Multi Time Machine, Inc. v. Amazon.com, Inc., No. 13-55575 (9th Cir. Oct. 21, 2015)
Reversing itself (with amicus advocacy from, among others, yours truly), the panel now by 2-1 holds that Amazon should have gotten summary judgment for its practice of responding to searches for “MTM Special Ops” with other, clearly labeled watches, without explicitly stating that none of its results are for that exact product. Because of the clear labeling of the results, “no reasonably prudent consumer accustomed to shopping online would likely be confused as to the source of the products”:
To whatever extent the Sleekcraft factors apply in a case such as this – a merchant responding to a request for a particular brand it does not sell by offering other brands clearly identified as such – the undisputed evidence shows that confusion on the part of the inquiring buyer is not at all likely. Not only are the other brands clearly labeled and accompanied by photographs, there is no evidence of actual confusion by anyone.
Sleekcraft isn’t a rote checklist, and different factors may be important in different circumstances. In search engine cases, an additional factor is particularly important, per Network Automation: “the labeling and appearance of the advertisements and the surrounding context on the screen displaying the results page.” By contrast, the multifactor test “is not particularly apt,” because it was developed for a different problem, “i.e., for analyzing whether two competing brands’ marks are sufficiently similar to cause consumer confusion.” [Comment: I’ve rarely seen a better admission that IIC is a bad idea. That “different” problem is called “whether there is trademark infringement.”]
But, the majority notes in a footnote, a panel can’t get rid of IIC. So, let’s go to the argument that’s not about real infringement: MTM’s argument isn’t that Luminox and other competitors’ brands have confusingly similar marks, but rather that Amazon’s search page creates IIC, because the search results page displays the search term used – here, “mtm special ops” – followed by a display of numerous watches manufactured by competitors, without explicitly informing the customer that Amazon does not carry MTM watches. The alleged confusion isn’t caused by competitors, but rather by the web page’s design.
Still, setting aside the multifactor confusion test, ultimately the question is whether a reasonably prudent consumer is likely to be confused. Here, evaluating the web page plus the relevant consumer was enough to answer that question as a matter of law. Because “the default degree of consumer care is becoming more heightened as the novelty of the Internet evaporates and online commerce becomes commonplace,” and because the watches at issue sell for several hundred dollars, reasonable consumers would be careful.
As for the webpage, even MTM agreed that clear labeling could entitle a defendant to summary judgment. Here, the products themselves were clearly labeled to avoid any likelihood of IIC on the part of a reasonably prudent consumer. There were even photos!
MTM argued that the use of the search term “mtm special ops” three times at the top of the search page might lead consumers to think that the products displayed were types of MTM watches. But that’s silly. I mean, “highly unlikely.” None of the watches use “MTM” or “Special Ops,” and some of the products listed aren’t even watches, like Jerry Ahem’s book Survive!: The Disaster, Crisis and Emergency Handbook. “No reasonably prudent consumer, accustomed to shopping online or not, would assume that a book entitled ‘The Moses Expedition’ is a type of MTM watch or is in any way affiliated with MTM watches.” The same was true of the other results. The standard, after all, is likelihood of confusion, not possibility.
MTM argued that Amazon needed to explain to consumers that it doesn’t sell MTM watches before offering alternatives. But anyone who can read English could tell that Amazon sold only the listed products. The results were “unambiguous – not unlike when someone walks into a diner, asks for a Coke, and is told ‘No Coke. Pepsi.’”
MTM argued that factual disputes precluded summary judgment, but none were material. “The likelihood of confusion is often a question of fact, but not always.” The Ninth Circuit has reached similar results in false advertising cases where no reasonable consumer could have been fooled. Anyway, consideration of the remaining Sleekcraft factors wouldn’t change anything. There was no evidence of actual confusion. As for intent, the design of the webpage “indisputably produces results that are clearly labeled as to the type of product and brand,” so that too didn’t favor MTM. As for strength, “[e]ven assuming MTM’s mark is one of the strongest in the world – on the same level as Apple, Coke, Disney, or McDonald’s – there is still no likelihood of confusion because Amazon clearly labels the source of the products it offers for sale.” Furthermore, Network Automation already said that the other factors were unimportant in an internet search case involving clear labeling and a high degree of consumer care.
Judge Bea, dissenting, would have sent the case to a jury on the hypothesized IIC about affiliation between MTM and Luminox et al. (Would consumers think that MTM had formed relationships with all the brands listed?) The dissent pointed out that other internet sites that didn’t sell MTM watches would give a “no results” result to the “mtm special ops” search, and seemed to think that was all Amazon should be doing too. Maybe MTM wasn’t right that consumers might think it sold watch parts to Luminox as a result of Amazon’s search functions, but a jury should decide that.
The majority, the dissent accused, implicitly added “at point of sale” to the “likely confusion” made actionable by the Lanham Act, and ignored “the possibility that a reasonably prudent consumer might initially assume that those brands enjoyed some affiliation with MTM which, in turn, could cause such a shopper to investigate brands which otherwise would not have been of interest to her.” [Note the continued use of “possibility”—yet to avoid summary judgment MTM needed to offer evidence that a factfinder could use to conclude that this was likely, even if you buy the doctrine of initial interest affiliation confusion.] [Also, how would “no results for MTM special ops found, but try these watches instead” dispel the possibility of a suspicion that MTM had sold parts to or had some other affiliation with the brands listed? As our amicus pointed out, “no results for MTM special ops, but what about MTM Black Cobra?” would be a perfectly sensible sentence. The “no results” doesn’t address the postulated type of confusion. I suppose Amazon could program in “no results for MTM special ops, but try these other completely unrelated, competing products with no affiliation with or parts-selling relationship to MTM,” but that seems even sillier.]
Network Automation was distinguishable because (1) it reversed a grant of a preliminary injunction, and (2) “the ‘diversionary’ goods were clearly labeled on the response page as ‘Sponsored Links,’ showing that the producers of those products were the ones advertising for themselves, not for the firm named in the search request.” This second reason is pretty bold, since the empirical evidence is that consumers don’t know very much about the division between paid and organic results despite labels of that type, and also since firms often bid on their own trademarks anyway.
“On this record, a jury could infer that users who are confused by the search results are confused as to why MTM products are not listed.” [Again: not trademark confusion!] And those confused users might “wonder” whether a competitor acquired MTM or was otherwise affiliated with or approved by it. [Also, this would be true even if MTM watches appeared in the results.] The risk of affiliation was especially great for a luxury brand, since “many luxury brands with distinct marks are produced by manufacturers of lower-priced, better- known brands—just as Honda manufactures Acura automobiles …, and Timex manufactures watches for luxury fashion houses Versace and Salvatore Ferragamo.” As for the Coke/Pepsi comparison, “[n]o shopper would think that Pepsi was simply a higher end version of Coke, or that Pepsi had acquired Coke’s secret recipe and started selling it under the Pepsi mark. [Why wouldn’t a jury have to decide that under the dissent’s theory? Also, what makes the second possibility even relevant to trademark infringement?]
Anyway, the real issue isn’t source confusion, it’s whether Amazon’s tactics caused IIC “by attracting potential customers’ attention to buy the infringing goods because of the trademark holder’s hard-won reputation.” [Hey, wait, what infringing goods? At what point did Luminox watches start to infringe MTM’s rights?] A jury could find IIC, not just mere diversion.
And here the dissent demonstrates one of the ways in which the multifactor confusion test has become more harmful than helpful: it contends that the majority erred in considering mark strength unimportant, because “[a] mark’s strength is a measure of how uniquely identified it is with a product or service, and therefore how deserving of trademark protection.” But one quite persuasive point of Network Automation, not to mention New Kids and Rogers v. Grimaldi and parody cases like Chewy Vuiton, is that sometimes mark strength doesn’t increase the likelihood of confusion, because of the way that a strong mark is being used. If we’re at all interested in being accurate about likely confusion, even if we’re committed to using circumstantial tests, then I don’t see why we should use a blanket rule that the stronger a mark is, the more likely confusion is no matter what the other facts are.
The dissent also was willing to treat Amazon’s intent as intent to confuse, because “MTM submitted evidence that Amazon vendors and customers had complained to Amazon because they did not understand why they received certain non-responsive search results when they searched for products that are not carried by Amazon.” Amazon didn’t address the complaints by explaining how its search function worked (it also accounted for related searches and other aspects of consumer behavior). Amazon did nothing to alleviate confusion, which provided “some evidence of an intent to confuse.” [Again, it would be nice if everyone could remember that “confusion” is not “confusion in the air” but trademark confusion. Intriguingly, the dissent doesn’t notice that these very complaints tend to disprove its initial interest affiliation confusion theory: customers understood that they were getting unrelated search results, rather than thinking that some hidden marketplace consolidation had taken place.]