- What are the intellectual property implications of paid keyword advertising?
On March 25, 2013 Google announced that on April 23, 2013, it would expand the ability of advertisers to purchase competitors’ trademarks as keywords. From April 23 forward, keywords Google previously restricted from purchase by third parties due to trademark issues and legal concerns would no longer be limited in China, Hong Kong, Macau, Taiwan, Australia, New Zealand, South Korea, and Brazil. That is, Google would not prevent the use of trademarks as keywords in the affected regions, although trademark owners could still take issue and contact Google for action regarding trademark use in advertisement text.1
For all intents and purposes, in the United States Google solidified and settled its legal bases for this same policy in October 2012, when language learning software company Rosetta Stone settled its trademark suit against the technology giant.2
The Internet has spawned a number of novel intellectual property issues. This article considers one of them: use of trademarks in the context of paid search advertising, which was addressed in 2012 in the Fourth Circuit’s Rosetta Stone Ltd. v. Google, Inc.3 Without expressing an opinion on Google’s paid search program, the Fourth Circuit reversed the grant of summary judgment for Google, finding Rosetta Stone had raised genuine issues of material fact regarding its direct and contributory trademark infringement and trademark dilution claims, and roundly rejected the district court’s application of the functionality doctrine in its initial decision.
As Rosetta Stone and Google released a joint statement in October 2012 through Reuters stating the companies planned “to collaborate to combat online ads for counterfeit goods” rather than continue the litigation, we cannot be sure how the district court would rule on these issues.4 However, the effect of the settlement, along with Google’s similar legal efforts and settlements around the world, facilitated Google’s implementation of its new, broader keywords policy, enabling it to provide what it calls, “A consistent policy and user experience worldwide benefits users, advertisers and trademark owners alike.”5 Most important for Google, its efforts appear to “legitimize” billions of dollars of Internet keywords advertising revenue.6
We will see how the Google keywords program policy develops. From a legal point of view, district and appellate courts have weighed the interests of the parties involved, the public, the law, and policy, to arrive at general conclusions, while leaving critical issues unclear or unresolved, and with only a slight indication there are answers forthcoming. Those issues remain unresolved today as Rosetta Stone settled its trademark suit against the search giant, with the terms of the settlement agreement kept confidential. Still, the three-year-old case was one of the highest-profile legal battles over a practice that has raised the ire of many trademark holders. In the United States, Google allows AdWords advertisers to bid on trademarked keywords, so someone searching for “Rosetta Stone” could be served ads for competitors. Now that’s also the case elsewhere.
This article starts with an overview of paid keyword searching, a discussion of how Google AdWords has been treated by the courts, and a summary of the Rosetta Stone litigation. Part Two, which will be released next month, will then look in depth at the Fourth Circuit’s ruling. It will also examine open questions and issues that may arise in the future in relation to keyword advertising.
I. Paid Search, Google AdWords, and Potential Issues
Obviously, one of the easiest ways to find information online is to use a search engine. Users enter a search term, or keyword, in a search field and submit it to trigger the search.7 The search engine then returns a list of websites or other results determined to be relevant based on the search engine’s proprietary search algorithm—these are “natural” or “organic” results.
A. Google AdWords Program
In addition to organic results, search engines like Google also display paid advertisements on the search results page. Like natural search results, keywords trigger paid advertisements, also known as sponsored links. After a Google search is completed, sponsored links are presented in a colored box just above and also to the right of the natural search results, separated by a vertical line or appearing in a lightly outlined box labeled “Sponsored.”
In the Google AdWords program,8 advertisers buy the right to have their ads triggered for display alongside organic search results when users submit relevant keywords. Advertisers generally pay per click, meaning they pay when users click through the sponsored links.9
To advertise on Google, advertisers need to register an account and agree not to violate any laws or engage in any illegal or fraudulent business practices.10 Until 2004, Google barred advertisers from using trademarks in ad text or as keywords upon request of the trademark owner. Google relaxed its guidelines in 2004 to allow advertisers to use others’ trademarks as keywords, even if the trademark owner objected.11 Google also developed a Keyword Tool that suggests trademarks for advertisers to purchase to trigger their ads.12 However, Google’s policy still was to bar the use of third-party trademarks in ad text if the trademark owner objected.
In 2009, Google implemented guidelines to permit trademark use in ads where (1) the sponsor is a reseller of a genuine trademarked product, (2) the sponsor makes or sells component parts for a trademarked product, (3) the sponsor offers compatible parts or goods for use with the trademarked product, or (4) the sponsor provides information about or reviews a trademarked product.13 Google adopted these guidelines after it developed technology designed to ensure that an advertiser’s use of a given trademark complied with AdWords policy.14
Beginning in 2003, in addition to its other marketing efforts, Rosetta Stone conducted millions of dollars in online advertising, including using Google’s AdWords program to great success.15 At the same time, Google worked diligently to address trademark and counterfeiting issues: it has maintained a “Safety and Enforcement Team,” headed by a former Secret Service Agent, to address counterfeiting and fraud issues, respond to notices of counterfeit ads, and remove ads confirmed to violate its AdWords Program—including many that Rosetta Stone brought to its attention.16
Between September 3, 2009, and March 1, 2010, Rosetta Stone claims to have reported to Google nearly 200 instances of paid search results that linked to counterfeit Rosetta Stone products, and it claimed that users looking for genuine Rosetta Stone software were confused by the paid search results and ended up buying counterfeit software.
II. District Court Proceedings
Rosetta Stone sued Google in the Eastern District of Virginia in August 2010, alleging direct trademark infringement, (2) contributory trademark infringement, (3) vicarious trademark infringement, (4) trademark dilution, and (5) unjust enrichment. Google filed a motion for summary judgment on all claims except unjust enrichment, which it moved to dismiss.
A. Direct Trademark Infringement
To prevail on a trademark infringement claim, a plaintiff must (1) establish that it has a valid mark entitled to protection, (2) that the defendant used the same or a similar mark (3) in commerce (4) in connection with the sale or advertising of goods or services (5) without the plaintiff’s consent, and (5) defendant’s use of the mark is likely to confuse consumers as to the origin of the goods or service.17
On summary judgment, the parties did not dispute the first four trademark infringement elements. The district court therefore turned its attention directly to whether Google’s policies and practices were likely to create consumer confusion as to source or origin of Rosetta Stone’s goods or services. Under Fourth Circuit law, nine factors generally are considered relevant to determine likelihood of confusion: (1) strength or distinctiveness of the mark, (2) similarity of the two marks to consumers, (3) similarity of the goods or services that the marks identify, (4) similarity of the facilities used by the markholders, (5) similarity of advertising used by the markholders, (6) defendant’s intent, (7) actual confusion, (8) quality of the defendant’s product, and (9) sophistication of the consuming public.18
The district court noted that only three of the confusion factors were in dispute (1) defendant’s intent, (2) actual confusion, and (3) sophistication of the consuming public—and it confined its analysis to those factors.19
The district court concluded that no reasonable juror would find that Google intended to confuse potential purchasers of Rosetta Stone’s products by auctioning Rosetta Stone marks as keywords or allowing their use in accordance with AdWords policy. The court also saw no evidence that Google was passing off its own goods or services as Rosetta Stone’s; it noted that Google sells advertising space.20
To demonstrate Google’s intent, Rosetta Stone pointed out that Google knew that including the Rosetta Stone marks in the AdWords program and in sponsored link text would generate revenue. Still, the district court noted that mere financial gain alone does not demonstrate intent to confuse potential buyers. The district court also found that Google had no financial interest in consumer confusion because its success depends on users finding what they seek in their searches. Even if Google profits from higher click-through rates, long-term financial losses would outstrip immediate gains if it provided search services without regard to counterfeiting or confusion. Confusion is bad for Google’s business, and counterfeiters tend to use stolen credit cards to secure ads and are a drain on Google’s resources. Finally, Google revised its trademark policy only after it developed the tool to verify AdWords policy compliance, which, the court found, indicated an intent to avoid consumer confusion.21 Add the Safety and Enforcement Team into the picture, and it is difficult, if not baseless, to find an intent to confuse consumers.
2. Actual Confusion
With respect to actual confusion, the district court observed that evidence of actual consumer confusion is the best evidence of likelihood of confusion.22 But it found Rosetta Stone’s evidence of actual confusion insufficient.
Rosetta Stone presented testimony from five individuals regarding actual confusion. The court found this de minimis given that since the AdWords policy was revised in 2004 more than 100 million ad impressions had been triggered on Google using the Rosetta Stone marks.23 The court also discounted this testimony because (1) none of the ads found to be confusing complied with Google’s policies and (2) the witnesses all testified to knowing they were not purchasing directly from Rosetta Stone, but were confused about whether the merchandise was genuine or counterfeit.24
While the testimony easily could be found de minimis, the court’s discounting it seems unsustainable in that (1) compliant with AdWords policy or not, confusion resulted from ads served by Google, and (2) whether confusion is related to product source or purchase source rings of a factual issue for a trier of fact.
Regarding Google’s 2004 AdWords policy and sponsored links that included third-party trademarks, Rosetta Stone introduced evidence of Google employees being confused as to the authenticity of products advertised. But the court reasoned that because the internal studies did not include Rosetta Stone marks, the evidence was not probative.25 The district court’s reasoning appears again to be more of a factual conclusion than a legal determination.
Rosetta Stone’s anecdotal evidence of actual confusion and the consumer confusion survey evidence it presented the district court also came up short. The court set aside Rosetta Stone’s anecdotal evidence of confusion on the ground that the ads in question were not on Google, which is understandable, and it found that testimony by Google lawyers regarding screenshots presented at their depositions reflected mere uncertainty, not confusion, as to the source of the merchandise. This begs the question of what the real difference is between the two.
Finally, the court deemed Rosetta Stone’s consumer confusion survey evidence unreliable because it found the question of whether Google endorsed a sponsored link did not relate to actionable confusion under the Lanham Act.26 This is a sound conclusion considering the confusion issue relates to source, affiliation, and sponsorship of the goods or services, not the advertisement or sponsored link itself.
The district court distinguished Sara Lee Corp. v. Kayser-Roth Corp.27 on the ground that it involved direct competitors, unlike Google and Rosetta Stone, and the record contained consumer surveys with confusion rates of 30 to 40 percent.
3. Consumer Sophistication
Sophistication of the consuming public can be important because it can mitigate or even negate potential confusion. The district court found that the relevant consuming public was only potential buyers, not the public at large, and that the consumer sophistication factor favored Google because of the relatively high cost of Rosetta Stone’s products and the time commitment required to learn a language. This type of well-educated consumer, the court concluded, would research these products before purchase and would not likely be confused by any AdWords-related issues.28
There is little evidence to support the likelihood of confusion element in Rosetta Stone’s direct trademark infringement claim against Google. Summary judgment may have been appropriate, but the district court also made conclusions that were more factual determinations, rather than strictly legal conclusions. Those missteps left the court’s order open for attack.
The functionality doctrine provides that trademark law cannot protect subject matter “if it is essential to the use or purpose of the article or if it affects the cost or quality of the article.”29 The district court concluded that keywords “have an essential indexing function because they enable Google to readily identify in its databases relevant information in response to a web user’s query.”30 It found that keywords affect the cost and quality of the AdWords program because without them Google would need a different paid search delivery system that might cost more and be less effective. Moreover, the AdWords program encourages market competition and delivery of the best content to search users. The court thus held that Google’s use of trademark keywords in paid search was functional, which precluded a finding of infringement.31
The district court’s conclusions here are inadequate in that Google’s use does not fall under the functionality doctrine. The question of functionality would relate to Rosetta Stone and its use of its marks. Here, the district court could have discussed issues such as AdWords and Google’s indexing being nontrademark use, nominative, referential, and/or fair use. Instead, by stretching the functionality doctrine the district court created an appealable issue.
C. Contributory Infringement
Contributory trademark infringement is a dangerous issue for Google. It requires a plaintiff to show that the defendant intentionally induced another to infringe a trademark or continued to supply products to a party whom the defendant knew or had reason to know was engaging in trademark infringement.32
As evidence of Google’s contributory infringement, Rosetta Stone pointed to Google’s Keyword Suggestion Tool, which recommends trademarks to AdWords users, and to Google’s alleged allowance of known infringers to bid on the Rosetta Stone marks. But the district court held that the existence of a tool to assist advertisers to optimize ads does not indicate inducement to infringe.
Before it provides a list of terms to AdWords users, Google informs them that they are responsible for keywords selected and for ensuring that their use does not violate applicable laws. Also, according to the court, there was no evidence that Google supplied its service to any party it knew or had reason to know was engaging in trademark infringement.33 The court cited Tiffany, Inc. v. eBay, Inc., 34 for the proposition that without specific knowledge of particular infringements or potential infringements, liability for contributory trademark infringement will not lie.
The court concluded that Rosetta Stone had failed to show Google knew of particular alleged infringing activity by AdWords advertisers without taking remedial action. What is more, Google lacks the ability to detect offers of counterfeit Rosetta Stone products.35 And it was undisputed that Google actively worked to remove ads when notified and by Rosetta Stone.
Still, outside of Google’s acknowledged efforts, the district court was unpersuaded by Rosetta Stone’s contention that after it notified Google of certain infringing sites Google continued to permit the same advertisers to use the Rosetta Stone marks as keywords and in ad text for other sites. The court never says how this notice from Rosetta Stone does not indicate Google knew or had reason to know certain advertisers were engaging in trademark infringement. The district court may have reached the right conclusion about Google not being liable for contributory infringement, but Rosetta Stone appeared to have raised a legitimate factual issue for a jury to explore.
D. Vicarious Infringement
Vicarious liability for trademark infringement by Google was a stretch for Rosetta Stone to even try to demonstrate. It can be found where parties have an agency relationship or where the defendant and infringer exercise joint ownership or control over an infringing product.36 For vicarious trademark infringement to apply, Rosetta Stone had to show that, beyond merely suggesting keywords to AdWords users, Google had joint ownership or control over the alleged infringing ads.
Rosetta Stone’s vicarious trademark infringement claim failed because the court found no agency relationship between Google and AdWords users, nor did Google control its advertisers’ Sponsored Links or their use of the Rosetta Stone marks in text.37
E. Unjust Enrichment
Rosetta Stone’s unjust enrichment claim also had little chance. Unjust enrichment is quasi-contractual and requires a plaintiff to show (1) it conferred a benefit on the defendant, (2) the defendant knew of the conferring benefit, and (3) the defendant accepted or retained the benefit under circumstances which render it inequitable for the defendant to do so without paying for its value.38 The district court found that Rosetta Stone adequately alleged that it conferred a benefit on Google through Google’s use of the Rosetta Stone marks in the AdWords program, that Google derived a benefit from that use, and that Rosetta Stone requested that Google stop the use. However, the court dismissed the claim on the ground that Rosetta Stone did not allege that Google knew of the benefit such that it would pay Rosetta Stone to use the marks as keywords.39
The district court also held that the claim was barred by the Communications Decency Act, which makes interactive computer service providers immune from non-IP liability based on content created by third parties.40 Either way, the claim was appropriately dismissed.
Under the Federal Trademark Dilution Act, a trademark owner may obtain an injunction against a party that, after the owner’s mark is famous, commences use of a mark in commerce in a way that is likely to cause dilution by blurring or by tarnishing the famous mark.41
The district court found that while the Rosetta Stone marks have been famous since at least 2009, they need not have been famous when Google revised its trademark policy in 2004 and began selling the marks as keywords. Rosetta Stone must show only that at any time after its marks became famous, Google began using them in a manner that was likely to dilute the marks.
The district court found that “Insofar as Google does not sell language learning software, it cannot be held liable for trademark dilution,” and that Rosetta Stone failed to show trademark dilution “where its brand awareness has only increased and where the reputation of its Marks has not been harmed since Google changed its trademark policy in 2004.”42
In any event, the court found in the Fourth Circuit no claim for dilution by blurring exists where the defendant’s product only increases public identification of the plaintiffs’ marks. And because Rosetta Stone brand awareness had only increased since 2004, the distinctiveness of the marks had not been impaired, and Rosetta Stone could not show the AdWords policy was likely to cause dilution by blurring.43 The district court also found no dilution by tarnishment where there was no evidence that consumers who bought the allegedly counterfeit software had a lower opinion of the Rosetta Stone marks.
In the case of dilution, the district court’s finding no dilution by blurring or tarnishment was right, but its reasoning was not altogether accurate. Dilution only applies where a defendant commences using a mark after the plaintiff’s mark is already famous.44 So, if the Rosetta Stone mark were not famous in 2004, then the claim could not lie years after Google already had commenced “using” the mark in the AdWords program.
Like the other discussion and decisions at the district court level, the conclusions on dilution, let alone direct infringement, vicarious infringement, and unjust enrichment, were not wrong, but the steps to arriving at those conclusions left the decision open to appeal and reversal.
1. See “Updates to AdWords trademark policy” https://support.google.com/adwordspolicy/answer/177578?hl=en-GB&uls=en-GB (last visited 6/28/13).
2. Terry Baynes, Rosetta Stone and Google Settle Trademark Lawsuit, Bob Burgdorfer and Leslie Adler, eds., Reuters (Oct 31, 2012) http://www.reuters.com/article/2012/10/31/us-usa-court-rosettastone-google-idUSBRE89U1GE20121031; see Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012).
4. http://www.reuters.com/article/2012/10/31/us-usa-court-rosettastone-google-idUSBRE89U1GE20121031. “Rosetta Stone Inc. and Google have agreed to dismiss the three-year-old trademark infringement lawsuit between them and to meaningfully collaborate to combat online ads for counterfeit goods and prevent the misuse and abuse of trademarks on the Internet.”
5. See Updates to AdWords Trademark Policy, https://support.google.com/adwordspolicy/answer/177578?hl=en-GB&uls=en-GB (last visited June 28, 2013).
6. Eric Goldman, With Rosetta Stone Settlement, Google Gets Closer to Legitimizing Billions of AdWords Revenue, Forbes, http://www.forbes.com/sites/ericgoldman/2012/11/01/with-rosetta-stone-settlement-google-gets-closer-to-legitimizing-billions-of-adwords-revenue/ (last visited June 22, 2013).
7. Web Search Engine, WikiPedia: The Free Encyclopedia, available at http://en.wikipedia.org/wiki/Web_search_engine (last visited July 20, 2012).
9. See The Cost of AdWords, http://support.google.com/adwords/bin/answer.py?hl=en&answer=1704424 (last visited July 20, 2012).
10. See What Is Google’s AdWords and AdSense Trademark Policy?, http://support.google.com/adwordspolicy/bin/answer.py?hl=en&answer=6118 (last visited July 20, 2012); See AdWords Policy Center, http://support.google.com/adwordspolicy/bin/answer.py?hl=en&answer=1316548&rd=1 (last visited July 20, 2012).
12. Using the Keyword Tool to Get Keyword and Ad Group Ideas, http://support.google.com/adwords/bin/answer.py?hl=en&answer=2470029&topic=1713918&ctx=topic&path=1713917-1713909 (last visited July 20, 2012).
13. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 151 (4th Cir. 2012); see http://www.searchenginejournal.com/google-updates-trademark-policy-on-keywords-ads/10400/ (last visited July 20, 2012).
15. Brief for Appellant at 9, Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012) (No. 10-2007). (Between July 2007 and March 2010, Rosetta Stone made more than $27 million from orders placed by customers who used Google paid and organic referrals).