Gersen & Hemphill: What's in a Bottle?

What's in a bottle?  If the bottle in question is shown below, the answer is protected trade dress. The Coca-Cola Company presently maintains several active registrations for many of its bottle shapes. See, e.g., Trademark Reg. No.'s 696147, 4242307, 4200433.

These registrations serve as "prima facie evidence" of, among other things, The Coca-Cola Company's "exclusive right to use the registered [trade dress] in commerce on or in connection with the goods or services specified in the registration[,]" subject to defenses such as abandonment, descriptive fair use, and functionality. Other sellers that use a confusingly similar bottle shape without authorization from The Coca-Cola Company could be sued under the Lanham Act, § 32 or § 43, even if they don't use the well-known Coca-Cola names or color schemes. That is what trade dress protection can accomplish.

In their fascinating new essay, "The Coca-Cola Bottle: A Fragile Vessel for Building a Brand," which is forthcoming in Legal Applications of Marketing Theory (Gersen & Steckel, eds), Jacob Gersen and Scott Hemphill tell the history of how The Coca-Cola Company (hereafter "Coke") secured lock-tight trade dress protection in the distinctive shapes of its bottles. The authors also previously authored a book chapter on the Coke bottle for A History of Intellectual Property in 50 Objects, edited by Claudy Op den Kamp and Dan Hunter.  

Digging into historical documents — including settlement agreements resulting from litigation that Coke brought against other sellers — the authors argue that the company's trajectory towards trade dress protection for the famous Coca-Cola bottle was not as smooth as we might think. In the authors' words, it was "fragile and contingent." (2). They reveal many reasons things might have come out differently for Coke.  

I won't summarize the whole essay here, because it's short and very fun to read. But I want to draw out what I see as one of the authors' most important insights. They show that Coke secured common law trade dress protection by initially using design patent law. They argue that, under contemporary legal principles, courts might have rejected this strategy and not permitted Coke to use unfair competition law to protect its bottle shapes following the expiration of Coke's design patents. This is, after all, what the Supreme Court did in 1938 when it did not let Nabisco stop other sellers from using the "pillow shape" of Nabisco's shredded wheat cereal following expiration of Nabisco's patents.  Things did not come out that way. Coke's strategy was a winner. But the authors suggest things could have gone very differently. 

Back in the 1880s, as the authors tell it, Coke deliberately designed a bottle shape with the hope that it would be distinctive and legally protectable against imitators. The idea was to design a bottle that was “so shaped that, even if broken, a person could tell at a glance what it was.” (4, quoting a contemporary source).  Coke's goal, the authors suggest, was to obtain exclusivity by any means necessary. It was not just to prevent other sellers from using similar names (e.g. Koke). For that, Coke could have relied on traditional word mark protection for "Coca-Cola" —though the authors describe how even this proved challenging given what was (or rather what was not) in the drink. (15-16).

Coke wanted more than words. It wanted to protect the bottle shape itself and it wanted to do so for as long as possible. According to the authors, this was a very deliberate strategy. Coke initially ran a $500 contest to settle on a unique bottle shape. Coke then secured three design patents for different versions of the Coke bottle shape and used those to sue other sellers like Whistle, which was using a somewhat similar bottle shape but selling "brightly colored soda" (8) instead of the dark colored soda Coke is known for. (For images of the Whistle's own design patent and comparison to Coca-Cola's, see the author's essay.). 

The Coke design patents, the authors write, were "granted in 1915, 1923, and 1937" and "collectively spanned nearly 36 years—more than double the 14-year duration of protection for a single patent." (7). After the design patents expired, Coke was left with trade dress protection under unfair competition law, and eventually federal trade dress protection under Lanham Act of 1946. It still has this trade dress protection today, and will continue to have it subject to abandonment. 

Under modern doctrine, trade dress is available for a bottle shape like Coke's.  However, there are two major caveats, either of which could have defeated Coke early on: distinctiveness and functionality.

Product Design, Secondary Meaning, and Design Patents

First, if the feature is classified as product design, which Coke's bottle probably is, there's a higher threshold for distinctiveness.   If a feature is deemed to be product design —part of what makes the product "useful or appealing" to consumers— then the trade dress owner needs to prove secondary meaning—i.e. that consumers associate the feature, in reality, with a single source. Justice Scalia specifically gave the glass Coca-Cola bottle as an example of a "hard case at the margin" where courts and trademark examiners might have trouble drawing the line between packaging and design, because the shape of the bottle is part of the product for at least some consumers. 

There will indeed be some hard cases at the margin: a classic glass Coca-Cola bottle, for instance, may constitute packaging for those consumers who drink the Coke and then discard the bottle, but may constitute the product itself for those consumers who are bottle collectors, or part of the product itself for those consumers who buy Coke in the classic glass bottle, rather than a can, because they think it more stylish to drink from the former. 

In the same breath Justice Scalia suggested that, to the extent this was a "close call," then the court should err on the side of classifying this as product design and require a showing of secondary meaning.  Obviously today the Coke shape would survive and remain protectable trade dress because it does have secondary meaning among consumers of soft-drinks. The Coke bottle is a shape that consumers recognize and associate with a single (and very famous) source.  

But how did Coke get this secondary meaning?  At the time, the authors show, there were other sellers, like Whistle, using similar bottle shapes. Coke had to distinguish itself from those sellers in consumers' minds. The easiest way to do this is to be the only one, and, to extent you are not the only one, place legal barriers in others' way. This is what Coke did. Specifically, the authors assert that Coke strategically used "bootstrapping" — which they define as using one type of claim "to compensate for a weakness or other deficiency in another"—to combine the protections of design patents and trade dress. The design patents Coke had secured gave it several decades' worth of protected use during which Coke could sue others using similar bottle designs for infringement. Coke wasn't always successful, but it did sue, and perhaps scared away, others.  Then, after the design patents' expiration, Coca Cola was left with trade dress protected through enormous amounts of secondary meaning that it had built up during its periods of design patent protection. 

In other words, Coca Cola used "de facto" secondary meaning (exclusivity thanks to patent rights) to gain "de jure" (legal) secondary meaning in the end based on real-world consumer impressions.

The authors do not explicitly condone or condemn this type of "bootstrapping"—though, as discussed below, they note the views of courts, especially early on, that this could disrupt the patent bargain. The basic concern is that, if a seller gets its 14 (or now 15) years under a design patent, why should it also get trade dress protection too? The whole point of the patent system is ostensibly to calibrate incentives to reward creators while preserving the rights of the public after patents expire. Tacking on trade dress to the expired patent seems to go against all of that.  

To play Devil's Advocate, one could argue there's not necessarily anything wrong with a seller garnering secondary meaning for a feature (e.g. the Coke bottle shape) by using a design patent, and then using that secondary meaning to protect, specifically, those marketing features that consumers come to see as identifying the seller and distinguishing them from others.  It's at least theoretically possible to separate out the policy functions of the design patent and the trade dress, and to posit that design patents help fill a gap in trade dress law that protects new companies that haven't yet developed secondary meaning. 

It's easy to say now that Coke "doesn't need trade dress in the bottle because it has a famous name and a dominant position in the soft drink market." But what about smaller businesses or other sellers who aren't as dominant?  The principle of "secondary meaning in the making" is controversial in trademark law. Sellers aren't typically supposed to gain protection for descriptive marks until they develop secondary meaning. But it's not unheard-of for trademark law to recognize the interests of newcomers to a market and give them breathing space to develop their reputations. Examples include: restaurant decor/non-product-design trade dress like in Two Pesos, where the Supreme Court held this type of trade dress can be protected as "inherently distinctive" prior to developing secondary meaning; similar cases like Fuddruckers where the Ninth Circuit suggested Fuddruckers might be able protect its trade dress in an area where it did not yet have proof of secondary meaning if there was intentional copying; and, more generally, the "natural zone of expansion" doctrine which allows unregistered users to protect marks in places where they don't yet operate but are likely to expand.  

Functionality Bar and Design Patents

The story for functionality is still more interesting. Even trade dress that is legally distinctive is vulnerable to the functionality bar.  In the wake of Two Pesos, which permitted trade dress protection for the appearance of a Mexican restaurant that lacked secondary meaning at the time, Congress amended the Lanham Act to specifically provide that in a case involving unregistered trade dress, the burden is on the person asserting protection to prove non-functionality. See 15 U.S.C. § 1125(a)(3). And even for registered and incontestable trade dress, functionality remains as a defense.

In 2001, the Supreme Court held in TrafFix that a feature is functional in a utilitarian sense if it's "essential" to the use or purpose of the article, or if it affects cost or quality.  Importantly, the Supreme Court indicated that an (active or expired) utility patent  is "strong evidence" of utilitarian functionality.  But the Court did not indicate that a design patent, versus a utility patent, would serve as strong evidence of functionality. So the bottle design is in a good place to survive a functionality challenge, irrespective of the fact that it was covered by design patents long ago. 

However, the authors suggest things could have come out very differently.  They characterize early case law as viewing patents as a "bargain between the patentee and the public" and enforcing a somewhat-strict rule that "after expiration, the patented invention or design is dedicated to the public." (17).  The authors point out that in the famous "Shredded Wheat" case, the Supreme Court had held that Nabisco could not get trade dress for the previously-patented "pillow shape" of Nabisco Shredded Wheat, and that the Court had observed that Nabisco previously had a design patent on the "the pillow-shaped form," not just a utility patent on the machine for making them. Justice Brandeis wrote, 

The plaintiff has not the exclusive right to sell shredded wheat in the form of a pillow-shaped biscuit — the form in which the article became known to the public. That is the form in which shredded wheat was made under the basic patent. The patented machines used were designed to produce only the pillow-shaped biscuits. And a design patent was taken out to cover the pillow-shaped form. Hence, upon expiration of the patents  the form, as well as the name, was dedicated to the public.

Kellogg Co. v. National Biscuit Co., 305 US 111, 119-120 (1938). (The authors observe, too, that a court in a similar a prior case, written by Judge Hand on the Second Circuit, had also mentioned that the shredded wheat pillow shape was the subject of a now-expired design patent. The citation is Shredded Wheat Co. v. Humphrey Cornell Co., 250 F. 960 (2d. Cir. 1918)).

These cases might have suggested that Coke should be "unable to assert an unfair competition claim that blocked others’ use of the bottle once the design patents had expired[,]" because "[t]he logic of the bargain theory[,]" under which the patented subject matter is dedicated to the public once the patent expires, "applied with equal force to utility and design patents[.]" (19). 

But Coca-Cola did not fall to this challenge and instead was able to rest on its design patents; generate breathing room to develop secondary meaning; and then capture indefinite trade dress protection thereafter. The reason, the authors suggest, is that the law took a different trajectory, favoring the functionality concern alone. TrafFix for instance looks to utility patents as an indicator of functionality, not necessarily as an indicator of a patent owner impermissibly extending protection after expiration.  They argue this distinction between design and utility patents makes sense, if functionality is the issue. (19) ("A concern with functionality provides a basis for distinguishing design patents from utility patents. Utility patents protect functional features, whereas design patents protect ornamental features."). But it would not make sense under the older ideas about a bargain between patentee and public or even based on a loose "no-bootstrapping" rule.   

One comment on this point. I don't necessarily disagree with the authors' assessment of modern functionality. They state that "[f]unctionality, and a version of functionality confined to utility patents, is the principal doctrine that mediates the interaction between patent and trade dress. Today, an expired design patent is generally held to pose no barrier to continued trade dress protection or injunctive relief." (20).  This is true in the sense that an expired design patent doesn't generate "strong evidence" of utilitarian functionality. As mentioned above, in TrafFix, the Court specifically held that "[a] utility patent is strong evidence that the features therein claimed are functional..." but did not mention design patents.  Of course, this doesn't mean the previously design-patented feature might not still be found functional, as, for instance, the Ninth Circuit found certain features of Apple's iPhone were. The presence of a patent alone isn't dispositive in either direction.  (Notably, the Supreme Court in TrafFix left open the possibility of protection, even for formerly-utility-patented features, where the claimed feature "is merely an ornamental, incidental, or arbitrary aspect of the device.").  

Another complication to this thesis is so-called aesthetic functionality.  This is a doctrine that has split off from utilitarian functionality since those early cases were decided. Aesthetic functionality (which should really have a different name) limits protection for features that, even if they don't strictly "do" anything, are too pleasing to consumers to place in the hands of a single seller. Examples: the shape of peanut butter cups, heart-shaped boxes, shoulder pads...maybe even the particular shape of a soda bottle. Today, courts will hold a feature is aesthetically functional if providing exclusivity would place other sellers of the same product at a "significant non-reputation-related disadvantage." 

The authors, perhaps wisely, decided to steer clear of aesthetic functionality in their essay. But the doctrine alters the calculus for trade dress protection at a practical level. Even if it were true that utilitarian functionality disregarded features previously protected by design patents, the aesthetic functionality bar might still apply to those features and at least limit the scope of protection to preserve market competition. Here, the Coke bottle design would still be subject to aesthetic functionality. True, the standard is laxer because, if alternative designs are available, this can be considered and weigh in favor of granting protection. The Coke bottle designs still would survive with a narrow sliver of protection, so long as there's a range of adequate alternative designs leftover.  The key word though is narrow. The scope of protection would be limited to the bottle shape for which Coke has secondary meaning and in which granting protection would not "significantly" limit competition in the market for colas.  

To be sure, the fact that aesthetic functionality exists as an additional barrier to over-broad trade dress protection does not contravene the authors' point about the functionality approach winning out over a bargain theory approach. If bargain theory were the basis for these doctrines, then there might be no trade dress protection at all for trade dress features like the Coke bottle shape that were initially protected under design or utility patents. Still, aesthetic functionality's existence does complicate the authors' assertion that "[f]unctionality, and a version of functionality confined to utility patents, is the principal doctrine that mediates the interaction between patent and trade dress. (20)."  While it's true that "[t]oday, an expired design patent is generally held to pose no barrier to continued trade dress protection or injunctive relief[,]" so-called aesthetic functionality is still there as a back-stop that at least limits the scope of trade dress protection even for features that survive utilitarian functionality. 

***

Long story short, read the essay. It's really good!    

For some other recent papers looking at the design patent/trade dress interface, see, e.g., McKenna & Silbey, Investigating Design, Fromer & McKenna, Claiming Design, McKenna, Fixing Functionality. For argument that design patents fill a gap in trade dress protection and serve as an "alternative rule of evidence" in trade dress protection, see Crouch, A Trademark Justification For Design Patents. On design patents, generally, see, e.g., Burstein, various papers.