article.
A. The History of Rescuecom v. Google
The history of the case is simple: Rescuecom sued Google for trademark infringement because Google had been selling Rescuecom’s eponymous trademark as a keyword advertising trigger. Google moved to dismiss the suit for failure to state a claim on the theory that Google had not made an actionable use of a trademark within the meaning of the Lanham Act. The district court granted Google’s motion, and Rescuecom appealed. %CODE2%
B. The Second Circuit Panel’s Decision in Rescuecom v. Google
A panel of the Second Circuit then vacated and remanded the decision of the district court, holding that Rescuecom had sufficiently alleged a claim of trademark infringement. The panel reasoned that Rescuecom had sufficiently alleged an actionable claim because Google’s positioning of “sponsored” advertisements gave rise to a possible impression of affiliation between Rescuecom and any advertisers who purchased Rescuecom’s trademark from Google. Because of this possibility, the panel found that Google had made a “use in commerce”—as defined in 15 U.S.C. § 1127—of Rescuecom’s trademark. The panel openly questioned, however, whether Rescuecom would ever be able to prove a claim of trademark infringement.
Additionally, the panel took issue with Google’s invocation of “product placement” as a talisman against trademark infringement lawsuits. Curiously, the panel circularly defines permissible product placement without citing to a single authority and concludes that Google was engaging in mischievous product placement. The panel then held that mischievous product placement can result in trademark infringement liability.
The panel, however, failed to set forth any clear delineation between permissible and impermissible product placement behaviors, giving rise to the possibility that the Second Circuit may revisit this holding and determine that some types of product placement, including some types of keyword advertising, are not actionable under the Lanham Act as a matter of law.
C. The Panel’s Interpretation of “Use in Commerce”
Turning to one of the more perplexing aspects of the decision, the panel held that the definition of “use in commerce,” as found 15 U.S.C. § 1127, governed whether Google had made a “use in commerce” of Rescuecom’s trademark for purposes of infringement. The panel added, as dictum, an appendix to its decision to explicate its declaration that section 1127’s definition of “use in commerce” applied to the infringement sections of the Lanham Act.
The appendix proffered a strained, but technically accurate, interpretation of
section 1127’s definition of “use in commerce.” The panel correctly noted that the first sentence of the definition applied to registration requirements, but then proceeded to state that the second sentence, standing alone, could apply to the infringement sections of the Lanham Act. Rather than construing the first and second sentences together and then finding this totality to be inapplicable to the phrase “use in commerce” for purposes of infringement, the panel bifurcated the statutory definition and stated that it would apply the second sentence, standing alone, to determine whether a trademark had been “use[d] in commerce.”
The panel found that this interpretation was more faithful to principles of statutory interpretation for the following reasons: (1) prior to the 1988 Amendments to the Lanham Act, only the second sentence appeared in the definition of the “use in commerce”; (2) prior to the 1988 Amendments, it was possible that the second sentence, standing alone, had been used to define the scope of infringing behaviors; and (3) although the 1988 Amendments added the first sentence to the definition, these amendments failed to make any changes to the second sentence of the definition, and thus the pre-1988 meaning of the second sentence still survives.
While thoughtful and historically accurate, the court’s interpretation is needlessly complex. Although the 1988 Amendments did not amend the second sentence of the definition, these amendments added the first sentence, altering the meaning of the definition because the context of the amended definition appears to require that both sentences be read together. If the entire definition was altered by the addition of the first sentence, which this article suggests is the correct interpretation, section 1127’s definition of “use in commerce” does not apply to the phrase “use in commerce” as found in the infringement sections of the Lanham Act. Indeed, the Senate Judiciary Committee’s Report concerning the 1988 Amendments supports this article’s interpretation of section 1127:[T]he . . . definition [of “use in commerce”] is intended to apply to all aspects of the trademark registration process, from applications to register, whether they are based on use or on intent-to-use, and statements of use filed under Section 13 of the Act, to affidavits of use filed under Section 8, renewals and issues of abandonment. Clearly, however, use of any type will continue to be considered in an infringement action.
The panel found this article’s interpretation viable, even though the panel did not recommend it. This article builds upon the idea that the addition of the first sentence altered the entire meaning of section 1127’s definition of “use in commerce” and suggests that: (1) section 1127’s definition of “use in commerce” should not apply to the Lanham Act’s infringement sections; but (2) that section 1127’s definition of “commerce” should, with the result that any use of a trademark within Congress’ Commerce Power is potentially an infringing use of a trademark. This interpretation is faithful to the intent of Congress, makes sense upon a contextual reading of section 1127, and avoids needless complexity.
D. The Impact of the Panel’s Decision on this Article
The panel’s decision in Rescuecom v. Google undermines the reasoning drawn from 1-800 Contacts, Inc. v. WhenU.com, Inc. that had been used by numerous district courts within the Second Circuit to dismiss trademark infringement claims concerning keyword advertising. As a result, the Second Circuit no longer appears to be in conflict with other Circuits on the issue of whether keyword advertising can be infringing as a matter of law.
For reasons of time and efficiency, however, Part V of this article presents the state of the law as it existed before the panel’s decision.
Additionally, although the panel attempted to diminish the viability of “product placement” as an absolute shield against trademark infringement lawsuits, this article still vigorously argues that “product placement” is a viable ground for limiting the concept of trademark infringement.
Finally, the panel’s attempt to undermine the reasoning of 1-800 Contacts attacks the essential thesis of this article. This article’s primary thesis is that a use of a trademark is not actionable where the mark is used internally in a way that does not communicate an impression of the mark to a consumer. Because this article relies on the reasoning of 1-800 Contacts in supporting its thesis, the panel’s decision weakens the likelihood that this article’s central recommendation will be accepted by the Federal Judiciary. Despite this, the author believes that the central thesis of this article remains sound and should be seriously considered as a viable solution to the problems presented by keyword advertising litigation.
This article is all about getting back to the basics of trademark infringement in the context of keyword advertising. Keyword advertising, a form of internet marketing, occurs where a merchant pays an internet search engine, such as Google, to display an advertisement in response to a web user’s search query. A review of federal case law concerning keyword advertising reveals that courts have largely abandoned the basics of trademark law and built precedent upon precedent without considering the underlying nature of internet advertising. As a result, some federal courts have created a body of trademark law that is overreaching in scope. In creating this body of unbalanced law, these courts have used the spacious language of the Lanham Act’s infringement provisions to stop what they perceived as unfair competition practices.
In pursuit of clarity regarding keyword advertising, this article will first examine the mechanics of keyword advertising and internet search engines. It will then consider the basic legal framework of trademark protection, as well as the policies underlying trademark law. From there, this article will set forth the current legal landscape concerning keyword advertising by considering the current split in federal case law. This article will then show that keyword advertising is a “use in commerce” under the Lanham Act by reviewing the textual language of the Lanham Act and the Senate Judiciary Committee’s Report on the 1988 Amendments to the Lanham Act. After acknowledging that keyword advertising is a “use in commerce,” this article then argues that almost anything concerning the use of a mark in commerce could be actionable under the Lanham Act, but that the interpretation of the Lanham Act should be confined by tying its protections to the fundamental purposes of trademark protection. Finally, this article considers the similarity between legal “product placement” practices and keyword advertising, and contends that uses of a mark akin to product placement should not be actionable under the Lanham Act.
And with that, back to the basics. . .
III. The Mechanics of Keyword Advertising and Search Engines
Keyword advertising is a form of internet advertising where a merchant pays a search engine to display an advertisement when a web user enters a particular keyword as part of a search engine query. The merchant’s advertisement usually consists of a small amount of text and a link to the merchant’s website, and is typically displayed next to or above the natural search results for the search engine query. Natural search results (also called organic search results) are results that are retrieved by a search engine on the basis of some sort of relevance algorithm, i.e., they are search results which are relevant to the web user’s inquiry.
No understanding of keyword advertising would be complete without an understanding of the underlying mechanics of search engines. Search engines conduct searches of the internet not by actually engaging in real time searches of the internet, but by searching vast indexes of web-content. The indexes are produced in a two-step process. First, web-crawling programs copy each webpage they can access and transfer that copy to a webpage repository. Second, an indexing program takes the files in the repository and breaks them down into their most important components; essentially, the indexing program identifies only the “vital descriptors” contained in the page, saves this unique content as a file, and places the file in an index.
There may be several different types of indexes for any given search engine, and the types of indexes employed depend upon the nature of the algorithm used to create search results. Generally speaking, however, there is almost always a “content index” that stores all of the unique content contained in the body of a webpage and in the meta-content of the webpage. Meta-content is content that is not viewed by the web user, but that is still part of a webpage; meta-content may include a list of metatags, a title for the webpage, or other information.
With the rise of Google, a second type of index has become popular: the “structure index.” A structure index contains information relating to hypertext links going into and out of webpages. The importance of the structure index is explained below.
Using indexes, search engines can perform searches of the internet in a short amount of time because the search engines only need to search for relevant information in the stored indexes, according to some sort of algorithm, rather than attempting to search the vastness of the internet in real time. Search engines almost always employ some sort of content-based search in producing search results. A content-based search merely determines whether any matches exist between the text of a web user’s search query and information contained in the content index. If there are matches, and if the search engine uses a purely content-based search algorithm, a content-based search also determines the putative content-based relevance of any matches found. Often, the degree of relevance in a content-based search is determined by the repetition with which a webpage’s index file contains a word entered as part of a search query.
Because relevance in solely content-based searches can be determined merely by the repetition of words in a webpage, content-based searches are susceptible to manipulation. For example, in the early days of the internet, a person could easily manipulate simplistic search engines by repeating a given search term thousands of times in the metatags of a website or by hiding a search term in the background of a website thousands of times (such as by employing a red background on a website and then typing Coca-Cola a thousand times in red lettering). This manipulation allowed people to make their websites the top search result for a given search term, even if their webpage was entirely unrelated to the search term entered.
Two young men at Stanford came up with a solution to the manipulation problem. Larry Page and Sergey Brin, the founders of Google, Inc., created a webpage relevance algorithm that relied on both content and “popularity.” Popularity in the search engine context is simply defined as the number and importance of hypertext links going into and out of a webpage. Popularity metrics are stored in the structure index.
Page and Brin essentially believed that the most relevant search results could be discerned by allowing the web, as a collective entity, to determine the relevance of a given webpage (a.k.a. its popularity) that also met the basic content criteria of a web user’s search query. To everyone’s great satisfaction, they were correct. Today, the three major domestic search engines—Google, MSN, and Yahoo—employ some form of content-plus-popularity search algorithm. These three search engines now conduct roughly ninety-four percent of all search engine queries.
To defray operating costs and to provide for technological improvements, search engines have turned to keyword advertising or variants of keyword advertising (rather than shifting the costs to its primary consumers, web users searching the internet). Concerning the business practices underlying keyword advertising, a search engine generally allows an advertiser to purchase the “use” of a specific keyword. The advertiser then submits an ad to the search engine, and the search engine displays the ad when the particular keyword is entered as a search query. The advertiser usually does not have to pay the search engine when its ad is merely displayed. Instead, the advertiser is only required to pay the search engine when a web user clicks on its ad and views the underlying link (this action is called a click-through). Although the amount of each click-through may be just a few pennies, the money at stake with keyword advertising amounts to billions of dollars.
The collision between trademarks and keyword advertising occurred because search engines have allowed advertisers to purchase the use of unrelated entities’ word marks as keyword advertising triggers. This purchase allows the competitor(s) of a trademark owner to display ads for their products alongside the results for a search query based on the owner’s word mark. Trademark owners decry this practice, alleging that it is trademark infringement. Consequently, the remainder of this article attempts to deconstruct the framework of trademark protection and argues that, generally, keyword advertising is not trademark infringement. To the contrary, it is a practice that is highly beneficial to consumers.
IV. The Basic Structure and Policies of Trademark Law
A. Trademark Law as an Extension of Passing Off
Trademark law is a form of unfair competition law that was derived from the concept of “passing off.” Passing off is perhaps best understood by a simple illustration. Two businesspersons are in competition with each other, each making the same product. One of them makes an exceptionally high quality product and is rewarded by a high volume of sales while the other person’s product is of inferior quality. The person producing the high quality product affixes a mark to her product to distinguish it from the products of others, thereby hoping to reap the reputational value—or goodwill—flowing from the quality of her product. The second person sees the value that the first person has created and seeks to exploit it by placing the first person’s mark on the second person’s products and selling them to the public. This is passing off.
Trademark law is essentially an extension of passing off. The hallmark cause of action under trademark law is trademark infringement. Generally speaking, a person is liable for federal trademark infringement when that person uses, without authorization, a device in commerce “in connection with the sale, offering for sale, distribution, or advertising of any goods and services” that is likely to cause confusion with a valid and protectable trademark of another person. Trademark infringement actions aim to stop confusion as to source, sponsorship, or affiliation—that is, confusion in the minds of consumers as to who is responsible for or affiliated with the production of a particular good or service.
While trademark law’s focus is most prominently on the prevention of certain types of consumer confusion, courts have found that trademark law also serves other related, albeit limited, policies such as: (1) protecting consumers from fraud; (2) increasing the quality of products by identifying the party responsible for a consumer’s experience; (3) reducing consumer search costs; and (4) protecting the goodwill of a mark holder.
B. Trademark Law Exists within the Boundaries of Free Competition
Courts have also recognized that trademark law exists within the boundaries of free competition and have therefore restricted trademark protection where it would unduly hinder such competition. Free competition provides a sufficient justification to limit trademark protection because free competition is necessary to our free-market system.
Free competition is necessary for free-market economies, such as the United States’, because such economies have no central planning agency to ensure an efficient distribution of economic resources, i.e., a distribution of physical resources and human capital that results in a politically palatable standard of living. Free competition fosters the efficient distribution of economic resources because it naturally results in the destruction of monopolies through increased competition, thus allowing consumers to individually choose the most efficient good or service for their needs. This individual choice results in an efficient distribution of economic resources because it allows society as a whole to determine where to allocate resources (and this collective determination aggregates and processes data on allocative need with a great degree of precision). Moreover, this efficient allocation of resources prevents the breakdown of economic activity by ensuring, at least to some degree, that producers are selling products at prices, and with features, that are attractive to consumers. A complete breakdown of economic activity results in socio-economic decline, and perhaps even revolution. Free competition, which prevents such a breakdown through the efficient allocation of resources, is therefore necessary for the continued functioning of free-market economies.
To the extent that trademark law threatens to severely disrupt the efficient allocation of resources by hindering free competition, courts have and should limit the scope of trademark law.
C. Inherent Historical Limitations of the Lanham Act
Before moving on to a review of case law concerning keyword advertising, it is also important to note one inherent historical limitation regarding the concept of trademark protection as embodied in the Lanham Act, as well as to note some of the ways that courts have stretched the Lanham Act (and the original paradigm in which it was enacted) to prevent what the courts have considered to be inequitable consequences.
The Lanham Act was written in the early 1930s and was passed into law in 1946. Although the Act has been amended several times, one of its two basic trademark infringement provisions has remained largely unchanged. Thus, it is easy to surmise one inherent limitation in the Lanham Act’s concept of trademark law that is still present today: the Lanham Act was designed for a physical marketplace. That is, trademark law under the Lanham Act was designed for situations where the consumer actually viewed the allegedly infringing mark. The Lanham Act did not anticipate the use of trademarks in metatags, keyword advertising triggers, and other forms of non-visible uses on the internet.
D. Contemporary Courts Endeavor to Prevent Inequitable Conduct Not Contemplated by the Lanham Act
Courts have endeavored, however, to prevent inequitable, non-traditional uses of allegedly infringing marks. Some of these attempts have been both valiant and necessary. For example, in Coca-Cola Co. v. Overland, Inc., the Ninth Circuit affirmed a district court’s finding of trademark infringement because the defendant-restaurant had been substituting orders for Coca-Cola with Pepsi. Although the consumers could not see any sort of mark affixed to the product given to them, the Ninth Circuit rightly held that this behavior was infringing because the defendant-restaurant essentially conveyed to its consumers that they were drinking Coca-Cola by failing to inform them that they were actually drinking Pepsi.
1. The Lanham Act in the Context of Metatags
Other recent attempts to extend trademark law have seemed questionable, however. Discounting for a moment the fact that purely content-based search engines are a thing of the past, courts have recently engaged in a rash of decisions that find the use of another’s trademark in metatags to be trademark infringement. These courts were attempting to prevent the manipulation of search engine results where the manipulation took the following form: a competitor of a trademark owner would incorporate the owner’s trademark into the metatag of the competitor’s website without permission, probably repetitively, in order to fool the search engine into putting the competitor’s website above or below the trademark owner’s website when the search engine returned search results. These courts apparently believe that such practices result in a misappropriation of goodwill through the creation of impermissible confusion.
Upon closer inspection, however, such rationales seem suspect. Although some form of confusion could arise from the use of metatags, the concept of trademark infringement as embodied in the Lanham Act would seem to suggest that such behavior is not infringing so long as the competitor’s search result contains only content that is distinct from the owner’s mark. That is, although the web user may be confused by the presence of a search result that does not match the query entered, the web user almost certainly suspects that the competitor’s search result is not affiliated with the mark’s owner. Instead, the user will probably demand a different search platform that produces more relevant search results. And, in fact, such a demand is partly what gave rise to Google’s search platform.
2. The Rise of the Initial Interest Doctrine
Another suspect, albeit necessary, attempt to prevent inequitable results in trademark cases has resulted in the rise of the initial interest doctrine. The initial interest doctrine is sometimes necessary to ward off highly sophisticated attempts to divert goodwill where the diversionary mechanism is the creation of temporary source or sponsorship confusion, which occurs and is dispelled before the ultimate point of sale. Initial interest confusion, however, has been used to justify some questionable decisions in the context of the internet.
The initial interest doctrine was essentially created in the context of brick and mortar sales (non-internet sales) where a consumer could view a product’s labeling and be initially confused as to its source or sponsorship. As a result of this confusion, the consumer would attempt to purchase the product or at least consider purchasing the product. Before the consumer could complete the transaction, however, she would realize that the source of the product was not the same as the desired source. Yet because the consumer had already expended resources (time, energy, attention, etc.) in attempting to purchase the product, the consumer might settle for the substitute product.
This rationale does not translate well to all areas of the internet, in part, because the underlying reason for the initial interest doctrine is greatly diminished in the context of the internet. In the physical world, a consumer’s expenditure of resources in attempting to obtain a product might actually cause the consumer to settle for a substitute product rather than expending more resources to obtain the desired product. On the internet, however, transactional costs are typically much less and indeed may only be as great as the click of a mouse.
Courts have overlooked this glaring problem and have issued decisions that extend the initial interest doctrine to the internet. For example, in Brookfield Communications, Inc. v. West Coast Entertainment Corp., a Ninth Circuit panel held that the use of a confusingly similar metatag is infringing under the Lanham Act because it is likely to cause initial interest confusion, even assuming that the content displayed in search engine results is not confusingly similar to the mark of another. The panel reasoned that such behavior is infringing because of the willingness of the consumer to settle for a substitute good or service upon finding an alternative source.
The panel analogized the situation in Brookfield to the following: a consumer is driving down the highway and sees a billboard sign for a well-known video store. The billboard has been put up by the video store’s competitor and directs the consumer to an exit at which the consumer will find the competitor’s store and not the well-known video store. Once off the highway, the consumer would be unable to locate the desired store and might instead settle for the competitor’s store, knowing that it is such.
The analogy employed in Brookfield is problematic because it does not adequately take into account the significantly lower physical resources expended in the internet search context as compared to the real world. In the physical world, the consumer might be distraught, having expended physical resources to reach the exit. The consumer might not want to expend any more resources, especially considering the fact that the consumer has no guidance as to the true location of the well-known video store. In the internet context, however, if a consumer arrives at the website of the competitor, the consumer can merely click the “back” button on her browser and enter a different and/or more developed search. The consumer will likely spend only a few seconds on this process.
Moreover, the Brookfield panel assumed that search engines conducted purely content-based searches, and thus assumed that the use of metatags alone could result in search engine manipulation. Yet Google’s revolutionary search technology has obviated this assumption. This technological development further undercuts the reasoning of Brookfield.
The Brookfield decision also demonstrates another reason why the initial interest doctrine does not translate well to the internet context with regard to meta-information such as hidden metatags, keywords, etc. The panel affirmed the grant of a preliminary injunction that stopped the defendant from using the plaintiff’s marks as metatags in any website, regardless of whether the website displayed content that was confusingly similar to the marks of the plaintiff. The panel based this holding on the above analogy. Interestingly, though, the panel’s analogy is seriously flawed because the analogy assumes that the consumer can see the defendant’s use of the plaintiff’s mark. Yet the science behind search engine technology shows that the ordinary consumer does not see the metatag use of the plaintiff’s mark.
In essence, then, the Brookfield analogy should have gone something like this: a consumer is driving down the highway. She thinks to herself, “I’d like to rent a movie from a specific well-known movie store.” In a worst case scenario, a billboard pops up that displays only directions to a competitor’s store. The consumer is annoyed that she is presented with no directions to the well-known video store and she may settle for the competitor’s store, but she will never be confused as to the source or sponsorship of the billboard advertisement.
In a more realistic scenario, the consumer will be presented with two billboards side-by-side: one for the well-known video store and one for the competitor’s store. Both billboards will provide directions to each respective store and the time it takes to get to each store (from the billboards to the store and back to the billboards) will be de minimis (or de micromis). As long as the competitor’s billboard contains only content that is not confusingly similar to the mark of the well-known video store, the consumer is never confused and should settle for the competitor’s store only because of some other consideration such as price advertised, availability advertised, etc.
When viewed as such, it is hard to find behavior that is legally objectionable in this situation. Yet the doctrine of initial interest confusion has become a foothold for courts in their quest to prevent uses of trademarks that merely feel unfair, even though the consumer (1) can easily choose between a mark-holder’s product and a competitor’s product and (2) is almost certainly never confused as to the source or sponsorship of both products.
Having set out the current legal landscape for trademarks in general, a review of federal keyword advertising precedent can now be undertaken appropriately.
V. The Key To Keyword Advertising Cases—Use In Commerce
A split exists in the federal courts over the question of how to approach the issue of keyword advertising. Courts within the Second Circuit have adopted the position that the mere use of another’s mark as a keyword advertising trigger is not a “use in commerce” under the Lanham Act, and thus is not actionable. The Ninth Circuit and courts within the Ninth Circuit have assumed that the use of another’s mark as a keyword advertising trigger is a “use in commerce” and have focused on whether such use creates a likelihood of confusion. Of the other federal courts that have addressed the issue of keyword advertising, either directly or by necessary implication, all have essentially fallen in line with the Ninth Circuit’s position.
Many federal courts, however, have yet to address whether keyword advertising can constitute trademark infringement. In fact, besides the Ninth Circuit, no appellate court has addressed whether keyword advertising is actionable within the meaning of the Lanham Act.
This section will review the two basic positions taken by courts on each side of the split, but will not advocate for either approach because both approaches are incomplete.
A. The Second Circuit Approach
The Second Circuit has taken the position that the use of another’s mark as something akin to a metatag is not a “use in commerce” within the meaning of the Lanham Act. Courts within the Second Circuit, relying on this precedent, have likewise found that a defendant does not make an actionable use of another’s mark when the defendant merely uses the other’s mark as a keyword advertising trigger, i.e., when the defendant purchases another’s mark as a keyword advertising trigger but does not incorporate any content into the keyword advertisement that is confusingly similar to the other’s mark (“mere keyword advertising”).
These courts reason that uses akin to metatags and keyword advertising triggers are not “uses in commerce” because these uses are purely internal uses of another’s mark. This rationale is likely limited, however, to instances where another’s mark is used “in a way that does not communicate [the mark] to the public.” The essence of this cryptic phrase is that another’s mark is used internally “in a way that does not communicate it to the public” when that mark is used solely as a referential device, i.e., when a mark is merely used to refer a consumer to possible alternatives that are clearly labeled as such.
The Second Circuit and courts within the Second Circuit, however, base their rationale, in part, on the definition of “use in commerce” found in 15 U.S.C. § 1127 (2008). As will be explained, such reliance is misplaced. Nonetheless, the basic point is this: under Second Circuit precedent, the use of another’s mark as something akin to a metatag or keyword advertising trigger is not actionable so long as any content displayed in the triggered advertisement or webpage is not confusingly similar to the other’s mark. Thus, under Second Circuit precedent, trademark infringement claims based on keyword advertising and metatag use may be disposed of with a motion to dismiss for failure to state a claim.
B. The Ninth Circuit Approach
The Ninth Circuit takes the opposite view. While a panel of the Ninth Circuit has never directly addressed the “use in commerce” issue, Ninth Circuit precedent, which sanctions the imposition of trademark infringement liability (or a likelihood thereof) concerning keywords and metatags, has led its district courts to conclude that such uses are “uses in commerce” under the Lanham Act, and thus actionable. A review of Ninth Circuit cases leads to the conclusion that the Ninth Circuit relies largely on the initial interest doctrine in finding trademark infringement liability in the context of keywords and metatags. As has been explained, an undisciplined application of the initial interest doctrine can lead to a theory of trademark infringement that is overreaching in scope.
Wisely though, at least one district court and one Circuit Judge in the Ninth Circuit have recognized the problems inherent in the Ninth Circuit’s positions concerning the initial interest doctrine and hidden uses of trademarks on the internet. The most poignant of these critiques comes from Ninth Circuit Judge Marsha Berzon’s concurrence in Playboy Enters., Inc. v. Netscape Commc’ns Corp. In that concurrence, Judge Berzon recognizes the distinction between initial interest confusion, which depends upon a consumer being given an initial impression that goods or services come from or are affiliated with one source when they in fact come from another source, and initial interest distraction (or so I will call it), wherein a merchant (or search engine) provides a consumer with a range of product choices that includes both a requested product and alternative products clearly labeled as such. Judge Berzon doubted that initial interest distraction could be actionable under the Lanham Act, and for good reason. As further developed in Part VIII of the article, initial interest distraction is an essential part of our free-market system because it is the foundation for “comparison shopping” and “product placement,” behaviors that are clearly—by themselves—not illegal.
Nonetheless, the courts within the Second Circuit stand largely diametrically opposed to the main position of the Ninth Circuit on the issue of keyword advertising. The former have created a mechanism that disposes of keyword advertising cases at the earliest possible stage of a lawsuit while the latter seeks to impose liability for such claims on the basis of perceived unfairness.
VI. Use of Another’s Mark As a Keyword Advertisement Trigger Is a Use In Commerce Under the Lanham Act
Unfortunately for the courts within the Second Circuit, their opinions have not yet won the day even though they contain the better approach. This lack of acceptance is likely because of the Second Circuit’s reliance on the Lanham Act’s definition of “use in commerce.”
The Lanham Act sets out several definitions in 15 U.S.C. § 1127 that are relevant to the keyword advertising issue. First, “commerce” is defined as “all commerce which may lawfully be regulated by Congress.” The section 1127 definition of “commerce” is broad, and it is highly probable that the use of a mark as a keyword advertising trigger would be subject to regulation under that definition.
Next, “use in commerce” is generally defined as “the bona fide use of a mark in the ordinary course of trade, and not made merely to reserve a right in a mark.” section 1127 then provides two examples of uses that are “deemed” to be uses in commerce. The first example provides that a use in commerce occurs on goods when a mark:
is placed in any manner on the goods or their containers or the displays associated therewith or on the tags or labels affixed thereto, or if the nature of the goods makes such placement impracticable, then on documents associated with the goods or their sale, and . . . the goods are sold or transported in commerce.
The second example provides that a use in commerce occurs in connection with services when a mark:
is used or displayed in the sale or advertising of services and the services are rendered in commerce, or the services are rendered in more than one State or in the United States and a foreign country and the person rendering the services is engaged in commerce in connection with the services.
Finally, section 1127 defines a mark (either a trademark or a service mark) generally as any device that is used to distinguish the goods or services of the owner from the goods and services of others.
The scheme set out by section 1127 specifically concerning “use in commerce” seems to set forth requirements relevant to the registration of trademarks on the principal register of the United States Patent and Trademark Office. This, in fact, is the clear legislative intent behind this definitional scheme. As was stated in the Senate Judiciary Committee’s Report concerning the 1988 Amendments to the Lanham Act:[T]he . . . definition [of “use in commerce”] is intended to apply to all aspects of the trademark registration process, from applications to register, whether they are based on use or on intent-to-use, and statements of use filed under Section 13 of the Act, to affidavits of use filed under Section 8, renewals and issues of abandonment. Clearly, however, use of any type will continue to be considered in an infringement action.
Thus, it appears that section 1127’s definition of “use in commerce” does not control whether a particular use of an allegedly infringing mark or device is a “use in commerce” under the infringement sections of the Lanham Act. Because section 1127’s definition of “use in commerce” relates only to the acquisition of trademark rights and because section 1127’s definition of “commerce” is not similarly limited, it seems appropriate to default to section 1127’s definition of “commerce” in determining the scope of an infringing “use in commerce.” Under this plain language analysis, almost any use of an allegedly infringing mark or device is actionable under the Lanham Act, including the unobservable use of another’s mark as a keyword advertising trigger.
However, such an unbounded interpretation does not seem consistent with the basic purposes and principles of trademark law as embodied in the Lanham Act. As argued below, “use in commerce,” under the infringement sections, should extend only to uses of an allegedly infringing mark or device that give a consumer the impression that the person using the allegedly infringing mark or device is somehow associated with the mark owner; but only where the impression arises in a manner that is inconsistent with the basic system of free market enterprise in which the Lanham Act arose and developed.
Courts in the Second Circuit have often placed some reliance on the text of section 1127’s definition of “use in commerce” in finding that keyword advertising is not actionable under the Lanham Act. These courts have reasoned that because the allegedly infringing mark is used as a keyword advertising trigger and is not affixed to goods or displayed in the course of rendering services, such use of a mark is not a “use in commerce” under the Lanham Act.
This reliance on section 1127’s definition of “use in commerce” is flawed for two reasons. First, the plain language of section 1127 makes it clear that a use in commerce is greater than the two illustrations presented in connection with the definition. That is, although section 1127 presents two situations in which a mark shall be deemed to have occurred—one for goods and one for services—the plain language of its general “use in commerce” definition is clearly broader than its two illustrations. Therefore, courts within the Second Circuit have misconstrued the illustrations by interpreting them to be the outer limit of “use in commerce.”
Second, as has been shown, the definition of “use in commerce” in section 1127 is inapplicable as a limiting principle in determining whether a particular use of a mark or device is an infringing use under the Lanham Act. Thus, the approach taken by courts within the Second Circuit turns out to be partially unsatisfactory. This, however, does not mean that their approach is totally meritless.
VII. Is Any Use of a Mark or Device In Commerce Actionable Under the Lanham Act?
The short answer appears to be yes, as long as it is within the bounds of Congress’ Commerce Power. The long answer requires reflecting on: (1) the history of trademark law, specifically the environment in which the Lanham Act was born; (2) the policies of trademark law; and (3) issues of judicial economy.
A. The Historical Underpinnings of the Lanham Act
The Lanham Act was developed and enacted in a world of physical marketplaces. For this reason, the Lanham Act targets source or sponsorship confusion that arises when a consumer views a mark on a product’s label(s) or in advertising. It is important to note, however, that the Lanham Act targets only confusion arising from the use of similar marks. Thus, the Lanham Act does not target source or sponsorship confusion arising merely from the placement of similar but distinctly labeled products next to each other. If the Lanham Act did target this type of confusion, our domestic system of supermarkets would be illegal because similar products could not be shelved side-by-side. Imagine a world where one would have to look down 12 different aisles to compare brands of toothpaste!
This type of proximity-based confusion can be best understood with a simple illustration. Suppose a consumer walks into a grocery store and asks for deodorant. She is directed to the deodorant aisle where she encounters nearly identical products that are distinctly labeled. The nearly identical yet distinctly labeled products are shelved side-by-side. The consumer may become confused as to who is actually producing the several types of deodorant she is viewing. This confusion, however, is not actionable under the Lanham Act because the Lanham Act assumes that the distinct labeling will be a sufficient indication of source and sponsorship for each type of deodorant. Thus, although the consumer is actually confused in the abstract, the Lanham Act posits a limit on actionable confusion: confusion must arise from the similarity of the marks employed in the sale of goods or services, and not from the fact that the consumer is overwhelmed by product information arising from the proximity of products.
B. The Policies of Trademark Law
The policies of trademark law reinforce that there is a limit on actionable confusion. Courts have found at least four main policies that underlie trademark law: (1) protection of consumers from fraud; (2) increasing the quality of products by fixing the responsibility for a consumer’s experience; (3) reducing consumer search costs; and (4) protecting the goodwill of a mark holder.
These policies have been recognized as the policies of trademark law because they help further our domestic free-market system in important ways. First, protecting consumers from fraud creates consumer confidence that domestic markets convey accurate and honest information about products. This consumer confidence encourages consumers to participate in the markets. Second, reinforcing the use of trademarks as a quality assurance mechanism promotes free-market enterprise because it allows consumers to allocate their resources among only quality products, thereby weeding out inferior products. Third, reducing consumer search costs also promotes free-market enterprise because it facilitates consumer participation in our markets: that is, when consumers are better able to find products in organized markets, they will rely on those markets in the future. Fourth and finally, by protecting the goodwill of a mark holder, trademarks encourage supply side innovations by giving producers an incentive to make the highest quality products possible. Thus, the recognized policies of trademark law dovetail to support our free-market system.
Free-market systems by their nature, however, rely on free competition to function. One major mechanism necessary for free competition is the ability of consumers to compare products against each other before buying any given product. Thus, if products, especially less well-known products, could not be compared alongside well-known products because trademark law prohibited source confusion arising merely from product proximity, then our free markets would face severe operational obstacles on a macro level. Essentially, such a prohibition on source or sponsorship confusion would require significant modification, if not outright annihilation, of grocery stores, department stores, and other methods of retail shopping that allow for comparison shopping.
Thus, the policies of trademark law support the following limitation on the actionable use of a mark or device in commerce: the use of a mark or device is not actionable where source confusion arises merely from the proximity of the products.
C. Judicial Economy
Underlying the keyword advertising issue, at least in the Second Circuit, is the search for a justification for disposing of keyword advertising claims with a motion for failure to state a claim. If a keyword-triggered advertisement contains only content that is not likely to cause confusion with another’s mark, any confusion arising from the display of the ad would seem to be merely the result of the proximity of the search results displayed in response to the consumer’s search query. Thus, a court conducting a proper likelihood of confusion analysis should find that the keyword advertiser has not infringed the other’s mark.
Since the basic fact pattern set forth directly above should produce a finding of non-infringement in every instance, it seems odd to prohibit the resolution of keyword advertisement claims until summary judgment (at the earliest) and force courts to expend scarce resources conducting futile likelihood of confusion analyses. Accordingly, courts within the Second Circuit find that the mere use of another’s mark as a keyword advertising trigger is not a “use in commerce” under the Lanham Act.
Although these courts were wrong concerning their specific interpretation of “use in commerce,” their search for a legal theory that could allow them to dispose of keyword advertising claims as early as possible is commendable because it preserves precious judicial resources for meritorious claims. Thus, preservation of precious judicial resources also supports a limit on infringement such that the mere use of another’s mark as a keyword advertising trigger is not actionable.
VIII. Product Placement and the Concept of Referential Fair Use
The term “product placement” refers to the practice of placing one product next to another because such placement results in higher sales of the first product, either because the first product is a suitable alternative to the second product or because the first product is desirable to consumers who buy the second product. Product placement is ubiquitous in our free market system. Product placement occurs in drug stores, in movies and TV shows, and elsewhere. Sometimes product placement is consensual, as in movies, and sometimes product placement is non-consensual, as is generally the case in drug stores. Most important, non-consensual product placement is generally legal.
Non-consensual product placement, however, is essentially a form of trading off the goodwill of another’s mark. In fact, such product placement can occur only if a producer of goods or services uses another’s mark as a referential tool. That is, the producer must say, “I want to be next to good or service X.” The producer may convey this to a brick and mortar retailer, as in the case of an independent soft drink manufacturer requesting that a grocery store place its product next to a well-known soft drink, or the producer may be a brick and mortar retailer, as in the case of a grocery store placing its store brand products next to well-known national brand products. Because product placement is a form of preying on another’s goodwill, product placement would seem to be a prime candidate for trademark liability, but that has not been the case.
So long as the products involved in the product placement scheme are distinctly labeled, our society has found product placement to be legal. As has been explained, such schemes are legal because there is no actionable source or sponsorship confusion when confusion is due only to the proximity of the products. Moreover, such schemes are beneficial to our free-market system because product placement allows for comparison shopping, and comparison shopping is a major mechanism that helps free-market systems operate.
Keyword advertising is essentially a technologically advanced form of product placement. Although there are some differences between product placement that occurs in a brick and mortar store and product placement that occurs as a result of keyword advertising, only one difference seems material: decontextualization.
Product placement in brick and mortar stores occurs in a contextualized consumer marketplace. Consumers observe both products and product labels. Meanwhile, in the internet search engine context, consumers may observe only product descriptions, word marks, and hypertext links. Essentially, the internet consumer is only able to observe simple product labels and descriptions in an environment where there are minimal spatial cues that provide separation between labels.
Take, for example, the following Google search for “hastings entertainment store”:
Objectors to keyword advertising will decry the minimal separation between the “Coke Cola Store” sponsored link and the “GoHastings.com” search result, and will decry the fact that the sponsored link is placed above the desired search result. Objectors may also say that this type of labeling is inadequate to stop source or sponsorship confusion, even though the sponsored link’s content is distinct from the search results below it. These objections overlook the history of trademarks. Moreover, giving these objections controlling weight in the debate surrounding keyword advertising would be contrary to good policy.
History shows that early trademarks were simple, sometimes consisting of nothing more than simple words or simple “geometric symbols.” History also shows that simply but distinctly labeled products incurred no trademark infringement liability. Since the Lanham Act was born out of this history, it seems difficult to imagine that the Lanham Act would now impose trademark infringement liability on such simple yet distinct labeling. If liability were to be imposed on simple but distinct labeling, it would mark a radical shift in terms of the normative underpinnings of trademark law, enlarging the scope of activities that can be held to cause consumer confusion.
Further, finding this type of simple yet distinct labeling insufficient to avoid trademark infringement liability would significantly hamper internet search technology (at least until such technology could provide the contextualization that people encounter in retail stores). This is because keyword advertising provides an enormous amount of revenue to search engines, revenue that allows search engines to operate at little-to-no cost to web users and that is necessary for the advancement of search engine technology.
Moreover, a dominant trend in American commerce is the rise of online commerce. Internet search engines have contributed to this trend because they allow consumers to filter through the overwhelming amount of information on the internet. As such, search engines are becoming a fundamental part of American commerce. Limiting the ability of search engines to engage in product placement practices could significantly hinder our free-market system because consumers may be unable to easily compare products on the internet (without performing multiple searches), and thus consumers might make inefficient consumption choices based on incomplete market information. Therefore, policy supports finding this type of simple yet distinct labeling to be lawful in the context of product placement.
Because the mere use of another’s mark as a keyword advertising trigger should be legal, the question arises: how should federal courts dispose of keyword advertising claims? The ideal mechanism would be to allow federal courts to dismiss such claims on a motion for failure to state a claim. This is because, as was explained above, in a keyword advertising claim, the keyword advertisement displays only content that is not confusingly similar to another’s mark. Thus, a likelihood of confusion analysis would be futile and would waste precious judicial resources.
Therefore, this article argues that federal courts should adopt the concept of “referential fair use.” Unlike the recognized fair use defenses, which do not result in dismissal upon a motion for failure to state a claim, “referential fair use” should function as a limit on the types of use that are actionable under the Lanham Act. Essentially, the Second Circuit has laid out the boundaries of this limitation: (1) a mark must be used internally (2) in a way that does not communicate an impression of the mark to a consumer. In developing this concept, courts should find that a person does not communicate an impression of a mark to a consumer when that person presents a distinctly labeled product to the consumer in response to a request for another trademarked product, as can happen in keyword advertising.
Failing to recognize the concept of referential fair use doctrine in the internet context would have two primary effects. First, failing to recognize the referential fair use doctrine would undermine the basic premise of trademark law that simple, yet distinct marks are not capable of trademark infringement because such a failure would allow for the maintenance of trademark infringement suits (at least until the point of summary of judgment) based on a defendant’s use of simple, yet distinct trademarks. Second, a failure to recognize the referential fair use doctrine would hinder American commerce by prohibiting the dissemination of product information by search engines, eliminating the possibility of comparison shopping through search engines. Because internet commerce is increasing at a rapid rate, the prohibition of comparison shopping through search engines may become a significant obstacle to the operation of our free-market system, a system which trademark law endeavors to support.
Moreover, the brick and mortar equivalent of keyword advertising, product placement, has long been considered legal. Thus, it seems basic that adopting this concept is desirable and comports with traditional understandings of trademark law.
This article can be reduced to three basic points. First, the mere use of another’s trademark as a keyword advertising trigger should not be actionable under the Lanham Act. Second, courts should recognize something akin to a “referential fair use” limitation in order to dispose of claims based on the mere use of keyword advertising triggers. Third, keyword advertising claims should be actionable to the extent that the content displayed in the keyword advertisement is confusingly similar to another’s mark.
* Kevin Zeck, J.D. Candidate 2009, Gonzaga University School of Law. I would first and foremost like to thank Adam Leskis and my brother, Bryan Zeck, for developing the way in which I see the world; on the shoulders of these intellectual giants I stand. Among my legal peers, I would like to thank Megan New, Brian Sniffen, and Jefferson Boswell for developing the way in which I see the law. Among my legal teachers, I would like to thank Sheri Engelken for introducing me to intellectual property in such a clear and entertaining manner. Among my legal mentors, I would like to thank the Honorable Kevin Korsmo for providing me with invaluable advice whenever I asked. Finally, I must give thanks to the following authors for helping to shape the way in which I think: Dr. Jared Diamond, whose works taught me that anything can be studied scientifically, including law; and Dr. Richard Dawkins, whose works taught me that placing faith in science is worth it, even if the explanations derived therefrom are not wholly satisfactory on an emotional level.
. No. 06-4881-cv, 2009 WL 875447 (2d Cir. Apr. 3, 2009).
. Rescuecom Corp., 2009 WL 875447, at *1, *3.
. See id. at *1.
. See id. at *7.
. Id. at *5 (“Google’s utilization of Rescuecom’s mark fits literally within the terms specified by 15 U.S.C. § 1127.”).
. See id. at *7.
. See id. at *6.
. See id. (“It is not by reason of absence of a use of a mark in commerce that benign product placement escapes liability; it escapes liability because it is a benign practice which does not cause a likelihood of consumer confusion.”).
. See id.
. See id. at *6-*7.
. See id. at *5.
. See id. at *7-*16.
. See id. at *9.
. See id. at *15.
. See id.
. See id. at *14-*15. Another important reason that the panel chose this interpretation is because it is consistent with prior Second Circuit jurisprudence. See id. at *15 (“At least for this Circuit, especially given our prior 1-800 precedent, which applied the second sentence of the definition to infringement, the latter solution, according a different scope of application to the two sentences of the definition, seems to be preferable.”).
. See id. at *15.
. See 15 U.S.C. § 1127 (“The term ‘use in commerce’ means the bona fide use of a mark in the ordinary course of trade, and not made merely to reserve a right in a mark. For purposes of this chapter, a mark shall be deemed to be in use in commerce–(1) on goods when–(A) it is placed in any manner on the goods or their containers or the displays associated therewith or on the tags or labels affixed thereto, or if the nature of the goods makes such placement impracticable, then on documents associated with the goods or their sale, and (B) the goods are sold or transported in commerce, and
(2) on services when it is used or displayed in the sale or advertising of services and the services are rendered in commerce, or the services are rendered in more than one State or in the United States and a foreign country and the person rendering the services is engaged in commerce in connection with the services.”).
. S. Rep. No. 100-515, at 44-45 (1988), as reprinted in 1988 U.S.C.C.A.N. 5577, 5607-08 (emphasis added).
. See Rescuecom Corp., 2009 WL 875447, at *15.
. See infra Parts VI, VII.
. 414 F.3d 400 (2d Cir. 2005).
. See Rescuecom Corp., 2009 WL 875447, at *6 (“Google, supported by amici, argues that 1-800 suggests that the inclusion of a trademark in an internal computer directory cannot constitute trademark use. Several district court decisions in this Circuit appear to have reached this conclusion. See [sic] e.g., S & L Vitamins, Inc. v. Australian Gold, Inc., 521 F. Supp. 2d 188, 199-202 (E.D.N.Y. 2007) (holding that use of a trademark in metadata did not constitute trademark use within the meaning of the Lanham Act because the use ‘is strictly internal and not communicated to the public’); Merck & Co., Inc. v. Mediplan Health Consulting, Inc., 425 F. Supp. 2d 402, 415 (S.D.N.Y. 2006) (holding that the internal use of a keyword to trigger advertisements did not qualify as trademark use). This over-reads the 1-800 decision.”).
. See id.
. See infra Part VIII.
. See infra note 181 and accompanying text.
. See infra note 181 and accompanying text. It is interesting to note, however, that the panel in Rescuecom Corp. never took a direct shot at the primary reasoning of 1-800 Contacts, but merely holds that several district courts have “over-read” 1-800 Contacts. See Rescuecom Corp., 2009 WL 875447, at *6.
. FragranceNet.com, Inc. v. FragranceX.com, Inc., 493 F. Supp. 2d 545, 547 n.2 (E.D.N.Y. 2007).
. See, e.g., Brookfield Commc’ns, Inc. v. W. Coast Entm’t Corp., 174 F.3d 1036, 1065-66 (9th Cir. 1999).
. See Kurt M. Saunders, Confusion is the Key: A Trademark Law Analysis of Keyword Banner Advertising, 71 Fordham L. Rev. 543, 543-45 (2002).
. See id.
. Wikipedia entry for “Organic Search,” http://en.wikipedia.org/wiki/Organic_search (last visited on Feb. 8, 2009) (specifically noting the difference between organic search results in response to a user’s query and advertisement results).
. See Amy N. Langville & Carl D. Meyer, Google’s PageRank and Beyond: The Science of Search Engine Rankings 11-12 (2006) (describing elements of the web search process).
. See id.
. Id. at 12. The web crawling programs continuously access and copy webpages according to some sort of programming criteria. See id.
. See id.
. See id.
. See Wikipedia entry for “Meta element,” http://en.wikipedia.org/wiki/Meta_tag (last visited on Feb. 8, 2007). Metatags are the key descriptors of a webpage and are used to help web crawling programs “locate title, description, and keyword information in the page.” Langville & Meyer, supra note 35, at 203.
. See Langville & Meyer, supra note 35, at 12.
. See id. at 12-13.
. See id. at 4 (“Nearly all major search engines now combine link analysis scores, similar to those used by Google, with more traditional information retrieval scores.”).
. See id. at 12 (describing a “query module”).
. See id. at 5-7 (outlining Boolean models, vector space models, and probabilistic models).
. See id. at 10 (describing how early search engines were fooled by massive repetition of key search terms).
. See id. at 25-26.
. See id. at 26-28. The importance of a given hypertext link is determined by the popularity of the webpage containing the hypertext link. See id. at 36 (describing “importance” as the amount of times a web user finds herself at a given page after clicking on the hypertext links of other pages). Thus, determining the importance of any given link is a circular process, entirely dependent on the popularity of the underlying website, which is, in turn, dependent upon the number and importance of links going into and out of that website. Page and Brin made determining the popularity of any given website workable (and thereby made determining importance of a link possible) by using an iterative mathematical process that starts from the assumption that all webpages have equal popularity. See id. at 32.
. See id. at 12 (explaining the importance of the “structure index”).
. Their idea is brilliant and can be understood with the following analogy. A stranger arrives in a town one night and is looking for the best restaurant in town. He can look in the yellow pages of the phone book and see which restaurants are listed. He may choose a restaurant with the most eye-catching advertisement; and if he goes to that restaurant, he may discover that the restaurant is some chain restaurant with poor food quality. Alternatively, he might try to ask the residents of the town about their favorite restaurants. Since these residents are putatively more familiar with the town’s restaurants than he is, this method of inquiry is likely to be more accurate than looking at the yellow pages alone.
Popularity in the search engine context is essentially asking the town’s residents about restaurant quality but multiplied billions of times over. The mathematics behind popularity in the search engine context are stunningly complex for a lay person, but for a full explanation see generally Langville & Meyer, supra note 35, at 33-111.
. See Aaron Wall, How Search Engines Work: Search Engine Relevancy Reviewed, Seobook, June 13, 2006, http://www.seobook.com/relevancy/ (reviewing the relevancy algorithms used by Google, MSN, and Yahoo).
. Hitwise US – Leading Search Engines – December, 2008, http://www.hitwise.com/datacenter/searchengineanalysis.php (last visited Feb. 8, 2009) (showing that Google, MSN, and Yahoo conducted 93.96% of search engine queries in December of 2008).
. See, e.g., Google, Inc., Google Annual Report 2007 (Form 10-K), at 71 (Feb. 15, 2008) available at http://investor.google.com/pdf/2007_google_annual_report.pdf (reporting that approximately ninety-nine percent of Google’s revenue in 2007 came from advertising, i.e., approximately 16.4 billion dollars of Google’s 16.5 billion dollars of revenue came from advertising).
Much of Google’s advertising revenues come from two sources: its AdWords program and its AdSense program. Greg Lastowka, Google’s Law, 73 Brook. L. Rev. 1327, 1348 (2008). The AdWords program is responsible for keyword advertising that appears on Google’s proprietary webpages, and the AdSense program is responsible for keyword advertising on independent third party websites. See id. at 1348-49. Under the AdSense program, Google pays independent third party websites to display keyword advertising generated by Google’s AdWords program. Id. at 1349. A significant majority of Google’s revenue come from keyword ads displayed on Google’s proprietary websites. Id. at 1340, 1348.
. See, e.g., Lastowka, supra note 55, at 1339-40 (describing Google’s AdWords program and GoTo.com’s same system); see also FragranceNet.com, Inc. v. FragranceX.com, Inc., 493 F. Supp. 2d 545, 547 n.2 (E.D.N.Y. 2007).
. See, e.g., Lastowka, supra note 55, at 1339-40.
. See, e.g., id. at 1340 (describing Google’s AdWords program).
. See, e.g., id. (describing Google’s AdWords program). Additionally, it is not uncommon for search engines to have an auction mechanism for providing keyword advertising. See, e.g., id. What this means is that once an advertiser has selected a keyword trigger, the advertiser must bid against other advertisers to obtain the right to have its advertisement displayed when a web user enters the keyword in a search query. See, e.g., id. at 1339-40 (comparing the now defunct GoTo.com’s system of selling search results with Google’s AdWords program). Advertisers conduct this auction by bidding up to the amount they are willing to pay the search engine for each click-through. See, e.g., id. at 1340 (describing the “fierce bidding war” for the keyword “mesothelioma”).
As an interesting sidenote, Google added a “wrinkle” to this system of bidding: even if one advertiser greatly outbid another advertiser, Google might display the lower bidder’s advertisement more often than the higher bidder’s advertisement if the lower bidder’s ad is clicked on more often than the higher bidder’s. See id. at 1340 n.63. Google has thus tried to maximize its revenues by adding a relevancy component as a backstop to its auction mechanism. Id. As a result of Google’s auction-plus-relevancy system, a totally unknown advertiser is unable to obtain a monopoly on the display of its keyword triggered ad by bidding up the price of a click-through to a prohibitively expensive level for other advertisers. If Google did not place this relevancy limitation on the auction function, an unknown advertiser could bid up the price of a click-through to a prohibitively expensive level and obtain a high level of visibility on the web because the advertiser would be sure that its ad would receive very few click-throughs.
. See, e.g., Google, Inc., Google Annual Report 2007 (Form 10-K), at 71 (Feb. 15, 2008).
. See, e.g., FragranceNet.com, Inc., 493 F. Supp. 2d at 547 (plaintiff alleged that defendant purchased plaintiff’s trademarks as keyword advertisement triggers).
. See id. at 547 & n.2.
. See id. at 547.
. See generally Stacey L. Dogan & Mark A. Lemley, Trademarks and Consumer Search Costs on the Internet, 41 Hous. L. Rev. 777 (2004) (outlining the benefits of keyword advertising to our civilization).
. 1 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 5:2 (4th ed. 2008).
. See id. § 1:12 (“[T]he term ‘passing off’ or ‘palming off’ is more properly reserved for those cases where the defendant has made an unauthorized substitution of the goods of one manufacturer when the goods of another manufacturer were ordered by the customer.”).
. See id. § 5:2.
. See 15 U.S.C. §§ 1114, 1125(a) (2006).
. See, e.g., id. § 1114.
. Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 157 (1989) (“The law of unfair competition has its roots in the common-law tort of deceit: its general concern is with protecting consumers from confusion as to source.”) (emphasis in original); Elvis Presley Enters., Inc. v. Capece, 141 F.3d 188, 193 (5th Cir. 1998) (stating that to prove likelihood of confusion, a plaintiff must show that a defendant’s mark is likely to cause confusion among “potential consumers as to the source, affiliation, or sponsorship of the [d]efendants’ [product]”).
. See, e.g., Ty Inc. v. Perryman, 306 F.3d 509, 510 (7th Cir. 2002).
. See, e.g., Two Pesos, Inc. v. Taco Cabana, Inc. 505 U.S. 763, 775 (1992) (recognizing only non-functional trade dress as protectable under the Lanham Act and holding that functionality occurs where trade dress “is one of a limited number of equally efficient options available to competitors and free competition would be unduly hindered by according the design trademark protection”).
. See Charles E. Lindblom, The Market System: What It Is, How It Works, and What to Make of It 23 (2001) (explaining that market systems provide the coordination necessary for the maintenance of a society); id. (explaining that market systems operate in the absence of a central planning mechanism); id. at 36-41 (explaining that market systems provide social coordination through the allocation of resources that result in long “chains” of cooperation); id. at 126 (explaining that poor allocation of resources results in social breakdowns); id. at 140 (explaining that market systems achieve efficient allocation of resources when they simultaneously allow many buyers and sellers to set prices and prohibit price-fixing).
. See id. at 140.
. See id. (explaining the mechanics of free-market systems generally); id. at 41 (stating that the market system is “the world’s broadest and most detailed organizer of social cooperation – with the longest arm and the most dexterous fingers”). It is interesting to note the similarities between free-market economic theory and the core idea behind the use of “popularity” in Google’s search algorithm. Cf. supra note 52 and accompanying text (explaining how Google’s use of “popularity” in its search algorithm allows society—as a collective entity—to determine the relevance of a website).
. See Lindblom, supra note 73, at 126, 140 (explaining that efficient distribution of resources results where “the presence of many buyers and sellers constrains monopoly and . . . no one has the power to fix prices” and that in the absence of such distribution, economies collapse).
. See id. at 126 (citing the collapse of communist economies and the Great Depression as resulting from economic breakdowns caused by the inefficient allocation of resources).
. See, e.g., Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 775 (1992).
. 1 McCarthy, supra note 65, § 5:4.
. Compare 15 U.S.C. § 1114(1) (1946) (“Any person who shall, in commerce, (a) use, without the consent of the registrant, any reproduction, counterfeit, copy, or colorable imitation of any registered mark in connection with the sale, offering for sale, or advertising of any goods or services on or in connection with which such use is likely to cause confusion or mistake or to deceive purchasers as to the source of origin of such goods or services . . . shall be liable to a civil action by the registrant . . .”), with 15 U.S.C. § 1114(1)(a) (2006) (“Any person who shall, without the consent of the registrant- (a) use in commerce any reproduction, counterfeit, copy, or colorable imitation of a registered mark in connection with the sale, offering for sale, distribution, or advertising of any goods or services on or in connection with which such use is likely to cause confusion, or to cause mistake, or to deceive . . . shall be liable in a civil action by the registrant . . .”).
Admittedly, the structure of 15 U.S.C. § 1125(a) has changed markedly. Compare 15 U.S.C. § 1125(a) (1946), with 15 U.S.C. § 1125(a) (2006). But the effect that this change might have on the premise that the Lanham Act was developed for a physical world is not assessed in this article.
. Cf. Edward S. Rogers, The Lanham Act and the Social Function of Trade-Marks, 14 Law & Contemp. Probs. 173, 182 (1949) (quoting Senator Claude Pepper as stating in a Senate subcommittee report on the Lanham Act in 1946 that one of the purposes of the Lanham Act was “[t]o modernize the trade-mark statutes so that they will conform to legitimate present-day business practice”).
. Coca-Cola Co. v. Overland, Inc., 692 F.2d 1250, 1252 (9th Cir. 1982).
. Id. at 1252 (explaining that the defendant’s behavior appeared to present a “clear-cut” case of trademark infringement).
. See, e.g., Brookfield Commc’ns, Inc. v. W. Coast Entm’t Corp., 174 F.3d 1036, 1066 (9th Cir. 1999); Promatek Indus., Ltd. v. Equitrac Corp., 300 F.3d 808, 812 (7th Cir. 2002).
. See, e.g., Promatek Indus., 300 F.3d at 810.
. See, e.g., Brookfield Commc’ns, 174 F.3d at 1062 (finding that the use of another’s trademark as a metatag was an infringement of trademark rights under the initial interest doctrine, even though the court acknowledge that no confusion would be likely if the web user were presented with search results that were distinct vis-à-vis each other).
. See Playboy Enters., Inc. v. Netscape Commc’ns Corp., 354 F.3d 1020, 1034-36 (9th Cir. 2004) (Berzon, J., concurring); see also infra Part VII.A (noting that the history of trademark law shows that simple yet distinct labeling is sufficient to avoid trademark infringement).
. Google went from non-existence in 1998 to controlling roughly seventy-two percent of the domestic search query market in December 2008. See Langville & Meyer, supra note 35, at 25-26; Hitwise US – Leading Search Engines – October, 2008, http://www.hitwise.com/datacenter/searchengineanalysis.php (last visited Feb. 28, 2009).
. The initial interest doctrine makes actionable source confusion that arises temporarily before the ultimate point of sale of a good or service. 4 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 23:6 (4th ed. 2008) (providing an overview of the initial interest confusion doctrine).
. See infra notes 92-93 and accompanying text.
. See infra notes 95-108 and accompanying text.
. See, e.g., Grotrian, Helfferich, Schulz, Th. Steinweg Nachf. v. Steinway & Sons, 365 F. Supp. 707, 717 (S.D.N.Y. 1973) (finding initial interest confusion between the marks “Grotrian-Steinweg” and “Steinway” as applied to pianos).
. See McNeil-PPC, Inc. v. Guardian Drug Co., 984 F. Supp. 1066, 1074 (E.D. Mich. 1997) (“That the customer might realize that Defendant’s product is not Plaintiff’s before he gets to the check-out counter to pay for it is irrelevant. Even if the consumer realizes that the . . . product is not the same . . . , Defendant has already accomplished what it set out to do, which is to confuse the consumer at the point when he first reaches for the product on the shelf. It is at that point that the damage is done. Defendant has already succeeded in utilizing the Plaintiff’s trade dress to get the customer to consider buying its product.”).
. Admittedly, the area of domain names is one in which the initial interest doctrine does translate well because the consumer always sees the domain name to which he or she navigates. See, e.g., Interstellar Starship Servs., Ltd. v. Epix, Inc., 304 F.3d 936, 944 (9th Cir. 2002).
. 174 F.3d 1036, 1065 (9th Cir. 1999).
. Id. at 1064.
. See Playboy Enters., Inc. v. Netscape Commc’ns Corp., 354 F.3d 1020, 1036 (9th Cir. 2004) (Berzon, J., concurring) (“I note that the billboard analogy has been widely criticized as inapplicable to the internet situation, given both the fact that customers were not misdirected and the minimal inconvenience in directing one’s web browser back to the original list of search results.”).
. See Brookfield Commc’ns, 174 F.3d at 1064 (assuming that the use of metatags will result in the display of a search result containing a link to a competitor’s website by positing that such use would result in the posting of a billboard along a highway for a competitor’s store).
. See supra notes 49-53 and accompanying text.
. See Brookfield Commc’ns, 174 F.3d at 1066-67.
. See id. at 1064.
. See supra note 41 and accompanying text. This observation is limited to the extent that tech-savvy consumers decide to investigate the source code of every website they visit.
. This worst case scenario assumes that the consumer uses a purely content-based search engine and that the competitor of the well-known movie store has effectively manipulated the search engine so as to exclude the well-known movie store’s website from being displayed among the search results retrieved by the search engine.
. See Playboy Enters., Inc. v. Netscape Commc’ns Corp., 354 F.3d 1020, 1034-36 (9th Cir. 2004) (Berzon, J., concurring).
. Indeed, this type of behavior resembles product placement techniques. See infra Part VIII, for an overview of product placement and its legality in American society.
. As noted in the Preface to this article, the Second Circuit’s decision in Rescuecom Corp. v. Google Inc., No. 06-4881-cv, 2009 WL 875447 (2d Cir. Apr. 3, 2009), substantially impacts this part of the article. For a detailed explanation of this impact, see supra Part I.D.
. See, e.g., FragranceNet.com, Inc. v. FragranceX.com, Inc., 493 F. Supp. 2d 545, 555 (E.D.N.Y. 2007) (finding that the use of another’s mark as a keyword advertising trigger and as a metatag is not actionable under the Lanham Act); Site Pro-1, Inc. v. Better Metal, LLC, 506 F. Supp. 2d 123, 127 (E.D.N.Y. 2007) (finding that the use of another’s mark as a keyword advertising trigger and as a metatag is not by itself actionable under the Lanham Act); Rescuecom Corp. v. Google, Inc., 456 F. Supp. 2d 393, 403 (N.D.N.Y. 2006) (finding that Google’s practice of selling trademarks as keyword advertising triggers is not by itself actionable under the Lanham Act).
These courts have relied heavily on the Second Circuit’s opinion in 1-800 Contacts, Inc. v. WhenU.com, Inc., 414 F.3d 400, 412 (2d Cir. 2005), which found that the use of another’s mark in a pop-up advertisement directory is not actionable under the Lanham Act. See, e.g., FragranceNet.com, Inc., 493 F. Supp. 2d at 549-50, 552-55.
. See, e.g., Playboy Enters., 354 F.3d 1020 (finding that the use of another’s mark as a keyword advertising trigger is actionable in the abstract, both under an infringement theory and under a dilution theory); Finance Express LLC v. Nowcom Corp., 564 F. Supp. 2d 1160, 1173-74 (C.D. Cal. 2008) (finding that the use of another’s mark as part of a website’s source code, including metatags, and as a keyword advertising trigger is an actionable use in commerce under Ninth Circuit precedent); Designer Skin, LLC v. S & L Vitamins, Inc., 560 F. Supp. 2d 811, 817-20 (D. Ariz. 2008) (conducting a likelihood of confusion analysis despite explicitly failing to address the issue of whether the use of another’s mark as a keyword advertising trigger or metatag is actionable under the Lanham Act; and finding no likelihood of initial interest confusion).
. See T.D.I. Int’l, Inc. v. Golf Preservations, Inc., No. 6:07-313-DCR, 2008 WL 294531, at *4 (E.D. Ky. Jan. 31, 2008); Boston Duck Tours, LP v. Super Duck Tours, LLC, 527 F. Supp. 2d 205, 208 (D. Mass. 2007), overruled on other grounds by Boston Duck Tours, LP v. Super Duck Tours, LLC, 531 F.3d 1, 7-8 (1st Cir. 2008); J.G. Wentworth, S.S.C. Ltd. P’ship v. Settlement Funding LLC, No. 06-0597, 2007 WL 30115, at *8 (E.D. Pa. Jan. 4, 2007); Int’l Profit Assocs. v. Paisola, 461 F. Supp. 2d 672, 676-77 (N.D. Ill. 2006); 800-JR Cigar, Inc. v. GoTo.com, Inc., 437 F. Supp. 2d 273, 285 (D.N.J. 2006); Kinetic Concepts, Inc. v. Bluesky Med. Group Inc., No. SA-03-CA-0832-RF, 2005 WL 3068223, at *8 (W.D. Tex. Nov. 1, 2005); Rescuecom Corp. v. Computer Troubleshooters USA, Inc., 464 F. Supp. 2d 1263, 1266-67 (N.D. Ga. 2005) (denying a motion to dismiss a keyword advertising claim under Fed. R. Civ. P. 12(b)(6) until further development of the case’s facts); Gov’t Employees Ins. Co. v. Google, Inc., 330 F. Supp. 2d 700, 703 (E.D. Va. 2004).
. 1-800 Contacts, Inc., 414 F.3d at 412 (finding that the use of another’s mark in a pop-up advertisement directory is not actionable under the Lanham Act).
. See, e.g., FragranceNet.com, Inc., 493 F. Supp. 2d at 555.
. Id. at 549 (quoting 1-800 Contacts, Inc., 414 F.3d at 409)).
. Id. Compare id., with Hamzik v. Zale Corp., No. 3:06-cv-1300, 2007 WL 1174863, at *3 (N.D.N.Y. Apr. 19, 2007) (finding that the use of another’s mark as a keyword trigger is actionable where that mark is displayed in the ad triggered).
. Compare FragranceNet.com, Inc., 493 F. Supp. 2d at 550 (finding non-actionable a defendant’s purchase of a plaintiff’s marks for use as keyword advertising triggers because the plaintiff had failed to allege that the advertisements triggered incorporated the plaintiff’s mark or anything similar), with id. at 554 (distinguishing Hamzik, 2007 WL 1174863, at *2-3).
Although the Second Circuit has never spelled out this precise idea, it seems legally correct. For example, a defendant should be held liable for the internal use of another’s mark where such use is really a means to divert a consumer to an alternate source of goods or services and the consumer is unaware ab initio that the goods or services do not come from the mark owner. See, e.g., Coca-Cola Co. v. Overland, Inc., 692 F.2d 1250, 1252 (9th Cir. 1982) (affirming a finding of trademark infringement where a restaurant substituted Pepsi-Cola for Coca-Cola without orally notifying consumers of the substitution and where consumers had specifically ordered Coca-Cola); see also FragranceNet.com, Inc., 493 F. Supp. 2d at 554 (distinguishing Hamzik, 2007 WL 1174863, at *2-3).
. E.g., FragranceNet.com, Inc., 493 F. Supp. 2d at 549.
. See supra notes 113-117 and accompanying text.
. See, e.g., Site Pro-1, Inc. v. Better Metal, LLC, 506 F. Supp. 2d 123, 123-24, 128 (E.D.N.Y. 2007).
. See, e.g., Playboy Enters., Inc. v. Netscape Commc’ns Corp., 354 F.3d 1020, 1031, 1034 (9th Cir. 2004).
. See, e.g., Finance Express LLC v. Nowcom Corp., 564 F. Supp. 2d 1160, 1173-74 (C.D. Cal. 2008) (finding that the use of another’s mark as part of a website’s source code, including metatags, and as keyword advertising triggers is an actionable use in commerce under Ninth Circuit precedent).
. See, e.g., Playboy Enters., 354 F.3d at 1024-25.
. See supra Part IV.D.2.
. See Playboy Enters., 354 F.3d at 1034-36 (Berzon, J., concurring); Designer Skin, LLC v. S & L Vitamins, Inc., 560 F. Supp. 2d 811, 817-20 (D. Ariz. 2008) (finding no likelihood of initial interest confusion concerning the defendant’s use of plaintiff’s mark as a keyword advertising trigger because the defendant’s alleged actions were not a “bait and switch” tactic).
. 354 F.3d at 1034-36 (Berzon, J., concurring).
. Id. at 1035.
. Id. at 1035-36.
. See infra Part VIII.
. As noted in the Preface to this article, the Second Circuit’s decision in Rescuecom Corp. v. Google Inc., No. 06-4881-cv, 2009 WL 875447 (2d Cir. Apr. 3, 2009), substantially impacts this part of the article. For a detailed explanation of this impact, see supra Part I.D.
. See supra note 112 and accompanying text.
. See, e.g., FragranceNet.com, Inc. v. FragranceX.com, Inc., 493 F. Supp. 2d 545, 549-50 (E.D.N.Y. 2007).
. 15 U.S.C. § 1127 (2006).
. Cf. Wickard v. Filburn, 317 U.S. 111, 121-25 (1942) (finding that purely intrastate conduct, which has a substantial effect on interstate commerce in the aggregate, is subject to Congress’ regulatory powers under the Commerce Clause).
. 15 U.S.C. § 1127.
. S. Rep. No. 100-515, at 44-45 (1988), as reprinted in 1988 U.S.C.C.A.N. 5577, 5607-08 (emphasis added).
. See, e.g., Rescuecom Corp. v. Google, Inc., 456 F. Supp. 2d 393, 400, 403 (N.D.N.Y. 2006).
. See, e.g., id. at 403 (“Thus, in the absence of allegations that defendant placed plaintiff’s trademark on any goods, displays, containers, or advertisements . . . plaintiff can prove no facts in support of its claim which would demonstrate trademark use.”).
. See supra notes 135-136 and accompanying text.
. See supra notes 135-138 and accompanying text.
. See S. Rep. No. 100-515, at 45 (1988), reprinted in 1988 U.S.C.C.A.N. 5577, 5608 (stating that “[c]learly, however, use of any type will continue to be considered in an infringement action”).
. See supra Part VI.
. See supra notes 79-81 and accompanying text.
. See 15 U.S.C. §§ 1114, 1125(a) (2006) (making the touchstone of trademark infringement the use of a mark or device that is “likely to cause confusion, or to cause mistake, or to deceive” as to source).
. See, e.g., Ty Inc. v. Perryman, 306 F.3d 509, 510 (7th Cir. 2002).
. 1 McCarthy, supra note 65, § 2:3 (quoting former attorney general William Baxter as stating: “Trademarks play a crucial role in our free market economic system.”).
. See Stefania Fusco, Geographical Indications: A Discussion on the Trips Regulation after the Ministerial Conference of Hong Kong, 12 Marq. Intell. Prop. L. Rev. 197, 234 (2008) (generally discussing the economic value of consumer confidence to free-market systems).
. See id.
. See 1 McCarthy, supra note 65, § 2:4.
. See id. § 2:5.
. See id. § 2:4.
. See supra notes 73-77 and accompanying text.
. See Jeff Sovern, Toward a Theory of Warranties in Sales of New Homes: Housing the Implied Warranty Advocates, Law and Economics Mavens, and Consumer Psychologists Under One Roof, 1993 Wis. L. Rev. 13, 68-69 (1993) (discussing home warranties and stating, “Economists have suggested that imperfect information gives market power to sellers; by contrast, more information in the hands of consumers increases competition. A corollary is that when consumers are unable to evaluate information properly, competition suffers. Studies show that increased comparison shopping leads to lower market prices, suggesting that greater comparison shopping promotes more competitive equilibria. More specifically, one theorist has argued that when sellers have more information about quality than buyers, the market will reach equilibrium at a suboptimal quality level. A scheme that corrects the information asymmetry should lead to a better allocation of resources.”).
. See supra Part V.A.
. I offer a caveat to this observation: if the search engine provided insufficient separation between search results, any confusion that would arise would be due to a “blending” of labels, which should be actionable under the Lanham Act. The story of Google, however, tells us that this should be a non-issue. That is, consumers would demand a different search engine if search results and keyword advertisements were blended together. Given the enormous profits earned from internet advertising, a different search engine that provides sufficient separation between search results would be expected to arise.
. Some courts in the Ninth Circuit, however, have found a likelihood of confusion in metatag and keyword advertisement cases using a likelihood of confusion analysis that is highly flawed. The flaw in their analysis specifically concerns the “similarity of marks used” factor. See, e.g., Playboy Enters., Inc. v. Netscape Commc’ns Corp., 354 F.3d 1020, 1028 (9th Cir. 2004).
In these cases, a defendant uses a competitor’s exact word mark as a metatag or keyword. See, e.g., id. The courts then find that the mark used by the defendant and the mark owned by the plaintiff are identical, regardless of the content actually displayed to the consumer, and that this weighs in favor of a finding of likelihood of confusion. See, e.g., id. Because the ordinary consumer never views the mark used by the defendant as a metatag or keyword, however, the similarity of the marks used should be determined by comparing the content actually displayed to consumers with the owner’s mark; that is, the analysis should be conducted from the perspective of the consumer. Cf. Malletier v. Burlington Coat Factory Warehouse Corp., 426 F.3d 532, 537-38 (2d Cir. 2005) (stating that the similarity of the marks should be evaluated according to the way consumers ordinarily view the marks at issue).
. See supra notes 141-145 and accompanying text.
. See Rescuecom Corp. v. Google, Inc., 456 F. Supp. 2d 393, 402-03 (N.D.N.Y. 2006).
. See id.
. See, e.g., Teenage Mutant Ninja Turtles (New Line Cinema 1990) (displaying ubiquitous references to Domino’s Pizza).
. See id.
. See Rescuecom Corp., 456 F. Supp. 2d at 402-03 (“For example, a drug store typically places its own store-brand generic products next to the trademarked products they emulate in order to induce a customer who has specifically sought out the trademarked product to consider the store’s less-expensive alternative.”).
. See, e.g., id. (implying that such practices are clearly legal).
. See supra Part VII (outlining, inter alia, the necessity of comparison shopping to our free-market economy).
. See supra notes 156-157 and accompanying text.
. See Rescuecom Corp., 456 F. Supp. 2d at 402-03 (comparing keyword advertising to product placement).
. As an interesting aside, the author entered this same search query a month after writing this article and was unable to obtain search results that contained any sort of sponsored link. This screen shot is reproduced in black and while with Google’s permission (email on file with Gonzaga Law Review).
 Editor’s Note, Wal-Mart Stores, Inc. v. Semara Bros., Inc., 90 Trademark Rep. 561, 561 (2000) (“The history of the Lanham Act and its protection of our nation’s consumers from confusion in the market place is chronicled by court decisions reflecting the rapid growth and expansion of trademarks, from simple geometric symbols to a multiplicity of different forms – symbols, words, packaging, product configurations, colors, sounds, smells and even holograms.”); see also 1 McCarthy, supra note 65, § 5:1.
. See, e.g., Sterling Remedy Co. v. Eureka Chem. & Mfg. Co., 80 F. 105, 107 (7th Cir. 1897) (finding no trademark infringement between two tobacco “cures” sold using the same channels of trade, in part because one was named NO-TO-BAC and the other was named BACO-CURO, and in part because of the dissimilarity of color used on labels).
. See supra note 55 and accompanying text.
. U.S. Census Bureau, E-Stats 1 (2008) (focusing on the year 2006 and noting that there were marked annual increases in all sectors of e-commerce), available at http://www.census.gov/eos/www/2006/2006reportfinal.pdf.
. See supra Part III.
. Cf. Charles Wheelan, Naked Economics: Undressing the Dismal Science 84-85 (2002) (explaining how market systems break down when consumers are unable to obtain minimum levels of product information by examining the used car industry).
. See supra Part VII.C.
. This term and concept can be attributed to Professor Eric Goldman of Santa Clara University School of Law. See Posting of Eric Goldman to Technology & Marketing Blog, http://blog.ericgoldman.org/archives/2007/06/ (June 28, 2007).
. There are two recognized fair use defenses in trademark law: (1) classic fair use; and (2) nominative fair use. The classic fair use defense requires that a defendant use a name, term, or device, in a manner otherwise than as a mark, that “is descriptive of and used fairly and in good faith only to describe the goods or services of [the defendant], or their geographic origin.” 15 U.S.C. § 1115(b)(4) (2006).
Turning to nominative fair use, there is both a statutory fair use defense and a non-statutory fair use defense. The statutory nominative fair use defense requires that the defendant use a name, term, or device, in a manner “otherwise than as a mark, of the party’s individual name in his own business, or of the individual name of anyone in privity with such party.” Id. The non-statutory nominative fair use defense was developed by the federal courts and has three elements that must be satisfied:[f]irst, the 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 product or service; and third, the user must do nothing that would . . . suggest sponsorship or endorsement . . .
Playboy Enters., Inc. v. Welles, 279 F.3d 796, 801 (9th Cir. 2002).
. See 1-800 Contacts, Inc. v. WhenU.com, Inc., 414 F.3d 400, 409 (2d Cir. 2005); see also FragranceNet.com, Inc. v. FragranceX.com, Inc., 493 F. Supp. 2d 545, 549-50 (E.D.N.Y. 2007) (applying the reasoning from 1-800 Contacts, Inc., 414 F.3d at 409).
. I offer a caveat to this observation: if the person offering the defendant’s distinctly labeled product affirmatively told the consumer that the defendant’s product was affiliated with the plaintiff’s business, this would be a form of trademark infringement.
. See supra notes 172-173 and accompanying text.
. See supra notes 175-177 and accompanying text.
. See supra note 175.
. See supra notes 72-77, 177 and accompanying text.
. See supra notes 162-169 and accompanying text.