Ten years ago - 2014 - the Supreme Court decided Alice Corp v. CLS Bank, holding that - yes indeed - the expansive language of Mayo v. Prometheus (2012) applies equally to software and technology patents. A few weeks later, the USPTO began a dramatic transformation - pulling back notices of allowance and issuing thousands of supplemental office actions. The Federal Circuit's May 2024 en banc decision in LKQ v. GM is perhaps as dramatic a change for the design patent arena as Alice was for utility patents. The old Rosen-Durling test made it almost impossible to reject a design patent as obvious except for extreme cases involving either direct copying or extremely broad claims. The key difficulty was that precedent required the obviousness inquiry to begin with a single prior art reference that is "basically the same" as the claimed design - a roughly 1-to-1 relationship. Further, any secondary references had to be ‘so related’ to the primary reference that features in one would suggest application of those features to the other." In LKQ, the court found those requirements "improperly rigid” under principles of KSR which require a flexible obviousness inquiry. The overall effect is to make it easier to find a design patent obvious.
On May 22, 2024, the day after the Federal Circuit's en banc LKQ v. GM decision, the USPTO issued a memorandum to its examiners providing updated guidance and examination instructions in light of the court's overturning of the long-standing Rosen-Durling test for determining obviousness of design patents. The memo, signed by USPTO Director Kathi Vidal, aims to immediately align USPTO practices with the more flexible approach outlined by the Federal Circuit, which eliminated the rigid requirements that: (1) a primary reference be "basically the same" as the claimed design, and (2) secondary references be "so related" to the primary reference that features in one would suggest application to the other. This is a major shift in examination practice for design patents so it will be important to watch the developments to see whether the office ramps up design examination.
In a highly anticipated en banc decision, the Federal Circuit has overruled the longstanding Rosen-Durling test for assessing obviousness of design patents. LKQ Corp. v. GM Global Tech. Operations LLC, No. 21-2348, slip op. at 15 (Fed. Cir. May 21, 2024) (en banc). The court held that the two-part test's requirements that 1) the primary reference must be "basically the same" as the claimed design, and 2) any secondary references must be "so related" to the primary reference that features from one would suggest application to the other, "impose[] limitations absent from § 103's broad and flexible standard" and are "inconsistent with Supreme Court precedent" of both KSR (2007) and Whitman Saddle (1893). Rejecting the argument that KSR did not implicate design patent obviousness, the court reasoned that 35 U.S.C. § 103 "applies to all types of patents" and the text does not "differentiate" between design and utility patents. Therefore, the same obviousness principles should govern. This decision will generally make design patents harder
Seirus has petitioned for writ of certiorari in its long-running design patent dispute with Columbia Sportswear. The petition asks two questions related to the comparison process for design patent infringement -- in particular, the questions focus on what can qualify as "comparison prior art" used to provide context for the infringement analysis.
The Schedule A litigation phenomenon continues apace in the Northern District of Illinois, a court that has become, in the words of Judge Seeger, “an assembly line for TROs.” But Schedule A litigation is not confined to Chicago. It has spread, perhaps most notably to the Southern District of Florida and the Southern District of New York.
One recent decision out of New York merits closer attention. In this case, as in most Schedule A cases, the plaintiff was able to obtain an ex parte TRO that included an order instructing Amazon to freeze the defendants’ seller accounts. The order also required the plaintiff to post a bond of $20,000 “for the payment of any damages any person may be entitled to recover as a result of an improper or wrongful restraint ordered.”
The plaintiff sued 163 defendants, alleging that each was liable for infringing a utility patent directed towards “a rectangular-shaped buckle-and-belt mechanism” for “an outdoor exercise product.”
At the TRO stage, the plaintiff’s “infringement evidence chart” consisted of a series of screenshots (many of them low-resolution screenshots) of the accused products. Here’s an example of all of the evidence submitted in that chart with respect to one of the defendants:
As per usual in a Schedule A case, the defendants did not find out about the case until after their accounts were frozen. When the plaintiff moved to extend the TRO, multiple defendants appeared to object. Judge Rochon refused to extend the TRO and the plaintiff voluntarily dismissed the case.
Two of the defendants, Hyponix and NinjaSafe, moved for bond damages, sanctions, and fees. They argued that the plaintiff had filed to conduct a sufficient pre-suit investigation and had committed various acts of litigation misconduct. They further argued that they each suffered damages from being wrongfully enjoined.
Judge Rochon granted the defendants’ motions for bond damages but denied their motions for sanctions and fees. She agreed that the moving defendants had been wrongfully enjoined because “Hyponix has pointed to at least four elements of claim 1 of the ’673 patent that are not present in its product” and “Ninja Safe has also shown that its products may not infringe claim 1 and has raised questions of invalidity.” Accordingly, Judge Rochon granted each moving defendant bond damages, though less than they asked for: $3,682.28 for Hyponix and $14,641.51 for NinjaSafe.
But she refused to grant sanctions or fees, despite being “troubled by Plaintiff’s conduct in this case.” In particular, Judge Rochon noted the “‘clear discrepancies’ between the protected elements of the ’673 Patent and the products of many of the parties against which Plaintiff secured a TRO.” She also noted other instances of “possible misconduct,” including:
“Plaintiff represented that most of the 163 parties were difficult to find and contact. In practice, however, contact information for many of the parties was readily available. . . . Plaintiff does not indicate that it tried with any diligence to locate these parties before seeking a TRO.”
“The pace and prevalence of Plaintiff’s dismissals suggest to the Court that Plaintiff used Rule 41 as part of a broader strategy to freeze the accounts of its competitors, then withdraw its claim against any party that happened to object.”
“Plaintiff failed to provide Hyponix with documents necessary for its defense. . . . . Plaintiff claims, falsely, that Hyponix did not request these documents.”
(Emphasis added.) Despite all of this, Judge Rochon refused to sanction the plaintiff:
Despite these concerns, the Court does not lightly award sanctions and will not do so in this case. Plaintiff holds a valid patent for its Hanging Exercise Product, its claim was colorable against at least some of the parties, and it dismissed its lawsuit voluntarily at a very early stage in the litigation (presumably in light of the issues raised by the Court at the order to show cause hearing), before any of the defendants responded to the Amended Complaint. Defendants here were made whole for their losses under the bond. Although a close question, the Court exercises its discretion to deny Defendants’ request for sanctions under its inherent powers and 28 U.S.C. § 1927. . . . To the extent that Plaintiff and its counsel engage in similar misconduct in the future, however, the Court will not hesitate to impose sanctions.
As Professor Eric Goldman noted in this blog post, “it would not be lightly awarding sanction when a plaintiff has committed so many violations.” Two additional points stand out as well.
First, the fact that some of the infringement claims might be colorable does not change the fact that the plaintiff brought numerous claims that were not—including the claims brought against the moving defendants. If the plaintiff had sued Hyponix and NinjaSafe separately, would that have changed the court’s analysis? If so, why should the fact of mass joinder insulate the plaintiff from sanctions? Especially in light of the fact that it’s far from clear that any—let alone all—of the defendants were properly joined, as they sell different products and do not seem to be actually connected in any way. See 35 U.S.C. § 299. In any case, the fact remains that this plaintiff brought many claims that were not colorable and used the machinery of the federal judiciary to wrongfully enjoin competitors. That is what should matter in the sanctions calculus, not the fact that some of the other claims (against apparently unrelated defendants) might have potentially had merit.
Second, it is true that the plaintiff dismissed the case at what would be, in a regular case, “a very early stage in the litigation.” But in a Schedule A case, the TRO seems to be the whole game. The plaintiff gets a TRO with an asset freeze, then starts making settlement demands. At that point, the defendants generally either settle or default. It appears that these cases aren’t meant to proceed any further. And as the defendants’ submissions show, significant damage can be done in these cases, even in a short period of time. (One also wonders how much money the plaintiff may have been able to extract in settlements before dismissing the case.)
In the end, the decision to sanction and to award fees is left to the discretion of the judge. And while it is encouraging to see Judge Rochon recognize the damage caused by acts that have become common in Schedule A cases (e.g., using FRCP 41 to dismiss defendants who fight back), it is discouraging to see a result that will only serve to further disincentivize Schedule A defendants from fighting back.
Once a judge grants a TRO with an asset freeze, the deck is heavily stacked against the Schedule A defendants. Defendants have strong incentives to settle, even when the cases against them lack merit. In many cases, it’s just too expensive to fight back, especially when your assets have been frozen.
If judges were willing to sanction plaintiffs—or at least shift fees—when Schedule A defendants were wrongfully restrained, that would do a lot to help level the playing field and incentivize the plaintiffs to bring better claims.
Without fee shifting or sanctions, the cost of bringing a nonmeritorious claim in a Schedule A case is virtually zero, while the harms to defendants who are wrongfully restrained—even for a short time—can be devastating. As Judge Hunt has noted, “the extraordinary remedy of freezing all [the defendants’] assets without notice” can “potentially ruin[] a legitimate business.”
Plus, as Casey Hewitt noted on Mastodon, Schedule A “defendants have no choice but to litigate, have no option to meet and confer and avoid a lawsuit . . . They did not ignore demand letters or refuse to negotiate or discuss alleged infringement.” But once they find out that their assets have been frozen, they have to “hire expensive IP litigators or they will lose their businesses.” In these circumstances, it seems like fee shifting for wrongfully enjoined Schedule A defendants should be the norm, not the exception.
Yes, it’s true that a presumption in favor of fee shifting would be a departure from normal federal court practice. But courts routinely use their discretion to grant procedural departures to Schedule A plaintiffs—e.g., email service, ex parte asset freezes, mass joinder upon conclusory (and in many cases, dubious) allegations. Perhaps it is time for judges to start using their discretion to make routine departures for Schedule A defendants, too.
Additional observations:
This case is a good example of why patent litigation is a poor fit for the Schedule A litigation model. I’ve written here before about how design patent infringement is ill-suited to ex parte adjudication; so too is utility patent adjudication. If judges are going to keep allowing the Schedule A model in patent cases (and they don’t have to do so, in these cases or in any others), they should consider making it a regular practice to special masters to help analyze the infringement evidence at the TRO stage. And they should, at a minimum, require an individualized claim chart for each and every defendant.
It is far from clear that Judge Rochon actually had the power to freeze these defendants’ assets in the first place. As Judge Kendall noted in a recent order, 35 U.S.C. § 284 “does not provide for the equitable relief of accounting and profits,” which is seems to be the standard basis for asset freezes in other types of IP cases. Furthermore, as Judge Seeger has noted, “Schedule A plaintiffs typically don’t request and receive equitable monetary relief” at the end of their cases, even when equitable relief is available. Other judges might be well-advised to start questioning whether they should use their discretion to keep granting these types of asset freezes, even in cases where a remedy of equitable disgorgement is actually available.
As Professor Crouch has noted, the Federal Circuit has granted rehearing en banc in the design patent case of LKQ v. GM. The main question raised in the petition for rehearing—and in the court’s order granting that petition—is the continuing viability of the Rosen primary reference requirement for evaluating whether a design is obvious under § 103.
The court has ordered briefing on the issue of whether Rosen and Durling (the Federal Circuit case that expanded upon the Rosen approach) were overruled, abrogated, or otherwise affected by the Supreme Court’s decision in KSR and, if so, what the test should look like going forward. For my own thoughts on those issues, see this article and this post. (Tl;dr: The primary reference requirement is good but the Federal Circuit has applied it too strictly.)
In this post, however, I wanted to discuss another issue raised by the court. In granting LKQ’s petition for rehearing, the court asked:
To the extent not addressed in the responses to the questions above, what differences, if any, between design patents and utility patents are relevant to the obviousness inquiry, and what role should these differences play in the test for obviousness of design patents?
These are good questions and well worth discussing. They also stood out to me in light of some of the arguments that were made at the petition stage.
In support of LKQ’s petition for rehearing, some of my friends and colleagues submitted an amicus brief wherein they argued against what they called “design patent doctrinal exceptionalism.” According to these amici, “except where Congress has explicitly specified otherwise, there should be no ‘exceptional’ approach to design patent law doctrines that render them different from utility patent law doctrines. Congress, not the Supreme Court, has imposed this requirement of consistency.” I respectfully disagree.
Section § 171(b) provides: “The provisions of this title relating to patents for inventions shall apply to patents for designs, except as otherwise provided.” But the statute says nothing about legal “doctrines.” It refers only to the statutory “provisions.” Saying that a statutory provision applies to design patents is very different than saying that judicial decisions about how to apply that provision to utility patents also apply to design patents.
Indeed, due to the differences between utility patent and design patent claims, neither courts nor the USPTO could directly import those rules anyway. Design patents cover different things (ornamental designs as opposed to useful inventions). They are claimed differently (using central as opposed to peripheral claiming). They are fundamentally different types of patents. General principles may be transferrable but the actual tests are generally not.
Consider novelty. Section 102 says a patentable invention must be novel. Section 171(b) says that this requirement of novelty applies to designs. But that doesn’t mean that the judicially-created tests for determining when a useful invention is novel also must be used to determine when a design is novel.
A useful invention is deemed not to be novel if all of its elements can be found in a single prior art reference. But design patent claims don’t have elements in the same way that utility patent claims have elements. (A design patent claim consists of a pro forma verbal portion that incorporates by reference drawings that show the claimed design.) So courts can’t simply apply the “all elements” rule to design patents. Instead, courts have, quite rationally and fully consistent with § 171, developed a different anticipation test for design patents (for more on that test, see this short piece). That’s not “design patent exceptionalism,” it’s a thoughtful and appropriate application of a generally-applicable statutory requirement to a fundamentally different type of patent.
A similar problem occurs with § 103. Because design patent claims don’t have “elements” in the way utility patent claims do, we can’t just apply the utility patent mix-and-match approach to design patents.
Some might argue that we could just let litigants chop up design patent claims into “elements” in litigation and let them fight it out utility-patent style. But that would be inconsistent with longstanding case law emphasizing that the thing protected by a design patent is the design as a whole. It would also be inconsistent with design theory. Not to mention the significant litigation and uncertainty costs such an approach would entail.
A visual design isn’t simply a collection of visual pieces. It is, as a group of amici explained during Apple v. Samsung, a “cohesive and integrated whole.” While it might be technically obvious to take existing visual pieces and recombine them into new shapes or surface designs, that doesn’t mean that doing so always (or even often) creates a visually obvious result.
It’s true that actual designers don’t (often and definitely not always) start designing by taking a primary reference and modifying it. But the question of what we should deem obvious is ultimately a policy question, not a factual question about the actual processes of invention. Cf. The “Winslow tableau.” There’s nothing irrational about saying that, as a policy matter, we don’t think a design should be deemed visually obvious when there’s nothing that looks “basically the same” in the prior art.
This, of course, means that we need a concept of what looks “basically the same.” In particular, there needs to be a meaningful difference between what is deemed “basically the same” (i.e., similar enough to be a primary reference) and “the same” (i.e., similar enough to anticipate/infringe). The way the Federal Circuit has been applying Rosen doesn’t seem to leave much blue sky between those two concepts. But that is a problem with the application of Rosen, not a problem with Rosen itself.
One more thing: It’s true that it’s difficult to invalidate design patents. (If you’re tempted to ask about those old studies, read this.) It’s also true that the Rosen approach leaves some plainly uncreative designs immune from § 103 attacks—or at least, from successful ones. But maybe § 103 isn’t the best way to address that issue. Maybe, as I argue in this forthcoming article, courts should start taking the originality requirement of § 171(a) seriously.
By Sarah Burstein, Professor of Law at Suffolk University Law School.
Columbia Sportswear North America, Inc. v. Seirus Innovative Accessories, Inc., 21-2299 (submitted but not decided) (oral argument recording available here)
The Federal Circuit heard oral arguments yesterday in the second round of Columbia v. Seirus. (Prior Patently-O coverage of this appeal is available here.) My 2015 article, The Patented Design, was mentioned several times during the argument.
In that article, I argued that a design patent’s scope should be limited to the design as applied to a specific type of product. In making that argument, I acknowledged that this approach could create some line-drawing problems, including the type that have arisen in Columbia v. Seirus. I suggested that, with respect to infringement:
[O]ne solution would be to put the burden of proof on the patent owner to show that the accused device should be considered the same type of product. Courts have been tasked with determining which products are and are not the same “type” of product in the trademark context; there is no obvious reason why they should not be able to do the same in the design patent context.
If, however, the line-drawing problem proves to be intractable, an alternative would be to determine product “types” according to the Locarno Agreement Establishing an International Classification for Industrial Designs. Specifically, “type of product” could be defined to map onto Locarno sub-classes. For example, Locarno Class 10 includes subclasses for “Clocks and Alarm Clocks,” “Watches and Wrist Watches,” and “Other Time-Measuring Instruments.” The Locarno classification system is not perfect for this use but it may provide a second-best solution if judicial common law development proves unworkable.
Sarah Burstein, The Patented Design, 83 Tenn. L. Rev. 161, 219–20 (2015) (footnotes omitted). So my argument was not that design patent scope (and in this case, the scope of comparison prior art) must be limited to the exact, specific type of article named expressly in the verbal portion of the design patent claim.
Instead, I argued that courts should look to whether the accused product (or reference) is the same general type of article.
I also suggested that the product type could be determined with reference to Locarno sub-classes. How might that work in practice?
Looking at the dispute in Columbia v. Seirus, the most relevant class would seem to be Class 5, “Textile piece goods, artificial and natural sheet material.” Here are the relevant class subheadings:
So which subclass(es) would apply here? This case is an example of how the “substantial latitude” that design patent applicants are given in describing the relevant article in their claim language, see MPEP § § 1503.01(I), might sometimes complicate later attempts at classification.
The asserted design patent, U.S. Patent No. D657,093, claims “the ornamental design of a heat reflective material.” The only Class 5 subclass heading that uses the word “material” is 05-06, “Artificial or Natural Sheet Material.” But the patented design seems to be directed more to something like the examples mentioned in subclass 05-05, “Textile Fabric.” (Indeed, if you look closer at the full Locarno Classification, subclass 05-05 specifically includes ID No. 100480, “Insulating fabrics.”)
So which bucket would the D’093 fit in best? Where classification is disputed, the burden of persuasion should be on whichever party wants to prove something is the same type of product.
For infringement, then, the patent owner should have to prove the accused product is the same type as what is claimed. For comparison prior art, the accused infringer should have to prove the reference is the proper type. (This is because the use of the prior art in evaluating infringement is a one-way ratchet; it can only be used to narrow the presumptive scope of a claim, not to enlarge it. See Sarah Burstein, Is Design Patent Examination Too Lax?, 33 Berkeley Tech. L.J. 607, 612 (2018).)
Evidence relevant to this inquiry could include information about any commercial embodiments, including information the patent owner has provided to regulators about their product or market. And if a patent owner has registered their design internationally, the design may already have a Locarno sub-class.
Who should decide? Judges would be well-suited to make these types of determinations. Indeed, these issues are likely to arise at the Markman stage or, as in Columbia v. Seirus, in motions in limine. This is not the kind of credibility or historical-facts determination that we normally leave to juries. Indeed, it’s difficult to imagine a way to effectively instruct juries on this issue without injecting undue confusion into the infringement analysis.
Guest Post by Sarah Burstein, Professor of Law at the University of Oklahoma College of Law
Wepay Global Payments LLC v. PNC Bank, N.A. (W.D.Pa. June 1, 2022) [wepayDecision]
Companies associated with William Grecia have filed over a dozen cases alleging infringement of design patents for “animated graphical user interfaces.” A judge in one of those cases, Wepay v. PNC Bank, recently issued a decision dismissing the case.
The patent asserted in that case, U.S. Patent No. D930,702, was issued in 2021 and claims a “design for a display screen portion with animated graphical user interface.” Wepay alleges infringement of the second embodiment, which includes three images:
Consistent with the USPTO’s rules for “changeable computer generated icons,” see MPEP § 1504.01(a)(IV), the patent specifies that, “[i]n the second embodiment, the appearance of the sequentially transitions between the images shown in FIGS . 3 through 5. The process or period in which one image transitions to another image forms no part of the claimed design.” Consistent with the USPTO’s general drawing rules, the patent disclaims all matter shown in broken lines “form[] no part of the claimed design.”
So what does this patent cover? According to the court, Wepay seemed to think its patent covered any app with a QR code that lets you send money to other people:
WPG maintains that both its patent and the PNC app include an icon array of three squares that simulate a QR code; both cycle to a functional screen where the user may choose to whom the money will be sent; and both conclude with the display screen with a display of a zero value, where the user may input the amount of money that the user wishes to send.
The court rejected that interpretation—and rightly so. It’s well-established that design patents cover the visual designs that are actually claimed, not the larger design concepts.
The court held that the claimed designs and accused design, shown below as shown in the complaint, were plainly dissimilar and granted PNC’s motion to dismiss.
In doing so, the court recognized that placement and proportion matter in design, noting that “[a]ny similarity between the two designs is limited to basic geometric shapes, but with notable differences in shape size and spacing such that no ordinary observer would mistake the Accused Design with the Asserted Design or vice versa.”
Some may argue that the court erred in considering the placement and proportion of the claimed design fragments, because matter shown in dotted lines “forms no part of the claimed design.” But even if other visual portions of the design are disclaimed, that doesn’t mean courts have to treat the claimed portions as free-floating motifs. (Indeed, sophisticated design patent owner Nike has recognized that placement matters, even when dotted lines are used. See this PTAB decision at 10-13.)
I’ve explained here, placement and proportion are essential elements of design. The differences between the claimed and accused designs here are an excellent example of how different a single motif (e.g., each of the three squares in figure 3 and in the accused design) can look in the context of the design, depending on its proportion to the other elements and where it is located on the screen.
Before diving into the merits, however, the court expressed some doubt over whether infringement was even possible in this case:
[T]he Court notes that the ordinary observer test focuses on a hypothetical purchaser induced to buy a product with an accused design over an asserted design. The cases cited by the parties likewise address designs where there exists a consumer and/or purchaser at play. However, neither the cited cases nor this Court’s research reveals a case where, like the instant matter, the consumer has not voluntarily chosen the design at issue. Instead, the Accused Design is incidental to the PNC customers utilization of the mobile application. The Complaint does not allege that a purchaser would have been or has been involved with the Accused Design. Further, the Complaint does not allege a purchase or transaction by a consumer with regard to the Accused Design over the Asserted Design. Thus, the ordinary observer test would seem not to fit squarely with the designs at issue, and WPG would not be able to assert a design patent infringement claim.
Even though it’s dicta, the court’s suggestion that a design patent owner must plead “a purchase or transaction by a consumer with regard to the Accused Design over the Asserted Design” is worth further discussion.
Like other patents, a design patent is infringed when someone makes, uses, sells, offers to sell, or imports the patented invention without permission. The test for design patent infringement doesn’t change that. It just tells us when someone is making, using, selling, offering to sell, or importing the patented design.
The test for design patent infringement is a test of visual similarity. The “ordinary observer” part just tells us how similar the designs must be. (Seethis article at page 177.)
So, under current law, no sale is required, let alone—as the court seems to suggest—a direct sale to the end user, “with regard to the Accused Design over the Asserted Design.” Indeed, because there is no working requirement for design patents, there may not be a product embodying the claimed design available for sale.
All the defendant has to do is make (or sell, etc.) a product that looks “the same” as the claimed design. (For more on how the test for design patent infringement works, seethis essay or Chapter 12 of this free casebook.)
Because Grecia-related companies have filed so many similar cases, it seems likely this decision will be appealed. It is definitely one to watch.
By Sarah Burstein, Professor of Law at the University of Oklahoma College of Law
In December 2020, the USPTO put out a request for comments on “The Article of Manufacture Requirement.” (For prior PatentlyO coverage, see here.) Last week, at Design Day 2022, the USPTO announced that it had completed its summary of those comments. That summary is available here.
In the request for comments, the USPTO floated the idea of issuing design patents for what they are now calling “designs for projections, holograms, and virtual and augmented reality (PHVAR).” Presumably, this acronym is meant to suggest that these are new types of designs, uncontemplated by Congress when it passed statutes that—according to some—are in need of “modernization.”
But projected designs aren’t new.
As we explained in our law professor comments, designs for “magic lantern” projections go back at least as far as the 17th century. So the concept of projected designs was well-established by the time Congress passed the first design patent act in 1842—not to mention when it codified the Patent Act in 1952. Projected designs are simply not some new, unforeseen concept.
There may be new types of projection machines today, but this type of design is not new.
The USPTO suggests that the scope of design patentable subject matter “should be reevaluated in order to incentivize and protect innovation in new and emerging technologies.”
But it is not—and should not—be the role of the design patent system to incentivize or protect new machines. We have utility patents for that.
Furthermore, projected designs are not as the USPTO seems to suggest, unprotected today. Projected images qualify as pictorial, graphic, or audiovisual works that are already protected—costlessly and immediately upon fixation—by copyright. It can’t be seriously suggested that Congress was unaware of projected designs when it passed the broad definitions of protectable subject matter in the Copyright Act of 1976. It’s clear from the definitions that one form of projected creativity, cinema, was very much on Congress’ mind. See, e.g., 17 U.S.C. § 102(a)(6) (providing protection for “motion pictures and other audiovisual works.”).
Again, the type of projector may be new, but the type of work is not.
So why push for design patents when copyright is already available? There may be some who just want to accrue as many belts and suspenders as their IP budgets will allow. But that alone is not a good enough reason to expand design patentable subject matter.
It may be that, for at least some potential applicants, the goal is to avoid the requirements of copyright law. In Feist v. Rural, the Supreme Court held that a work must have “minimal creativity” to be protected. Thus, to gain copyright protection, a work must “possess some creative spark, ‘no matter how crude, humble or obvious’ it might be.” This is, according to the Supreme Court, a requirement of the Progress Clause.
Minimal creativity is a low bar. Yet the USPTO has granted numerous design patents for designs that fall below it—especially in the context of computer-generated designs. In theory, novelty and nonobviousness could be a higher bar but, in practice, that is not the case.
So let’s all be clear:
These are not new types of designs.
They’re not unprotected—unless they fail to meet the low constitutional threshold set in Feist.
In our new paper, The Truth About Design Patents, we debunk three widely held—but incorrect—views about U.S. design patents. Taken together, these myths paint a grim picture of design patents:
Half of all design patent applications are rejected.
Most asserted design patents are invalidated in litigation.
Most litigated design patents are not found infringed.
By this account, design patents are hard to get, hard to defend even if you do get them, and hard to recover on even if you do get and defend them. These beliefs are commonplace in the literature and often invoked without citation. But, when we traced these assertions back, we found an origin story of weak support and misunderstood data.
In our paper, we provide more current, transparent statistics to replace each myth that we debunk. However, we strongly caution against over-interpreting our findings.
Where do these myths come from?
The origins of these claims lie in three old and misunderstood sources: a 1985 study by Thomas Lindgren, a 1979 study by the USPTO, and a 1953 study by Raymond Walter. The Lindgren study contributed to all three claims, whereas the USPTO and Walter studies contributed mainly to the second claim—regarding frequent invalidation during litigation.
One big problem with these studies is that, as you can see from the summary above, they’re all really old. Walter’s data preceded the passage of the Patent Act of 1952. Even though the most recent one, the Lindgren study, was published in 1985, it only considers data through 1983. And the Federal Circuit hadn’t yet decided its first design patent case at that point. So these studies can’t tell us much about what’s going on in the Federal Circuit era, unless one assumes—unreasonably—that the Federal Circuit didn’t meaningfully change design patent law.
We also found other problems with these studies. For example, Lindgren noted in his article that his sample size was too small to draw statistically significant conclusions. But his data was cited by others, in citation networks we trace in much more detail in the paper, as if it were statistically significant. Moreover, a 1999 paper mis-cited Lindgren for a claim that the 1985 study didn’t actually make, and that mistake is what hardened into the first piece of conventional wisdom noted above: that half of all design patent applications are rejected.
What’s the real story?
Recognizing the best way to fight bad data is with better data, we also offer corrections and updates to the record. The following findings about the current state of affairs follow the same sequence: acquiring design patents, defending them, and enforcing them.
Acquiring Design Patents
For nearly the last quarter-century, the success rate of design patent applications seems to have been over 85%, not the 50% that Frenkel reported in 1999. (We note that our findings are consistent with what Professor Crouch found in his 2010 study of design patent examination.)
Allowance of Design Applications as a Share of Total Dispositions (1989–2020)
Allowance Rate of Design Applications as a Function of Total Design Applications Filed in the Same Year (1989–2020)
Our source is the USPTO’s own annual reports going back to 1989, the oldest data that is readily available. This data shows that, since 1989, the apparent grant rate for design patents has stayed at or above some 70%—whereas the volume of design patent filings has quadrupled over the same period. By the mid-1990s, that rate was nearly 80%.
But most regular design applications (i.e., those not filed through the Hague System) are kept confidential and unpublished unless and until they issue as patents. So how can we reach this finding if application-level data is all broadly unavailable? The USPTO annual report data reports three totals relevant for us: (1) design applications newly filed that year, (2) design applications allowed that year, and (3) design applications abandoned that year. From this, we can infer the success rate of design applications. It’s not perfect but it’s the best we can do right now.
Defending Design Patents
Similarly, our research did not support the assertion that most litigated design patents are invalidated in the courts. In fact, the opposite is true. For the period of 2008–2020, district courts making validity determinations about design patents upheld them (i.e., found them “not invalid”) 88.4% of the time—and only 11.6% of these determinations resulted in a patent being invalidated. Relatedly, over the same period, district courts making enforceability determinations about design patents upheld them 99.5% of the time—and only 0.5% were found unenforceable. The full paper explains how we reached these figures, which are based on a study of nearly 1200 case documents in U.S. district courts where a determination was made about patent validity or enforceability.
(We note that our findings are consistent with what Andrew Torrance found in his 2012 study of design patent litigation. To some extent, our findings are even more patentee-friendly than the Torrance study.)
We also looked at design patent outcomes in the International Trade Commission and the UPSTO Patent Trial and Appeal Board. There were fewer cases in these tribunals, but the data we found paints a picture that is far from grim.
For the period of 2011–2020, ITC determinations about validity upheld the patent 95% of the time and invalidated only 5%. Given the relatively small population of ITC design patent cases, we caution against strong conclusions about strategic behavior across tribunals as is often seen with utility patents. Still, the available data does suggest that the ITC is not invalidating two-thirds of the design patents asserted there.
Meanwhile, from when PTAB trial proceedings came online in 2012 through 2020, the survival rate of design patents in these administrative trials has been 79%, reflecting a survival rate of 78% among IPR cases and of 81% among PGR cases. We refer to an overall survival rate because patents can survive (1) when petitions challenging them are never instituted—i.e., taken up for review by the PTAB—as well as (2) when they are evaluated on the merits and are found not invalid. The overall PTAB survival rate we report accounts for both.
Enforcing Design Patents
Finally, our research did not support the assertion that most litigated design patents are found not infringed:
Findings of Infringement as a Share of District-Court Design Patent Infringement Findings (12-mo avg) (2011–2021)
Among district court cases decided between 2011–2021, the share of decisions in which the court found infringement showed some volatility, but generally, the share of design patent decisions in which the court found infringement has remained almost entirely above 50% since 2011. Since 2015, the share of design patent decisions in which the court found infringement has remained above 80%.
Again, we caution against over-interpretation of these findings. Our aim is to correct the record and reorient a conventional wisdom that rests on an inaccurate and incomplete statistical picture. Thus, we are acutely mindful of the risk in replacing one mistaken view with another.
What comes next?
A number of people have called for changes to the design patent system, relying frequently on the myths that we have now debunked. We hope that our paper will bring more light to these debates and, at a minimum, encourage more fact-based discussion on these important policy issues. We also call for more empirical research and for more transparency of data in this area. For example, we see no good reason for the USPTO not to publish all design patent applications. We know that the USPTO is bound by current law, and we call on Congress to reconsider that policy and to stop exempting design patents from the general publication requirements of 35 U.S.C. § 122.
Reasonable minds might differ over whether 18 months is the right time for publication as is the case for utility patents, but there is no good reason—at least none we’ve yet heard—to keep them secret forever (unless granted). Publishing these applications would not only provide more data for research and policy analysis but would also provide transparency on the issue of how the USPTO handles applications for this increasingly important category of patents.
= = =
Editors Note: We require disclosure of conflicts-of-interest for guest posts, including both financial interest and representation of interested parties. One of the authors (Vishnubhakat) indicated an attenuated potential financial conflict in that he owns stocks in a number of companies, including some tech companies that hold utility and/or design patents.
