Rosetta Stone Ltd. v. Google, Inc.: Deciphering Trademark Issues in Keyword Advertising (Part 2)

Vol. 3, No. 3


  • What are the intellectual property implications of paid keyword advertising??


This article is the second part of a two-part article. Click here to access the first part.


III. Fourth Circuit Proceedings

On appeal, the Fourth Circuit vacated the summary judgment ruling as to direct infringement, contributory infringement, and dilution, and affirmed with respect to unjust enrichment and vicarious infringement.


A. Direct Infringement

The Fourth Circuit assumed that Google used the Rosetta Stone marks in commerce, which Google did not dispute, and so it directly considered the likelihood of confusion issue.

Rosetta Stone argued the district court committed reversible error by not considering all nine likelihood of confusion factors, but the Fourth Circuit stated the factors are only a guide and that it was not necessarily error not to address every one, although where the court decides certain factors are not relevant, it should provide a brief explanation of its reasons for omitting them.1

In particular, the court stressed problems with the mechanical application of every confusion factor in a case involving a referential, non-trademark use,2 i.e., where a defendant uses the plaintiff’s mark to identify the plaintiff’s goods or services.3 Application of the traditional multi-factor test is difficult in such cases because several factors are not probative of confusion.4 For instance, when a car repair shop runs an advertisement containing the trademarked names of the brands it services, factors such as the similarity of the marks and similarity of the products would necessarily point toward confusion.5

It is interesting to have the Appellate Court discuss the referential or nominative type of use in relation to the confusion factors—and how use can avoid confusion—before turning to the district court’s assessment of the intent, actual confusion, and consumer sophistication factors. In regard to those factors, the Fourth Circuit found the district court had improperly resolved fact issues in Google’s favor.6


1. Intent

On intent, the court noted that even though Google’s 2004 internal studies showed there was significant source confusion among Internet searchers when trademarks were included in paid search ads, it still relaxed its trademark policy further in 2009 without conducting any new studies. The court concluded that a reasonable trier of fact “could find that Google intended to cause confusion in that it acted with the knowledge that confusion was very likely to result from its use of the marks.”7

Here, the court did not consider any of the steps that Google has taken to mitigate confusion, including the technology it developed to ensure that an advertiser’s use of a given trademark complied with AdWords policy, its efforts to remove offending listings, and its commitment to work with brand owners. Even if confusion were to go up, it’s still questionable whether a reasonable trier of fact could find that Google intended that result.


2. Actual Confusion

As for actual confusion, the Fourth Circuit rejected the district court’s conclusion that evidence of confusion as to whether an item was genuine was not probative of actionable confusion and that ads that conformed to Google’s policies could not be a basis for liability. The appellate court noted that consumers are not privy to Google’s policies, and that confusion as to authenticity is relevant confusion under the Lanham Act and should not have been disregarded.8

In the Internet context, initial interest confusion is where a user is distracted or diverted away from a site he or she initially sought to go to another site, based on the user’s belief that the second site is associated with the one he or she originally sought.9 It is well settled that, initial interest confusion is actionable under the Lanham Act,10 even though Google claimed that initial interest confusion is not actionable in the Fourth Circuit.11 However, the Fourth Circuit’s requirement of examining allegedly infringing use in the context in which it is seen by the ordinary consumer,12 looks to have resulted in the court’s rejection of the district court’s holding.

The Fourth Circuit also disagreed that Rosetta Stone’s evidence of actual confusion was de minimis. Rosetta Stone submitted evidence that it received some 123 complaints from purchasers of pirated/counterfeit software they believed to be genuine. The court held that this evidence was not de minimis and that it would be reasonable to conclude that many users have been confused by an apparent relationship between Rosetta Stone and the sponsored links purchased by counterfeiters served since Google relaxed its AdWords policies in 2009.13

With respect to Google’s internal studies, the court noted that one study showed 94 percent of consumers were confused at least once when shown ads with trademarks in them. Whereas the district court found these studies inapposite because they did not involve Rosetta Stone marks, the Fourth Circuit found them probative of confusion arising from Google’s trademark use and, citing evidence that two of Google’s in-house trademark attorneys could not tell which sponsored links from a “Rosetta Stone” Google search results page were authorized resellers, concluded that there were triable issue of fact as to actual confusion.14

Finally, the court of appeals criticized the district court for rejecting Rosetta Stone’s survey expert on the ground that it measured whether respondent thought sponsored links were endorsed by Rosetta Stone and that confusion as to endorsement was not the same as confusion as to source or origin. The court pointed out trademark infringement “creates a likelihood of ‘confusion not only as to source, but also as to affiliation, connection or sponsorship’”15 and that confusion as to endorsement was therefore probative of likelihood of confusion.

Google may not intend it, and the survey evidence may not carry great weight to show confusion as to source, but the court is right that there are issues of actual confusion for the district court to consider.


3. Consumer Sophistication

Considering the consumer sophistication factor, the Fourth Circuit emphasized that the district court can reach a conclusion about consumer sophistication by inference and the nature of the product and its costs in the context of a bench trial, but not on summary judgment.16

The court found evidence from deposition testimony and from Google’s internal investigations reflecting confusion as to the nature of sponsored links by “even well-educated, seasoned Internet consumers,” such that the factor did not favor Google as a matter of law.17

*          *          *

Having found triable fact issues as to likelihood of confusion, in a footnote the court rejected Rosetta Stone’s contention that it was entitled to a presumption of confusion on the infringement claim. The court clarified that such a presumption applies only in cases of intentional copying, and that there was absolutely no evidence that Google intentionally copied or adopted Rosetta Stone’s mark in an effort to pass off its own goods or services under the ROSETTA STONE mark.”18


B. Functionality

The functionality doctrine bars trademark protection of any product aspect that is functional or makes it more useful.19 With that, the Fourth Circuit next rejected the district court’s functionality analysis, finding that because Rosetta Stone’s marks are not a functional feature of Rosetta Stone’s products, they are not functional. Once it is determined that a product feature (e.g., “ROSETTA STONE”) is not functional, then the functionality doctrine has no application, regardless of another party’s (e.g., Google’s) use. 20 That is, the words “Rosetta Stone” merely identify the source of the language-learning product; Google’s use of the mark does not fall under the functionality doctrine because the question of functionality would relate to Rosetta Stone and its use of the marks, not Google’s.

The district court could have addressed matters such as AdWords and Google’s indexing from a different standpoint, including nominative, referential, and fair use. Instead, we are provided little insight on how the district court really sees the AdWords issue playing out.


C. Contributory infringement

As discussed, the district court concluded that Rosetta Stone had failed to show Google knew of particular alleged infringing activity by AdWords advertisers without taking remedial action. That is, the district court was unpersuaded by Rosetta Stone’s notifying of Google of almost 200 cases of sponsored links advertising counterfeit Rosetta Stone products, and that Google continued to allow those same advertisers to use the Rosetta Stone marks as keywords for other Web sites and in sponsored link ad text.21 Comparing Rosetta Stone’s notice to Google with the knowledge attributed to eBay in Tiffany,22 the district court concluded that Rosetta Stone had not met its summary judgment burden and awarded summary judgment to Google.

But the Fourth Circuit found that the district court misapplied the standard of review because even if Rosetta Stone did not meet the summary judgment threshold, there still may be sufficient evidence to permit a trier of fact to find contributory infringement.23 The Fourth Circuit wrote that “. . . granting summary judgment to Google because ‘Rosetta Stone has not met the burden of showing that summary judgment is proper as to its contributory trademark infringement claim,’” turns the issue upside down.24 Even as Rosetta Stone could not meet the requirements for summary judgment, the next step for the district court should have been considering whether a reasonable trier of fact could find in favor of Rosetta Stone.

The matter will be revisited on remand, with Rosetta Stone seeking to prove Google had specific knowledge of infringers and infringements and Google challenging Rosetta Stone’s ability to reach the high evidence bar set by Tiffany.


D. Vicarious infringement

The Fourth Circuit agreed with the district court that Rosetta Stone’s evidence failed to create a question of fact as to whether Google acted jointly to control the counterfeit Rosetta Stone products.25


E. Dilution

Turning to dilution, the court of appeals disagreed with the two grounds on which the district court granted summary judgment: (1) that Rosetta Stone needed to present evidence that Google used the Rosetta Stone marks to identify its own goods and services, and (2) that the claim was defeated by evidence that Rosetta Stone’s brand awareness had increased since Google relaxed its AdWords program guidelines.26

As for the first ground, the court observed that the trial court construed the FTDA’s fair use defense erroneously as requiring Rosetta Stone to demonstrate that Google’s use was not fair as a source identifier for its own goods.27 The court stated that “[w]hether Google used the mark other than as a source identifier and in good faith is an issue that Google, not Rosetta Stone, is obligated to establish.”28 The court stated: “In essence, the district court made nontrademark use coextensive with the fair use defense under the FTDA. The statute, however, requires more than showing that defendant’s use was ‘other than as a designation of source’—the defendant’s use must also qualify as a ‘fair use.’”29

As for the second ground, the Fourth Circuit faulted the district court for basing its likelihood of dilution conclusion solely on the fact that Rosetta Stone’s brand awareness had increased since Google revised its AdWords trademark policy in 2004.30 The FTDA provides a non-exhaustive list of six factors for courts to consider in assessing likelihood of dilution by blurring: (1) the degree of similarity between the marks; (2) the degree of inherent or acquired distinctiveness of the famous mark; (3) the extent to which the owner of the famous mark is engaging in substantially exclusive use of the mark; (4) the degree of recognition of the famous mark; (5) whether the user of the mark intended to create an association with the famous mark; (6) any actual association between the marks.31 Although not every factor will be relevant to every case, the district court erred, in the Fourth Circuit’s view, in relying solely on the “degree of mark recognition.” The court directed that on remand the district court should address additional factors that might inform whether Google’s use is likely to impair the distinctiveness of Rosetta Stone’s marks.32

Finally, the court considered whether Rosetta Stone’s marks became famous before Google began using the mark in commerce.33 The district court had found that though 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. That is, according to the district court, at any time after its marks became famous Rosetta Stone only had to show that Google began using its marks in a way that was likely to dilute them.

The Fourth Circuit clarified that the date a mark becomes famous is the date from which a diluting use can commence. Prior use from before that date would not qualify as diluting, nor would use after that date if the use commenced prior to the mark becoming famous. That being the case, the court ordered that on remand the district court should determine (1) when Google made its first potentially diluting use of the mark and (2) whether the Rosetta Stone mark was famous at that point in time. If Google commenced use of the mark before the Rosetta Stone marks became famous, its use would not be dilutive or likely to dilute.34

Once the evidentiary issues are sorted out, the Fourth Circuit’s guidance will certainly point to the viability of this claim for Rosetta Stone.


IV. What Next? Thoughts on Cracking the Code

There were a variety of issues for the district court to consider on remand, but, with the case settling, those issues were never taken up by the court. Even more, that settlement appears to have settled the law in this area—at least for the time being, namely until someone comes along with the funds and initiative to sue Google over the AdWords program.

So, while Google may be safe, a fundamental consideration went totally unaddressed by the district court and the Fourth Circuit: the question of “use in commerce” and direct trademark infringement. While relatively settled, or near settled, in the Ninth35 and Second Circuits,36 the matter is up in the air in the Fourth Circuit and it has bearing on whether Google could be determined a direct infringer. Still, even with a finding of “use in commerce,” a court would also need to find a likelihood of confusion, which would be difficult in itself. For example, Google does not use the Rosetta Stone or other marks to identify its own goods or services, and yet the direct infringement claim looms mostly because other parties are using third-party trademarks on Google’s site.

There are a variety of contexts to recognize where a trademark could be deemed “used in commerce” and the “user” subject to a direct infringement claim, and other contexts where the situation precludes a finding of “use in commerce,” preventing liability for direct infringement. Mary LaFrance, Professor at William S. Boyd School of Law, noted (1) Google’s “sale” of marks as keywords to trigger paid search; (2) Google’s Keyword Suggestion Tool, proposing that AdWords users purchase certain marks as keywords; (3) advertisers’ actual use of another party’s marks as keyword triggers; and (4) advertisers’ actual use of another party’s marks in their ads.37 It is possible that some of these activities might satisfy the use requirement while others might not.

But, even if the Fourth Circuit had found Google’s inclusion of the Rosetta Stone marks in it AdWords program comprises “use in commerce,” the next pivotal issue is to show there is a likelihood of confusion caused by that use. In the context of Google as a direct infringer, the apparently best possibility of finding liability is Rosetta Stone demonstrating Google’s activities are likely confuse consumers as to Rosetta Stone’s approval or endorsement of Google’s goods, services, or commercial activities involving the Rosetta Stone marks. This scenario does not seem likely. AdWords users know what the program entails and would not be confused by Google’s sale of trademarks.

Looking at contributory infringement, Google may have had greater exposure because Rosetta Stone may have had an easier time demonstrating the underlying infringement by any number of AdWords users it knows are selling counterfeit software and creating actual confusion for consumers. Still, Google generally has taken steps to mitigate its risk and shield itself from liability, stating in its AdWords Policy that, “If a trademark owner files a complaint with Google about the use of their trademark in AdWords ads, Google will investigate and may enforce certain restrictions on the use of that trademark in AdWords ads and as keywords.”38 Further, Google notes in its advertising policy that, “All advertisers bear their own responsibility for understanding the laws applicable to their businesses.”39

The Tiffany standard of actual knowledge sets a high bar for attacking Google. And Google takes important steps to fight counterfeiting, but if Google were to provide known past infringers access to the Keyword Tool, or of it were to shift resources away from preventing trademark and counterfeiting issues, one could wonder if Google’s actions or inactions in addressing trademark issues brought to its attention would be outweighed by considerations of what could be deemed willful blindness on its part. The Tiffany standard is high, but under certain circumstances one might reasonably conclude that Google could know or have reason to know that certain AdWords users were committing actual infringement, almost just by their being in the program.

How it would all play out from a legal or AdWords perspective, is still unclear, but we know that a single infringement is different from ten infringements, which is different from a pattern of infringing behavior. And even then, ten infringements in a short period of time could possibly reach the “know or have reason to know” threshold.

We will have to wait and see if any of these questions even comes up—if Google weren’t so dominant in the court battles and capable of driving settlement, they surely would. In the meantime, AdWords policy stands and we’ll wait to see if the matter is ever really addressed.



1. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 154-155 (4th Cir. 2012).

2. Id. at 155.

3. Id. at 154; see Tiffany (NJ) Inc. v. eBay Inc., 600 F.3d 93, 102 (2d Cir. 2010).

4. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 154 (4th Cir. 2012)

5. Id.; see New Kids on the Block v. News Am. Publ’g, Inc., 971 F.2d 302, 306-07 (9th Cir. 1992).

6. Id. at 155-160.

7. Id. at 156.

8. Id.

9. See Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036, 1064 (9th Cir. 1999).

10. Brookfield Communs. v. W. Coast Entm't Corp., 174 F.3d 1036, 1063 (9th Cir. Cal. 1999); see also, J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition §23:6, 50-53 (4th ed. 2012) (stating most courts now recognize the initial interest confusion theory).

11. Brief for Appellee at 8, Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012) (No. 10-2007).

12. Lamparello v. Falwell, 420 F.3d 309, 316 (4th Cir. Va. 2005).

13. Id. at 158.

14. Id.

15. Id., citing J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition §23:8 (4th ed. 2012); See also, 15 USC §1125(a).

16. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 160 (4th Cir. 2012), quoting Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986).

17. Id.

18. Id.

19. J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition §7:63, 173 (4th ed. 2012).

20. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 162 (4th Cir. 2012) (The functionality doctrine does not apply to whether a mark makes defendant’s product more useful. The issue is whether the mark was functional as Plaintiff used it.)

21. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 163 (4th Cir. 2012), quoting Rosetta Stone Ltd. v. Google Inc., 730 F. Supp. 2d 531, 547 (E.D. Va. 2010).

22. Id. at 548, citing Tiffany (NJ) Inc. v. eBay, Inc., 600 F.3d 93 (2d Cir. 2010).

23. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 164 (4th Cir. 2012).

24. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 164 (4th Cir. 2012), quoting Rosetta Stone Ltd. v. Google Inc., 730 F. Supp. 2d 531, 549 (E.D. Va. 2010).

25. Id.

26. Rosetta Stone Ltd. v. Google Inc., 730 F. Supp. 2d 531, 550-52 (E.D. Va. 2010).

27. Id. at 168.

28. Id. at 169.

29. Id., citing 15 USC § 1125(c)(3)(A).

30. Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 170 (4th Cir. 2012).

31. 15 USC § 1125(c)(2)(B).

32. Id. at 171.

33. Id. at 171.

34. Id. at 171–73.

35. See Network Automation, Inc. v. Advanced Sys. Concepts, 638 F.3d 1137 (9th Cir. 2011).

36. See Rescuecom Corp. v. Google, Inc., 562 F.3d 123 (2d Cir. N.Y. 2009).

37. Mary LaFrance, “Unauthorized use of Trademarks in Keyword Advertising,” 2012 Emerging Issues 6487, June 27, 2012.

38. See “AdWords Trademark Policy,” (last visited June 28, 2013).

39. See “Advertising Policies Help: Trademark, Copyright and Counterfeit,” (last visited June 28, 2013).


Zola Suite
ABA Member Discounts and Offers
  • About GPSolo eReport

  • Subscriptions

  • More Information

  • Contact Us

GPSolo -- Your Success, Our Mission