Trademark conflicts involving domain names on the Internet quickly led to the passage of new law -- the
Anticybersquatting Consumer Protection Act ("ACPA") -- to address problems of bad faith registration of
trademarked names by non-trademark holders. But the ACPA, the only law enacted to deal with trademarks on the
Internet, applies only to cybersquatting of domain names. Since trademark disputes related to the Internet arise in
a number of ways outside cybersquatting, it is important to understand how the historical doctrines of trademark
infringement and trademark dilution apply to situations outside the scope of ACPA.
The issues this article addresses are familiar questions to trademark law. As is the case outside the Internet, one must ask whether there is a likelihood of confusion and whether there is dilution. However, these questions are difficult to answer applying the old doctrines of trademark law, and raises new questions. For example, is initial interest confusion sufficient to find trademark infringement when trademarks are used in metatags? To what extent will the unauthorized use of a trademark in the path of a URL be considered a "fair use" of the trademark and therefore non-infringing? Can linking or framing cause dilution of a trademark? Will paid placement or keyed banner ads cause a likelihood of consumer confusion? When answering these and similar questions, the courts find themselves redefining the boundaries of the Lanham Act, and with the basic objectives of trademark law in mind, it is unclear where these boundaries should be set.
The Supreme Court has described the basic objectives of trademark law as twofold: 1) consumer protection, and 2) protecting trademark holders from competitors’ free riding.
Traditionally, the law’s emphasis was on consumer protection. However, both the latest legislative additions to the Lanham Act -- the Federal Dilution Act ("FDA") and ACPA -- are primarily concerned with free riding. Consequently, this has led to arguments that federal trademark law has shifted too far towards granting property rights in trademarks.
To protect consumers, the Lanham Acts holds liable any person who uses a registered trademark in connection with the sale of goods or services in a way that is "likely to cause confusion." Additionally, the so-called unfair competition provision of the Lanham Act (§ 1125 (A)) provides a similar protection from the use of an unregistered mark that is likely to cause confusion. For convenience, this article will not distinguish between unauthorized uses of registered or unregistered trademarks.
Traditional questions of trademark infringement are answered by the application of the court developed "likelihood of confusion" test. This test involves analyzing the facts of the case through a number of factors. These factors include: (1) the strength of the mark, (2) the proximity of the goods, (3) the similarity of marks, (4) evidence of actual confusion, (5) possibility of bridging the gap, (6) sophistication of the consumer and (7) good faith on part of the alleged infringer.
There is substantial case law relating to the application of the trademark infringement test in general, and the number of cases involving trademark infringement in cyberspace is also growing. Even though the Internet brings new elements into trademark infringement cases, old case law remains relevant to assessing whether an unauthorized use of a trademark on the Internet is an infringement.
According to § 1114 of the Lanham Act, a likelihood of confusion alone is insufficient to find trademark infringement. The trademark use must also be in commerce, and such use must be in connection with the sale, offering for sale, distribution, or advertising of any goods or services. Unless specifically argued by either of the parties, however, the courts seem to place little emphasis on these requirements.
While, the likelihood of confusion test mostly applies to a competitors’ unauthorized use of trademarks, the Federal Dilution Act ("FDA") also protects against a non-competitors’ use. The FDA applies where such use causes actual "dilution of the distinctive quality of the mark." It defines dilution as the "lessening of the capacity of a famous mark to identify and distinguish goods and services."
To establish dilution, a trademark holder must show that "(1) [defendants have] made use of a junior mark sufficiently similar to the famous mark to evoke in a relevant universe of consumers a mental association of the two that (2) has caused (3) actual economic harm to the famous mark’s economic value by lessening its former selling power as an advertising agent for its goods and services."
Moreover, there are additional ways an unauthorized use can be held to cause dilution of a famous mark, such as blurring or tarnishment. For the purposes of this article, no distinction will be made between the different forms of dilution. Further, unlike the trademark infringement test, the dilution doctrine does not require a likelihood of confusion. However, the protection granted by the dilution doctrine is limited to "famous" trademarks.
Before the Internet and the World Wide Web, identical or nearly identical trademarks could exist in different geographic areas at the same time without the issue of trademark infringement ever arising. This is no longer the case.
Imagine, before the Web, a west coast company selling spare car parts and doing business under the mark Thrifty. Another, completely unrelated company could use the same mark while engaging in the same kind of business on the east coast without causing consumer confusion, and therefore, without infringing on anyone’s trademark. Today, after taking the advice of computer savvy friends and family, both "Thrifty East" and "Thrifty West" have established an online presence, hoping to increase their sale of spare car parts. Thrifty West uses the domain name www.thrifty.com, while Thrifty East uses the domain name www.thriftycarparts.com. While not directly competing a decade ago, the two companies are now chasing the same customers.
This is one example of how the Web has brought people closer together; an east coast company is suddenly just a mouse-click away for a west coast customer. While this certainly has its advantages, it increases the risk of confusion by the use of similar trademarks in different geographic areas. In other words, in the past trademarks had geographic limitations based on where the mark was physically used. Today, due to the web, these same geographical limitations don’t necessarily apply.
In addition to contributing to erasing the geographic boundaries of trademarks, the Internet has also created new possibilities for unauthorized use of trademarks. The remainder of this article will address how traditional trademark law applies to the unauthorized use of trademarks in metatags (section 4), in third level domain names and in the path of a URL (section 5), related to links (section 6), in frames (section 7), and by search engines (section 8).
Can an invisible word create confusion? On the web, perhaps. Metatags, or other forms of buried or hidden code are not normally visible to the average Internet surfer. Still, some courts have held that use of metatags can result in trademark infringement. So what exactly is a metatag?
Webopedia defines a metatag as:
A special HTML tag that provides information about a Web page. Unlike normal HTML tags, meta tags do not affect how the page is displayed. Instead, they provide information such as who created the page, how often it is updated, what the page is about, and which keywords represent the page's content. Many search engines use this information when building their indices.
As this definition points out, some search engines use metatags to retrieve information about websites and their content. Therefore, website operators realize that metatags can be used to lure or deceive search engines by making unauthorized use of a famous trademark to attract attention to their website. In response, some search engines have chosen not to make use of metatags when indexing a web site.
The Ninth Circuit used the following analogy to describe the unauthorized use of another’s trademark in one’s metatag:
Using another’s trademark in one’s metatags is much like posting a sign with another’s trademark in front of one’s store. Suppose West Coast’s competitor (let’s call it "Blockbuster") puts up a billboard on a highway reading – "West Coast Video: 2 miles ahead at Exit 7" – where West Coast is really located at Exit 8 but Blockbuster is located at Exit 7. Customer’s looking for West Coast’s store will pull off at Exit 7 and drive around looking for it. Unable to locate West Coast, but seeing the Blockbuster store right by the highway entrance, they may simply rent there. Even customers who prefer West Coast may find it not worth the trouble to continue searching for West Coast since there is a Blockbuster right there.
The role of search engines is important when considering whether unauthorized use of a trademark in a metatag creates a likelihood of confusion. Only when using a search engine can a consumer be confused by a deceptive metatag. As search engines and their users become more sophisticated, the likelihood of confusion will decrease. For example, if I type in a query for Volkswagen hoping to find the official Volkswagen website, I likely will get over a million hits. Many will be official Volkswagen websites from different countries, but certainly not all of them. Further, the order in which the hits appear depends on the search engine. The "better" search engines tend to have the official websites listed towards the top of their search results. Therefore, the search result location is often an indication of whether the web site is official. Finally, most search engines include additional information regarding the websites they list, such as an excerpt from the content of the website and the title of the website. Such information will also decrease the potential for consumer confusion.
The two most often cited decisions when it comes to use of trademark in metatags are Brookfield Communications v. West Coast Entm’t Corp. and Playboy Enters., Inc. v. Welles. Neither case applied the traditional likelihood of confusion test to the issue of metatags. In Brookfield, the court applied an initial interest confusion test, while the court in Welles held that Welles’ use of Playboy in the metatags on her site was a nominative use, and therefore not infringing.
Further, in Brookfield,
the Ninth Circuit held that West Coast’s use of "moviebuff" in their domain
name created a likelihood of confusion.
However, the court also found that the resulting confusion from West
Coast’s use of movie buff in their metatags was not as great as where West
Coast used the "moviebuff.com" domain name.
According to the court it was difficult to say that a consumer was likely
to be confused about whose site he had reached or to think that Brookfield
somehow sponsored West Coast’s web site.
Nevertheless, the court found that West Coast had infringed Brookfield’s
trademark rights when they used "movie buff" in their metatags, because such
use would result in initial interest confusion.
The
Brookfield court
cited several earlier cases in support for the initial interest confusion test: Grotrian v. Steinway & Sons,
and Mobil Oil Corp. v. Pegasus Petroleum Corp. In Grotrian
the court declined to hold "that actual or potential confusion at the time of
purchase necessarily must be demonstrated to establish trademark infringement
under the circumstances of this case." In Mobil, the court held that a
successful plaintiff must show a "likelihood that an appreciable number of
ordinarily prudent purchasers are likely to be misled, or indeed simply
confused, as to the source of the goods in question." Similarly, additional cases relied upon by
the Ninth Circuit in Brookfield
seem to require a likelihood of confusion in the initial phases of the
deal. Thus, applying the reasoning of
these cases to the factual pattern in Brookfield,
confusion at the initial phase, when the consumer was searching for "movie
buff" at a search engine, should be required to find trademark
infringement. But rather than asking
whether there was a likelihood that consumers would be confused at that point,
the court chose to focus on West Coast’s improper appropriation of the goodwill
that Brookfield had developed in
its mark. Thus, in order to find initial
interest confusion after Brookfield,
it is unclear whether a finding of confusion is necessary, or whether improper
appropriation of goodwill is sufficient.
A
persuasive argument can be made that initial interest confusion should be
insufficient to find trademark infringement, at least in a situation where a
website makes unauthorized use of a trademark in its metatags. There are significant differences between the
real world and the cyberworld not clearly reflected in the analogy used by the
court in Brookfield. Accordingly, these differences should, under
certain circumstances, lead to different results. For example, in the analogy, the consumer is
driving on the highway looking for a West Coast store, when she sees a sign
that telling her where to go to find the store.
In the cyber world, the consumer sees no such sign. If the consumer does not know the name of the
West Coast store’s website, she might try to guess it’s URL, e.g. www.westcoastvideo.com. If successful, it doesn’t matter how many
times Blockbuster had put "west coast" in its metatags, she would never be
exposed to them. The consumer is exposed
to Blockbuster’s metatags only if she uses a search engine to search for "west
coast." Moreover, even under these
circumstances, search engines that don’t search metatags would fail to turn up
Blockbuster’s website on their search list.
Further,
even search engines that do search metatags will, for different reasons, often
list Blockbuster towards the bottom of their search list. Additionally, West Coast’s website would also
be on that search list, as would many other websites. The number of hits search engines retrieve on
general queries is often very high, and the consumer often does not initially
find the desired site. If she ends up at
the wrong site, the search results are still only as far away as the back
button on her browser. Therefore, a
consumer spends little effort to return to these results.
Finally,
most search engines include some information regarding the content of the
Blockbuster website (such as a small excerpt from the web site or the title of
the website) when listing it among their search results. In most cases, this information reveals the
true identity of the website. Therefore,
consumer confusion is often not the result of the unauthorized use of a
trademark in metatags, not even initial interest confusion.
Therefore,
the use of West Coast in metatags is arguably very different from posting a
false sign telling consumers where to exit to find West Coast. This difference between the real world and
the cyberworld magnifies the consequences between the use of a false road sign
and the use of a mark in metatags. In
the real world, the consumer would have to drive her car around looking to get
back to the highway, and then keep looking for West Coast. Conversely, in the cyber world she would only
have to push her back button on her browser.
These differences can be used to support two arguments: (1) that the
unauthorized use of trademarks in metatags is unlikely to create even initial
interest confusion, and (2) that the consequences of initial interest confusion
in the real world and the cyber world are different, and therefore, initial
interest confusion should be sufficient to find trademark infringement only in
the real world.
In
Brookfield, the
Ninth Circuit recognized that the fair use doctrine applies in cyberspace as it
does in the real world. This doctrine
allows for the unauthorized use of a competitor’s mark to truthfully identify
the competitor’s goods in a fair manner.
Typically, the doctrine allows for use of a competitors trademark in
comparative advertisements.
A
convincing argument can be made that, as a general rule, the unauthorized use
of trademarks in metatags should be held to a similar fair use standard as
comparative advertisements. Such an
argument is supported by the nature of metatags, and when and how a consumer is
exposed to them. A fair assumption is
that many people when they query about a trademark on a search engine, are not
only looking for the official site, but also are looking for competitors’ websites
or alternatives to the products/services offered by the trademark holder. The Blockbuster/West Coast example from Brookfield
is one such illustration. In that case,
the fair use doctrine would likely allow (1) West Coast to make a comparison
with Blockbuster’s rental prizes on their website, and (2) allow West Coast to
list their major competitors (with hyperlinks to their websites) and encourage
their visitors to compare their terms and conditions. If so, why shouldn’t the fair use doctrine
allow West Coast to list their competitors in their metatags? Should the list of competitors become unfair
just because it is invisible? Further,
is it logical to conclude that the unauthorized use of trademarks in metatags,
like that described in Brookfield,
normally lead to an improper appropriation of goodwill? Suppose a user chooses to convert from
Blockbuster to West Coast after having entered West Coast’s website to look at
their terms and conditions. The most likely
conclusion to be made is that the terms and conditions of West Coast, rather
than the goodwill of Blockbuster, is the main reason for her conversion. Therefore, the better argument is that the
change from Blockbuster to West Coast is not an example of actionable
misappropriation of goodwill, but rather a use that is fair and non-infringing under trademark law.
It
should be noted that the fair use doctrine defense is only available when the
mark is used fairly and in good faith.
This test permits trademark holders to prevent the unauthorized use of
their trademarks in metatags in inappropriate situations.
Further,
the Brookfield case briefly mentions the nominative use theory. The court held that "moviebuff" was not
descriptive of the content of West Coast’s website, because the term
"moviebuff" was not a descriptive term. It appears from the court’s decision that
West Coast could have used the term "movie buff" in their metatags. The court placed great weight on the fact
that "moviebuff" is not a word in the English language, while "movie buff" is
routinely used in the English language to describe a movie devotee. This distinction is not convincing. It is unlikely that someone would look up a
word in an English dictionary before entering a search term at a search engine. Conversely, many people would likely attach the
same descriptive meaning to "moviebuff" as "movie buff," or accidentally
misspell movie buff as moviebuff.
Contrary
to Brookfield, the nominative use
defense was successfully argued in Welles. In Welles, the Ninth Circuit held that
Welles’ use of "Playboy" and "Playmate" in the metatags of her website was a
nominative use outside the scope of trademark protection because it met all the
prongs of the New Kids test. As a former Playmate, she was not readily
identifiable without the use of "Playboy" or "Playmate,"
she did not use the marks more than necessary, and nothing in conjunction with
the use of the trademarks suggested an affiliation with or sponsorship by
Playboy. With regard to the second
prong, the court emphasized that their "decision might differ if the metatags
listed the trademarked term so repeatedly that Welles’ site would regularly
appear above PEI’s in searches for one of the trademarked terms." Similarly, even if the other requirements to
qualify for a fair use defense were met, the repeated use of a trademark in
metatags probably would not be a fair or in good faith use under the fair use
doctrine discussed above.
Another
logical question raised is, can the unauthorized use of a trademark in metatags
without the simultaneous use of the trademark on the website itself qualify for
a nominative use defense? Provided that
the term being used is descriptive of the content of the website, the answer
should be yes. For instance, the term
"Playmate" or "Playboy" would be no less descriptive of Welles’ site if they
were only used in the metatags of her site.
In
Welles, Playboy also argued that
Welles’ use of "Playboy" and "Playmate" diluted their trademark. The court held that for the same reasons uses
in comparative advertising are exempted from anti-dilution law, nominative uses
are also excepted. The court reasoned
that comparative advertising and nominative use of a trademark cause no harm.
Even outside the
comparative advertising and the nominative use cases, anti-dilution law will
not play an important role when it comes to the unauthorized use of trademarks
in metatags. A consumer would not
actually see the trademark that is used in the metatags, but only be referred
to a site from a search engine with the assumption that there is some
mentioning of the trademark on that site.
This would only in unusual circumstances result in an actual lessening
of the selling power of the trademark as required under the FDA.
The
use of a trademark in the path of a URL or in a third level domain name cannot
be easily compared to other real world uses. Courts have held that second level domain
names (SLD) are identifiers "somewhat analogous to a telephone number or street
address." Building on this analogy, the path of a URL
or a third level domain name could be described as something similar to an
extension number to a telephone, or a department or office within a street
address. However, an SLD, the path of a
URL and a third level domain name, is capable of conveying more information
than an address or a phone number.
Typically this is the case where the URL includes a trademark, a
business name or another descriptive term.
When considering whether a
trademark use causes a likelihood of confusion, the location of the mark’s use
within the URL is of great importance.
Consider the following example.
If someone made an unauthorized use of "Volkswagen" in an SLD, such as, www.volkswagen.com, such use would likely
lead to consumer confusion. However, should "Volkswagen" be used in the
third level domain name, such as a host name for a sub-network or computer
running under a computer hosting a different SLD (e.g. www.volkswagen.ucdavis.edu) the
likelihood of consumer confusion is lessened.
Further, when the trademark appears in connection with another source
identifier such as "ucdavis," the second source identifier puts people on
notice that they are not accessing an official Volkswagen website, but rather a
sub-network of the main UC Davis network.
The
most interesting questions concerning the unauthorized use of a trademark in
either a third level domain name or in the path of a URL involve the nominative
use or descriptive fair use defenses to trademark infringement.
In
New Kids, the Ninth Circuit explained that to qualify for a nominative
use defense, three requirements had to be met: (1) The product or service in
question must be one not readily identifiable without the use of the trademark,
(2) only so much of the mark or marks may be used as is reasonably necessary to
identify the product or service, and (3) the user must do nothing that would,
in conjunction with the mark, suggest sponsorship or endorsement by the
trademark holder.
In the Internet
context the question of nominative use has been addressed at the district court
level in Patmont Motor Werks, Inc. v. Gateway Marine Inc. The court held that the use of the trademark
"Goped" in the path of the URL of a website operated by a distributor of
Go-Peds was a "non-trademark use of a mark – a use to which infringement laws
simply do not apply …" The court also held that all three
requirements of the New Kids nominative use test were met because: (1)
Go-Peds were not identifiable without using the word "Go-Ped" (the defendant
was only referring to the brand of scooter that he had for sale); (2) Go-Ped
was used only to the extent necessary (the defendant did not, for instance use
the plaintiff’s logo on his website); and (3) Defendant’s website did not
indicate any sponsorship or endorsement.
Further, the court went on to "…find as a matter of law that such use
[www.idiosync.com/goped] does not suggest Patmont’s sponsorship or endorsement,
because the Go-Ped mark did not appear in the website’s domain name." The court explained that while a domain name
has significant importance in identifying the source of a website, the path of
a URL serves a different function. It
"merely shows how the website’s data is organized within the host computer’s
files. […] Nothing in the post-domain
path of the URL indicates a website’s source of origin …" On this basis the court entered summary
judgment for the defendant.
In
addition to Patmont, a "pre-cyber world" Ninth Circuit case sheds some
light on when the unauthorized use of a trademark qualifies as a nominative use
that is outside the scope of trademark infringement. In Volkswagen Aktiengesellschaft v. Church
the court upheld the district court’s finding of no trademark infringement. Douglas Church, the defendant in Volkswagen, ran an automobile repair shop under
the name "Modern Volkswagen Porsche Service."
Later he changed the shop’s name to "Independent Volkswagen Porsche
Service." The court held that the
defendant could only use the Volkswagen trademark to the extent such use would
not deceive the public. Specifically,
the court meant the use could not suggest to the defendant’s customers that he
was part of Volkswagen’s organization of franchised dealers and repairmen. The court emphasized the defendant’s
prominent use of the word ‘Independent,’ and the fact that he "did not use
Volkswagen’s distinctive lettering style or color scheme, nor did he display
the encircled "VW" emblem."
Based on the above, it is
logical to conclude that Church should be able to take his business online
under the URL www.churchautorepair.com/volkswagen
without becoming a trademark infringer.
Especially if (1) he has a disclaimer on the site stating that he is not
an authorized Volkswagen dealer or repairmen, or (2) if it is apparent from the
website that he is independent from Volkswagen.
Although
the nominative use defense may be the best defense to allegations of
unauthorized trademark infringement in the path of a URL, the fair descriptive
use defense can also be asserted. For
instance, if on their website, Pepsi compared their own products to
Coca-Cola’s, with frequently updated survey results and the like, Pepsi could
claim a fair descriptive use of the Coca-Cola mark as a defense. What if such comparisons were displayed under
the URL www.pepsi.com/cocacola? It is logical to argue that this use is also
a fair use, and therefore, Pepsi could make such use without becoming a
trademark infringer.
The
unauthorized use of a trademark in a third level domain or the path of a URL is
unlikely to cause dilution. However, there is a split among the courts
involving the registration of domain names that contain derogatory words in
connection with a trademark. Some decisions have favored the trademark
holders, while others favor the holders of the domain names. Consider whether the following website makes
an unauthorized use of McDonald’s trademark in the path of their URL: www.capitalistpigs.com/mcdonalds? First, such use of the McDonald’s mark is
certainly outside the scope of ACPA. The
true question is whether under the FDA such use causes actual dilution of the
McDonalds mark. While most disputes under the ACPA and
ICANN’s UDRP are resolved on the issue of whether there is a "bad faith" intent
to profit, this fictional dispute will have to be decided according to
traditional dilution law. Further, it is
likely to turn on whether such a site would actually lessen the selling power
of the McDonald’s mark.
In
Brookfield, initial interest confusion was found sufficient to find
trademark infringement when a trademark is used in metatags. Accordingly, if courts hold that initial
interest confusion is sufficient to find trademark infringement in cases
involving the unauthorized use of trademarks in the path of a URL or in a third
level domain name, they may determine that infringement can not be avoided merely
by using a disclaimer of sponsorship or endorsement. Further, the initial interest confusion
theory could potentially limit the application of the nominative use test. Also, where the trademark is used in the path
of a URL or in a third level domain name, one could argue that such use is
sufficient to indicate sponsorship or endorsement. This could lead to different results in
online cases with similar facts to Patmont and Volkswagen.
A
link is a connector used by a website operator to lead visitors to a website
somewhere else on the Web. It can be
used to link to another website operated by the same website operator, or to
one operated by someone else. Typically,
these links appear as blue underlined text, but they can also appear in
different colors or graphic images. When a links offers a connection to an
off-site web page that is not the home page of the site being pointed to, it is
typically called a deep link. Most web
site operators encourage others to link to their sites because it is thought to
increase traffic to their site. However, when a deep link bypasses the home
page, advertisements placed on that home page get less exposure. This has led people to argue that while most
linking should be lawful, deep linking should not. Theories of trademark infringement, trademark
dilution, copyright infringement and also different theories of unfair
competition, passing off and misappropriation are used to support the
unlawfulness of deep linking.
Ticketmaster
Corp. v. Tickets.com Inc. (a preliminary injunction case before the U.S.
District Court for the Central District of California) involved deep linking. In this case Ticketmaster ("TM") filed to
obtain an injunction against Tickets.com ("T.com"). Both TM and T.com are in the business of
providing tickets or information on how to obtain tickets. TM is the largest ticket brokeraga company in
the U.S., and sells tickets through retail ticket windows over the Internet. TM operates a main web page with several
directories that categorize events, venues, geographical areas and the like
that a customer can choose. To obtain
information on any specific event, a visitor must click through links to these
directories. T.com sells tickets, but to
a larger extent provides information regarding other locations where tickets
may be purchased. One way T.com
accomplishes this is to provide the user with a deep link to the TM site for
the specific event the user requested.
However, the deep link bypasses the TM main page. With regard to the various Lanham Act claims,
the court held that T.com did not pass itself off as TM, in part because T.com
on its website said that it could not sell the tickets but would refer the
buyer to another broker who could. It
seems clear that this court was of the opinion that TM would be unsuccessful
with any trademark claims against T.com.
Even
though linking, and especially deep linking in some cases, raises legal
questions regarding violating the rights of the linked party, a plaintiff is
more likely to be successful with traditional trespassing tort claims and
copyright claims than with claims of trademark infringement.
To understand whehter
framing can constitute trademark infringement, it is first necessary to
understand what frames are and how they work.
The use of "frames" allows:
a Web page creator to divide the Web browser
window into several separate areas. [...] The programmer of the Web page can
dictate what goes into each frame.
Commonly, a Web site designer creates a page that at all times displays
one frame containing the name of the Web site and other identifying
information. The other frames are then
controlled by the user. For example, a
Web site employing frames might always show the original Web site’s graphic
logo on the top of the page while allowing the user to view [a second] Web site
in a different frame. [...] The legal implications of this are complex. In the example just given, a Web surfer might
easily be confused concerning the relationship between [the second] and the
framing site.
By changing the facts
in Ticketmaster to include framing, we can illustrate how trademark
claims can arise in relation to the use of frames. Imagine that T.com, in addition to just deep
linking to TM’s websites, displayed the event purchase site of TM within a
frame on their own website. This arguably increases the likelihood that a consumer may think that TM has
endorsed T.com’s website, or that there is a relationship between the two
companies. Under this scenario, it is unclear whether T.com could avoid trademark infringement by posting a
disclaimer or notice informing the consumer of no such endorsement or
relationship. Most likely this use has
crossed the line of trademark infringement because (1) it suggest a commercial
sort of relationship between the two companies, and (2) T.com passes off on
TM’s goodwill by presenting their own advertisement and promotion in connection
with a service offered by TM.
However,
the unauthorized use of frames does not always result in trademark
infringement. Consider the following
example of a use of frames that probably does not create a likelihood of
confusion: the "Ask Jeeves" search engine.
When one clicks on a link provided by Ask Jeeves, the link is presented
within a frame on the Ask Jeeves’ website.
This technique allows Ask Jeeves to display their own logo and banner
ads, even after the user has clicked through to the website of his choice. However, Ask Jeeves would rarely be a competitor
of the operator of the website being displayed within their frame. Additionally, Ask Jeeves’ users are likely
aware of how it operates and, therefore, understand that it is not endorsed by
or related with the websites listed in their search results. Therefore, it is unlikely that Ask Jeeves is
a trademark infringer.
What may a search engine be paid to
do without becoming a trademark infringer?
For instance, may Yahoo! sell banner ads to Volkswagen that show up
every time someone searches for another brand of car such as Mazda or
Toyota. What if Volkswagen is
unsatisfied with a banner ad, and wants to be listed as a search result when
someone makes the same query for Mazda or Toyota? For the remainder of the paper, the situation
described in the first hypothetical will be referred to as "keyed banner ads,"
while the situation described in the second hypothetical will be referred to as
"paid placement."
When a search
engine allows for keyed banner ads or paid placements, a further question is
raised regarding who the possible trademark infringers are: the search engine,
the advertiser, or both? For our purposes, there is no need to
distinguish between the two possible infringing parties.
Finally, while
ISP’s are given limited protection from copyright liability under the Digital
Millennium Copyright Act, and protection from defamation and other
non-intellectual property state law claims arising from third party content
under the Communications Decency Act, no such immunity is given from trademark
liability. Therefore, there is no statutory bar form
holding search engines liable for either direct or contributory trademark
infringement.
The practice of
using keyed banner ads, by itself, is unlikely to cause a likelihood of
confusion as to source. However, a
distinction should be made from situations where the content of the keyed
banner ad causes a likelihood of confusion.
This latter situation is not addressed by this article. Even without source confusion, trademark
infringement can still be found when a search engine user is likely to think
that there is some form of sponsorship or affiliation between the product or
service displayed in the banner ad and the owner of the trademark. After Brookfield, even a finding of confusion at the
earlier stages of a purchasing process - leading the typical user to click on
the banner ad thinking that there is a connection - would be sufficient to find
trademark infringement. However, the
most convincing argument remains that keyed banner ads normally will not lead
to initial interest confusion, not even sponsorship or endorsement confusion. Further, for there to be a likelihood of
confusion regarding sponsorship or endorsement, the customer must know or
assume that the search engine makes use of keyed banner ads, and the customer
must believe that the search engine only allows the trademark holder or
affiliated companies to "key up ads to queries" to the trademark. This is unrealistic, because most people are
likely not to believe the latter.
A better, but
still unpersuasive argument for trademark infringement, regarding the practice
of keyed banner ads, is when the targeted advertisement is presented not as a
banner ad, but as a pop up advertisement.
Consider, for example, the situation where (1) a pop up ad appears when
a search is made for Volkswagen that states "Do you need parts for your
Volkswagen? Come to carparts.com," and (2) in order to view the results, the
user must click on the "Close This Window" button to make the pop up disappear.
A commonly used
search engine illustrates how paid placement works. If you go to www.google.com
and search for "Volkswagen" you are likely to encounter at least one "Sponsored
link," or in other words an example of paid placement. For example, my search for Volkswagen March
20, 2002, resulted in both www.carsdirect.com
and www.valuepricing.com listed as sponsored links above the
official Volkswagen website - www.vw.com. Even though these links are clearly presented
as "sponsored" by
Google, they enjoy a more prominent placement than other links. The reason for this is that the operators of
these websites have paid Google to have their websites placed at the top of the
search results.
Similar to keyed
banner ads, paid placement is unlikely to cause a likelihood of confusion as to
source. Therefore, the relevant question
is whether the practice of paid placement is likely to confuse consumers into
thinking that there is a relationship or affiliation between the trademark
searched for and the sponsored link. For
example, in the previous hypothetical, the relevant question should be whether
a user would be confused into thinking that there is a relationship between
Volkswagen and carsdirect.com or valuepricing.com. For two reasons the answer is no. First, it is clearly indicated that these
links are sponsored, or essentially paid for.
Second, a reasonable visitor to Google should understand that search
results necessarily include links to web sites that have no affiliation or
relationship with Volkswagen. While some
search engines are better than others in placing official websites at the top
of the search results, a reasonable user should still expect to review the
results in order to find the desired website.
However, what if a
search engine, in a search for Volkswagen, placed www.valuepricing.com at the top of their
search results without indicating that it is a sponsored link? Would this cause a likelihood of confusion? Depending on such factors as the strength of
the mark, the proximity of the goods in question, the possibility that the
trademark holder will bridge the gap, and of course whether the trademark
holder can prove actual confusion a finding of trademark infringement could
result.
One illustration
of a borderline case is a lawsuit filed early in 2002 by "Body Solutions"
against four different search engines. Body Solutions alleged that the search
engines, by allowing the purchase of "Body Solutions" as a keyword for paid
placement without regard to its trademark rights, were directly,
contributorially, and vicariously liable for the resulting acts of unfair
competition, trademark infringement, trademark dilution, misappropriation,
deceptive trade practices, unjust enrichment and tortuous interference with
prospective economic advantage.
Alta Vista was one
of the accused search engines. When I
conducted a search for "Body Solutions" on AltaVista, it provided me with four
hits in a category called "Products and Services." Below this category they listed their 770,
767 results to the search. All four web
sites listed in the "Products and services" category made comparisons between
Body Solutions and their own product.
Three of the websites even used the term "Body Solution" in the title of
their website. Also, three of the sites had a disclaimer
stating that they were not affiliated with Mark Nutritionals, the holder of the
Body Solutions trademark. Finally, the
official Body Solutions website (www.bodysolutions.com)
did not appear on the first page of results returned by Alta Vista.
Does the AltaVista
practice of paid placement cause a likelihood of consumer confusion? Unlike Google, AltaVista does not indicate
that the four links in the "Products and Services" category are paid
advertisements. While they are presented
in a separate category ("products and services"), there is no indication that
this is not a normal search category on AltaVista. Alta Vista’s use of a separate category, by
itself, is likely insufficient to prevent users from being confused. Rather, the opposite is likely true. When a search is made for a product and when
the search engine used has a category for "Products and Services," users would
reasonably expect that the product searched for would be among those listed in
that category.
Despite the
disclaimers on at least three of the websites listed by AltaVista, there could
still be at least initial interest confusion, e.g., Brookfield. However, AltaVista will likely argue that
since all the results in the "Products and Services" category clearly indicate
that a comparison with the Body Solutions product is made, it is apparent to
the searcher that there is no affiliation between Body Solutions and the
products promoted on the listed websites.
This case also illustrates how a
search engine may argue a fair use defense in a paid placement case. In fact, the AltaVista scenario seems
tailored for a fair use defense. As
previously mentioned, the paid placement advertisers make use of the Body
Solutions trademark in a comparative way. A question still remains whether the
comparative advertising that AltaVista facilitates goes beyond what can be
regarded as "fair" and in "good faith" within the meaning of the Lanham
Act. This is an unsettled question with
very little case law on this point.
Therefore, what it takes to qualify as a fair and in good faith is
uncertain. For example, a related example that is
clearly a non-infringing use is when search engines and other websites
facilitate comparison-shopping for their users.
In
addition to fair use, search engines using paid placement may also argue
nominative use as a defense. For
example, in Playboy Enterprises, Inc. v. Netscape Communications the
court held that Netscape’s and Excite’s co branded search engine’s practice of
selling keyed banner ads to the search terms "Playboy" and "Playmate" was a
non-trademark use of a mark to which the infringement laws simply do not apply. The court further found that neither the
defendants nor the advertisers with whom defendants contracted used Playboy’s
trademarks to identify their goods.
While questions remain as to whether the court was correct to hold that
"playboy" and "playmate" were used in a non-trademark way, or whether the court
was correct to emphasize that the search engines did not compete with Playboy (disregarding
the fact that the advertisers, in my opinion, were direct competitors of
Playboy), this case is a good example that the New Kids defense is available also to keyed banner ads and paid
placement cases.
In
Netscape, Playboy argued trademark dilution because its marks were
located near materials which it considered more sexually explicit than its own
material. The court held that the
factual premise behind this claim was incorrect because (1) the Playboy
material could not be sufficiently distinguished from the material offered for
sale in the keyed banner ad, and (2) Playboy’s material was sold near other,
allegedly more sexually, explicit material in adult bookstores. The court concluded that Playboy had failed
to produce any evidence that the capacity of its marks to identify and
distinguish goods and services had been lessened.
Assuming
the Body Solutions trademark meets the threshold requirement of being famous,
the facts of the Body Solutions lawsuit also raise a question of
dilution. When AltaVista listed four
competitors at the top of search results for Body Solutions, and failed to
display the Body Solutions website close to these listings, their behavior could
have diluted the selling power of the Body Solution trademark. Users of AltaVista who searched for Body
Solution could have associated plaintiff’s trademark with the brands of the
four listed competitors or with weight loss products in general. This could have disabled the holder of the
Body Solution trademark from distinguishing its product by the use of the
trademark. If so, the trademark holder
will likely succeed in a claim for trademark dilution.
It
is generally known that search engines, to a large extent, rely on advertising
to generate revenue. Further advertisers
will pay more for an effective advertisement.
Targeted or customized advertising appears to increase the effectiveness
of advertising. Therefore, issues
related to paid placement, keyed banner ads and other forms of targeted
advertisement are likely to continue to arise.
To generate
advertising revenue search engines, such as Alta Vista, rely on traffic to
their websites. Therefore, there is a
strong incentive to find a balance between their advertising practice and what
their users prefer. Whether this
incentive is sufficient for search engines to structure their websites in
compliance with trademark law is unclear.
This uncertainty will continue as long as it remains unclear whether
search engines’ paid placement and keyed banner ad practices are in compliance
with trademark law.