Five days before trial in May 2010, the district court in Rosetta Stone Ltd. v. Google, Inc., 730 F.Supp.2d 531, 540 (E.D.Va. 2010), ruled that Rosetta Stone could not hold Google liable for auctioning Rosetta Stone’s trademarks as “keywords” – i.e., search terms which generate advertisements or “sponsored links” on Google’s website. Specifically, the district court found no triable issue regarding any likelihood of confusion arising from Google’s practices. Id. at 540-45. Earlier this year, the Fourth Circuit rejected the trial court’s reasoning, remanding the case for further proceedings. 676 F.3d 144, 152-60 (4th Cir. 2012). The appellate court’s reasoning will bolster foes of consumer confusion but may discourage advocates of free competition, who view Google’s practices as enabling competitors to jockey for prominent “shelf space” in the online marketplace.
Like the district court, the appellate court found that three factors governed determination of the likelihood of confusion: Google’s intent, evidence of actual confusion of users, and sophistication of the consuming public. Id. at 153-55.
Regarding intent to cause user confusion, the court found a triable issue in that Google had loosened its keywords policy in 2009 to permit trademarked terms in the title or body of a sponsored link despite internal studies from 2004 showing that “94% of users were confused at least once” by such usage. 676 F.3d at 155-56. However, critics may question the court’s reliance of a 2004 study to prove intent in 2009 given the pace of change in internet practices. Critics may also question whether Google’s intent regarding the sixteen other brand names studied properly evidences its intent regarding the trademarks “Rosetta Stone” and “language library” in particular. Id. at 158. In any event, Google may be wary going forward of promulgating studies opining on the risk of litigation without taking additional measures to garb the investigation in privilege.
Regarding evidence of actual user confusion, the district court had rejected Rosetta Stone’s testimony from five individual deponents as de minimis. Id. at 157-58. On appeal, the Fourth Circuit noted that five was the maximum number of “actual confusion” depositions permitted by the district court – a point with which it is difficult to argue. Id. at 158. Nonetheless, the question remains whether five instances of confusion out of 100,000 impressions over six years properly justifies trial. Id. at 157-58. While Rosetta Stone also offered evidence of 262 complaints by customers in one year regarding counterfeit software, no evidence existed of any purchases from sponsored links at Google. Id. at 158.
The appellate court further found a triable issue regarding actual confusion on grounds that two of Google’s own in-house trademark attorneys testifying under Rule 30(b)(6) could not determine which sponsored links were authorized resellers of Rosetta Stone products without additional research. Id. at 158-59. However, the decision does not reflect the information conveyed to Google’s deponents. Confusion arising from a bare screenshot of a results page may warrant less credit in ascertaining triable issues of fact than a broader array of information available to internet users, absent reliance on the fraught doctrine of “initial interest” confusion.
The appellate court further discerned evidence of actual confusion based upon the customer survey evidence of Rosetta Stone’s expert, which showed that 17 percent of consumers suffered actual confusion about the origin, sponsorship, or approval of the sponsored links generated on Google when queried using Rosetta Stone’s trademarked terms. Id. at 159. The district court dismissed this evidence as bearing upon the “non-issue” of Google’s endorsement of sponsored links. Id. The appellate court rejected this reasoning as fundamentally wrong-headed: given that trademark law protects against confusion not only as to source, but also as to affiliation, connection, or sponsorship, Rosetta Stone’s evidence properly gave rise to a triable issue. Id. While a confusion rate of 17 percent may not appear overwhelming, especially given that a rate below 10 percent would “clearly favor the defendant,” id., the Fourth Circuit’s opinion suggests the difficulty to defendants of obtaining summary judgment in trademark cases in the presence of unfavorable expert testimony. Id. at 153.
Regarding the final factor – sophistication of the consuming public – the evidence showed that Rosetta Stone charged $259 for a single-level package and $579 for a three-level bundle. Id. at 159-60. The district court found that these prices, together with the fact that consumers of Rosetta Stone products stood ready to invest the substantial time necessary to learn a foreign language using Rosetta Stone’s products, indicated a sophisticated consumer unlikely to be confused. Id. The Fourth Circuit disagreed, citing the confusion of Google’s own trademark attorneys, its internal study of confusion, and the deposition of five Rosetta Stone customers. Id. Ultimately, the appellate court repeated a criticism which became a motif in its opinion: the trial court improperly weighed evidence as a finder of fact instead of merely deciding whether sufficient evidence existed to proceed to trial. Id.
Further issues remain on remand, notably including whether Google uses Rosetta Stone’s marks “in commerce” as required for infringement – an issue which the Fourth Circuit explicitly held open before citing the Second Circuit’s rejection of Google’s arguments on this point in Rescuecom Corp. v. Google Inc., 562 F.3d 123, 129-31 (2d Cir. 2009). Then, of course, a six-day trial remains to be scheduled. Questions may eventually be asked as to whether keyword controversies present facts best analyzed under a direct infringement theory as opposed to an indirect infringement theory, given the arguments that counterfeiters and not search engines are properly the targets of any direct infringement suits. Ultimately, like the artifact giving rise to this plaintiff’s name, Rosetta Stone may take time to decipher, and may prove to have far-reaching impact.