Amazon Doesn’t “Sell” Its Marketplace Goods–Milo & Gabby v. Amazon
Milo & Gabby is a small family business that designs and sells “animal-shaped” pillowcases. It discovered that knockoffs were listed for sale on Amazon’s website. The products were actually offered for sale by third party sellers, and all but one did their own fulfillment. One of the merchants used Amazon’s “Fufillment by Amazon” service which meant that the seller took advantage of Amazon’s logistics network. A seller who used this service ships its product to an Amazon fulfillment center where Amazon deals with storage, shipping, etc.
The record does not reflect whether M&G sent takedown requests prior to filing suit. It asserted trademark, copyright, and design patent claims. Upon receiving the lawsuit, Amazon suspended the sellers and prevented one of the rogue sellers from re-listing the product.
The district court granted summary judgment as to M&G’s copyright and trademark claims. It rejected M&G’s theory of direct infringement because of the lack of evidence that Amazon “actively reviewed, edited, altered or copied” M&G’s images. It rejected the trademark claims because there was “no evidence of any violation of a valid, enforceable mark entitled to protection under the Lanham Act.”
The remaining claim for design patent infringement went to trial. The parties agreed the key question is whether Amazon “offered to sell” the items in question and that this was a question of law. An advisory jury found in favor of Amazon, and the district court entered a verdict in its favor. M&G appealed.
Design Patent claims: On appeal, the Federal Circuit found that M&G advanced the theory on appeal that Amazon “sold” the product rather than its “offer to sell” theory it relied on below. M&G abandoned the former theory at summary judgment. Because it failed to meaningfully argue the “offer to sell” theory for patent infringement, the court affirmed judgment against M&G on the patent claim. The court notes that M&G did not raise alternate theories of infringement, such as inducement or contributory infringement.
Copyright claims: M&G offered two possible theories of copyright liability for Section 106 “distribution”: (1) Amazon was a ‘seller,’ and (2) it was a distributor.
The seller liability theory turned on what it means to “sell” an infringing product for copyright purposes. The court says usually that a transfer of title dictates who is the seller, but there are some exceptions. M&G argued that several Uniform Commercial Code provisions (dealing with consignment sales and the right to return certain goods bought primarily for resale) demonstrate that someone can be a seller even without taking title. The court says these exceptions are not applicable. While there may be some circumstances where someone can be considered a seller when not transferring title, M&G failed to make the case that those exceptions apply here.
The distribution right covers infringement by “sale or other transfer of ownership.” Since M&G don’t satisfy the definition of “sale”, The infringement by distribution argument depends on M&G showing that Amazon infringed on the distribution right by “another transfer of ownership.” The court says that M&G failed to devote meaningful attention to this argument and essentially waived it. In passing, the court says that, whatever the phrase “other transfer of ownership” means, it too would require some transfer of title.
Lanham Act claims: M&G did not plead a palming off claim in its complaint (it asserted a claim for false designation of origin). In response to Amazon’s motion for summary judgment, M&G raised for the first time a palming off theory. The district court rejected this argument, and the appeals court does the same.
This is a big win for Amazon since the case exposed Amazon to liability for sales by its marketplace vendors. (eBay weighed in as amicus on its side.) Any ruling that gives Amazon a clean bill of health for infringement that occurs on its platform is useful. It’s unclear whether this ruling is fact-specific or will have broader effects. I tend to think the latter.
The part of the case that is puzzling is that there is no doubt Amazon’s fulfillment service helps parties sell the products in question. In this respect, Amazon is no different from other third parties (some of whom are even more removed, such as flea market vendors) who can be found liable. Indeed, the court says Amazon’s services “made it easier for third parties to consummate a sale.” This theory does not warrant attention from the court because M&G did not argue derivative liability. The court’s recitation of facts makes it seem like that would have been a more viable argument.
The decision does not mention the DMCA at all. Amazon’s off-line activities would not be covered by the DMCA, but their reproduction of M&G advertisements would be eligible for protection under the DMCA (see, e.g., Corbis v. Amazon). Perhaps the court did not address this theory due to the narrow grounds on which M&G premised infringement? It is still noteworthy to see a copyright ruling involving Amazon not mention the DMCA.
The court discusses what constitutes a sale in some detail, and covers a breadth of issues, including UCC provisions. The opinion also briefly alludes to first sale doctrine, including the case involving promo CD’s (Augusto, blogged here).
Finally, while not a big part of the ruling, the court refuses to consider the district court’s award of fees for the Lanham Act claim. Apparently, the trial court sent a clear signal that this was a bogus claim, but the plaintiff continued to pursue it. The court hammered down and awarded fees.
It’s painful to see how many times the court says that the plaintiff waived an argument or otherwise made unforced litigation errors. For example, because the plaintiff changed midstream its theory of design patent infringement, we don’t really get guidance on the application of design patent law to online marketplace sales. That leaves holes in this ruling that could be explored in future cases.
Still, I agree with Venkat that this is a big ruling for Amazon. This court says that Amazon didn’t make a copyright “sale,” and it seems to indicate that the same analysis would apply to patent law’s references to “sale.” This further ameliorates the district court’s troubling initial ruling on design patents, which was already blunted in part by the jury ruling.
The opinion is unfortunately marked “nonprecedential.” Still, this ruling reminds me a little of the Cablevision Second Circuit ruling. That ruling only addressed one of several litigation theories over DVR-as-a-service, yet it has evolved into the de facto legal standard. I expect that, over time, this case will stand for the proposition that Amazon doesn’t “sell” marketplace goods, even though the court only addressed pieces of that question.
The Federal Circuit’s holding that the person making a “sale” (at least for copyright purposes) usually is the person transferring title is another example of the Federal Circuit’s preferences for rules over standards. It’s a logical conclusion, however, and ideally other courts won’t lard it up with too many exceptions that can swallow the rule. If the “seller = transferor of title” syllogism holds, it deemphasizes plaintiffs’ various possible arguments over the other marketplace services that marketplaces provide to vendors and buyers, such as advertising services, payment services, trust-and-safety/reputation services–and even storage and fulfillment services. So long as the title passes directly from vendor to buyer, these ancillary services provided by marketplaces shouldn’t turn the marketplace operator into a “seller” for IP purposes. This reinforces the legitimacy of marketplace operators providing these ancillary services and perhaps encourages them to provide even more ancillary services.
While that may be good news, I remain troubled by the overlay of the Airbnb v. SF ruling, which opened up the door for regulators to bypass Section 230 by regulating “booking services.” Even if marketplace operators can sidestep some IP liability by not being “sellers,” regulators might still hinge liability on the marketplace’s closing of marketplace deals. See also the McDonald v. LG case, exposing Amazon to potential product liability claims for marketplace sales.
While this case was on appeal, eBay won a highly relevant (but uncited) ruling in Blazer v. eBay. That case held that, for utility patent purposes, vendors, not eBay, made the “offers for sale.” The court also rejected inducement and contributory patent infringement claims against eBay. It would have been nice if the Federal Circuit ratified or even acknowledged the Blazer opinion, but instead we get two different courts independently reaching complementary–and defense-favorable–rulings.
Case citation: Milo & Gabby LLC v. Amazon.com, Inc., 2016-1290 (Fed. Cir. May 23, 2017)