The Three Properties of Patent-Eligibility: An Empirical Study

By Michael Borella and Ashley Hatzenbihler* --

Patent eligibility is a bit of a mess these days.  Ever since the Supreme Court handed down the Alice v. CLS Bank decision six years ago, the distinction between what might be subject matter that can be patented and what is not has been blurry at best.

Famously, the Court legislated its own two-part test to determine whether claims are directed to patent-eligible subject matter under 35 U.S.C. § 101.  One must first decide whether the claim at hand involves a judicially-excluded law of nature, a natural phenomenon, or an abstract idea.  If so, then one must further decide whether any element or combination of elements in the claim is sufficient to ensure that the claim amounts to significantly more than the judicial exclusion.  But elements or combinations of elements that are well-understood, routine, and conventional will not lift the claim over the § 101 hurdle.  While the Alice inquiry is generally carried out as a matter of law, factual issues can come into play when determining whether something is well-understood, routine, and conventional.

However, the Court declined to define what it meant by "abstract idea" or "significantly more," leaving that for the Federal Circuit and the rest of us to figure out.  By admission of several of its own members, the Federal Circuit has failed to do so in a cogent fashion.  As just one example (and this is a doozy of an example) Judge Plager wrote in 2018's Interval Licensing v. AOL case that "[t]he law . . . renders it near impossible to know with any certainty whether the invention is or is not patent eligible" and "a search for a definition of 'abstract ideas' in the cases on § 101 from the Supreme Court, as well as from this court, reveals that there is no single, succinct, usable definition anywhere available."

Meanwhile, the USPTO's interpretation of Alice has added more steps to the test, clarified some aspects, muddied others, and counts for little to nothing in the courts.

Simply put, the problem is that we are spending too much time trying to ascertain the unknowable contours of Alice.  The Federal Circuit has separated and combined and separated and combined the two parts to the point where some have questioned whether Alice is actually a one-part inquiry in practice.  Further, at the prompting of ill-reasoned language from Justice Breyer in Mayo v. Prometheus, the Federal Circuit has amplified that case's overlap between §§ 101, 102, 103, and 112.

Rather than approach the Alice test with the rigidity of an automaton, there is a more pragmatic approach.  As described in a previous article, what the Federal Circuit is actually looking for to rule that a claim is patent-eligible can be expressed as three properties:  specificity, technical character, and novelty.  And not only does a claim almost certainly have to exhibit each to clear the § 101 hurdle, but there needs to be a nexus between these properties.  In other words, if one element of a claim is specific, another has technical character, and yet another is novel, that probably will not cut it.  Instead, a single element should exhibit all three.

That is not to say that claims lacking one of these properties are automatically ineligible -- there is a certain degree of flexibility in how the Federal Circuit balances the three.  But, as a rule, the more specific, technical, and novel your claim, the more likely it is to be patent-eligible.  Conversely, the less specific, technical, and novel your claim, the less likely it is to be patent-eligible.  A synthesis of the Federal Circuit § 101 case law makes this clear, and we have the data to back it up.

But before we get into the empirical stuff, let's do something that the Supreme Court did not do and define our terminology.

Claim specificity is not the same thing as claim breadth, although in practice the two are correlated.  Specificity, though, refers to whether and how much the claim recites a series of steps or set of elements that can be used to achieve a goal, rather than the goal itself.  The latter style has been referred to as "result-oriented," "outcome-oriented," and "black box" claiming, where the claim recites what the inventor hopes to achieve without reciting how the inventor purports to have achieved that goal.  The Federal Circuit often lauds the specificity of claims that it eventually find eligible (see McRO, Inc. v. Bandai Namco Games America Inc., Enfish LLC v. Microsoft Corp., Trading Techs Int'l v. CQG, Inc., and SRI Int'l Inc. v. Cisco Systems Inc.) while disparaging claims that lack this property (see Electric Power Group, LLC v. Alstom S.A., Two-Way Media Ltd. v. Comcast Cable Communications, LLC, and American Axle & Mfg. v. Neapco Holdings LLC).

Technical character is a bit trickier to explain.  Notably, § 101 lacks any requirement that inventions be technical, and even Alice only hinted that such a requirement might be the case.  But the Federal Circuit has run with these sub-dicta suggestions, using the term "technical solution to a technical problem" among others to describe what it is looking for in eligible claims.  To understand this property, it helps to imagine all inventions on a spectrum of technical character.  Those at one extreme are clearly technical because they improve the operation of machines (engine control, error correction coding, image sharpening, memory management in a computer).  Those at the other extreme are clearly more business-oriented than technical (derivative trading, annuity calculations, human resource management, retail schemes).  Put another way, the former improves how a machine works, while the latter improves how a business works.  The fact that business-oriented inventions can be run on computers (and may even require computers to obtain meaningful results) is largely irrelevant to this analysis -- such claims are still "non-technical".  Finally, in the middle of the spectrum we have inventions that improve how a person works (graphical user interfaces, video game animations, livestock management).  These could be viewed as either technical or non-technical depending on the actual claim language at play.

Novelty is a simple enough concept, and is well-understood in patent law.  The exact nature of the de facto prior art analysis under § 101 is unclear.  Nonetheless, comparisons between claimed inventions and prior art are found in a vast number of § 101 decisions including the aforementioned KPN, Chamberlain, Enfish, and McRO cases, as well as Cellspin Soft, Inc. v. Fitbit, Inc., Data Engine Technologies LLC v. Google LLC, Core Wireless Licensing S.A.R.L. v. LG Electronics, Inc., and Internet Patents Corp. v. Active Network, Inc., just to name a few.

So are the three properties a better way of understanding patent-eligibility than the two-part test?  We think so.  Do we like it?  No.  Do we think it is what the law should be?  Of course not.  But don't just take it from us.  Anyone who balks at the three properties should take note of Dropbox Inc. v. Synchronoss Techs. Inc. where the Federal Circuit recently wrote "an inventive concept exists when a claim recites a specific, discrete implementation of the abstract idea where the particular arrangement of elements is a technical improvement over the prior art" (emphasis added).

To quantify the three properties, we selected 24 notable post-Alice Federal Circuit decisions and scored them on a scale of 0-2 for each property, where 0 is a low score and 2 is a high score.  This scoring was not done in a vacuum -- we adhered to the following process.  If the Court spoke positively of a particular property in a decision, the score for that property would be a 2.  If the Court spoke negatively of a particular property, the score for that property would be a 0.  If the Court spoke neither positively nor negatively about a property or did not discuss the property in sufficient detail, we would review the claim language and the Court's discussion of the invention to come up with a value.

Thus, a specificity of 2 indicates that the claim describes the elements or steps involved in achieving its goal.  A specificity of 0 indicates that the claim is outcome-oriented.  A specificity of 1 indicates that the claim is partially descriptive and partially outcome-oriented.

Likewise, a technical character of 2 indicates that the claim is focused on a technical solution to a technical problem (i.e., improving the operation of a machine or system).  A technical character of 0 indicates that the claim is focused on a non-technical problem (i.e., improving the operation of a business).  A technical character of 1 indicates that the claim exhibits both technical and non-technical aspects (i.e., improving human performance or facilitating human decision making).

Similarly, a novelty of 2 indicates that the Court has stated that the claim is focused on an advance over the prior art.  A novelty of 0 indicates that the Court has stated that the claim is focused on a long-standing procedure or uses well-understood, routine, or conventional elements or steps to achieve its goal.  A novelty of 1 indicates that the Court made little or no comparison between the claim and the prior art.

Once determined, the specificity, technical character, and novelty scores for each case were added, resulting in summed values from 0 to 6.  Intuitively, one would expect that the higher this sum, the more likely that a claim is to be held eligible.

So, on to the data.

Case	Specificity (0-2)	Technical Character (0-2)	Novelty (0-2)	Sum (0-6)	Eligible?
DDR Holdings, LLC v. Hotels.com, L.P. (Fed. Cir. 2014)	2	1	1	4	Yes
Internet Patents Corp. v. Active Network, Inc. (Fed. Cir. 2015)	0	1	1	2	No
Versata Dev. Group, Inc. v. SAP America, Inc. (Fed. Cir. 2015)	2	0	0	2	No
Electric Power Group, LLC. V. Alstom S.A. (Fed. Cir. 2016)	0	0	1	1	No
Enfish LLC v. Microsoft Corp. (Fed. Cir. 2016)	2	2	2	6	Yes
McRo, Inc. v. Bandai Mamco Games America, Inc. (Fed. Cir. 2016)	2	2	2	6	Yes
Shortridge v. Found. Construction Payroll Service, LLC (Fed. Cir. 2016)	2	0	0	2	No
Thales Visionix Inc. v. U.S. (Fed. Cir. 2017)	0	2	2	4	Yes
The Chamberlain Group, Inc. v. Techtronic Indus., Co. (Fed. Cir. 2017)	0	1	1	2	No
Trading Techs Int'l v. CQG, Inc. (Fed. Cir. 2017)	2	1	2	5	Yes
Two-Way Media Ltd. v. Comcast Cable Commc'n., LLC (Fed. Cir. 2017)	0	1	1	2	No
Berkheimer v. HP Inc. (Fed. Cir. 2018)	1	2	2	5	Yes
Core Wireless Licensing S.A.R.L. v. LC Elec., Inc. (Fed. Cir. 2018)	2	2	2	6	Yes
Data Engine Tech. LLC v. Google LLC (Fed. Cir. 2018)	2	2	2	6	Yes
SAP America, Inc. v. InvestPic LLC (Fed. Cir. 2018)	0	0	2	2	No
Cellspin Soft, Inc. v. Fitbit, Inc. (Fed. Cir. 2019)	2	1	1	4	Yes
ChargePoint, Inc. v. SemaConnect, Inc. (Fed. Cir. 2019)	0	1	1	2	No
Koninklijke KPN N.V. v. Gemalto M2M GmbH (Fed. Cir. 2019)	2	2	2	6	Yes
SRI Int'l Inc. v. Cisco Sys. Inc. (Fed. Cir. 2019)	1	2	2	5	Yes
Trading Techs Int'l v. IBG LLC (Fed. Cir. 2019)	2	0	1	3	No
Elec. Commc'n. Techs., LLC v. ShoppersChoice.com, LLC (Fed. Cir. 2020)	0	0	0	0	No
Ericsson Inc. v. TCL Commc'n. Tech. Holdings Ltd. (Fed. Cir. 2020)	0	1	1	2	No
Dropbox, Inc. v. Synchronoss Techs. Inc. (Fed. Cir. 2020)	0	1	0	1	No
Uniloc USA, Inc. v. LG Elecs. USA, Inc. (Fed. Cir. 2020)	1	2	2	5	Yes

The 24 scored cases include 12 that were ruled eligible and 12 that were ruled ineligible.  Of the eligible cases, the sums of all landed in the range of 4-6.  Of the ineligible cases, the sums of all landed in the range of 0-3.  Thus, our initial intuition appears to be correct -- eligible claims tend to be more specific, technical, and novel, while ineligible claims are less specific, technical, and novel.  Notably, there were five with the highest possible sum of 6, but only one with the lowest possible sum of 0.

Overall, there was a correlation of 0.91 between sum and whether the claim was patent-eligible (we converted the "yes" or "no" results of the rightmost column to 1's and 0's respectively to perform correlations).  While this is a very high correlation, we need to keep in mind that in most cases correlation does not mean causation.  However, we believe that our analysis reflects a causative link from the properties of specificity, technical character, and novelty to eligibility.  In particular, the Court often states explicitly that a claim is eligible because it exhibits each of the three properties to at least some degree, or that a claim is ineligible because it lacks at least one of these properties.

We also calculated correlations between each of the properties and the eligibility outcome.  For specificity, the correlation was 0.58, for technical nature, the correlation was 0.80, and for novelty, the correlation was 0.74.  These results have an interesting implication for patent practitioners drafting initial claim sets or making amendments.  The technical nature of an invention appears to have the biggest impact on whether it is found to be eligible, but the practitioner has the least control over this aspect.  An invention that solves a business problem may be difficult to recast as solving a technical problem even using the cleverest claim drafting techniques.  Specificity and novelty, however, are more under the control of the practitioner, as detailed elements directed toward features believed not to be in the prior art can always be added (modulo the applicant's needs of course).  Thus, for a claim that is on the fence with respect to its technical character, making sure that it has at least one element that is specific and novel may help.

To sum up, the three properties appear not only to be a reasonably useful way of thinking about patent-eligibility in a world where the two-part test is bewildering and obtuse, it is also highly predictive of patent-eligibility outcomes.[1]  Nonetheless, while we attempted to avoid any bias during the scoring process, we are aware that there is a degree of subjectivity in our evaluations.  Moreover, the set of case law should be expanded beyond these 24 samples.  Therefore, we welcome others to suggest adjustments to the scoring and/or to continue this study with further cases.

[1] Of course, when you are before the USPTO or a Court, you need to argue the two parts of Alice.  But keeping the three properties in mind can help you hone those arguments.

* Ashley Hatzenbihler is a student at Loyola University Chicago School of Law where she is focusing her studies on Intellectual Property.  Ashley is a staff member of the Loyola University Chicago Law Review and a Summer Associate at McDonnell Boehnen Hulbert & Berghoff.  Prior to law school, Ashley received her Bachelors in Mechanical Engineering with a minor in Mathematics from Marquette University.  Ashley also completed her Masters of Science in Mechanical Engineering for her research on the optimal conditions for auto-ignition testing of diesel fuels and surrogates from Marquette University.

Comments

[skeptical]

'Celebrating' a broken system or methodology will NOT fix what is broken.

Instead of trying to import 'credibility,' the ethical thing to do is to challenge the broken items.

This article turns into nothing more than a slow uptick on the 'boil the frog' slowly motif.

[Moondog]

In Vanda the Court defined five specificities for a patent-eligible method of treatment claim. Your approach might gain some traction if persuasively argued to the Court

[MaxDrei]

The "take home" from what is for me a very impressive empirical study is that:

"...if one element of a claim is specific, another has technical character, and yet another is novel, that probably will not cut it. Instead, a single element should exhibit all three....."

To me, this looks like the "contribution approach" used by the specialist patents courts in England. The Boards of Appeal of the EPO don't like it because it requires dissection of the claim in order to pursue the preliminary enquiry into "eligibility". Is it now established, in the courts of the USA that 101 eligibility requires picking the claim apart in order to determine the point of novelty?

[Warren Woessner]

I like you attempt to analyze the requirements for patent-eligibility. But I think you should consider that the Fed Cir. has effectively grafter an "improvement" requirement onto what you have reviewed as a "technical solution to a technical problem." In fact you recognize this throughout your articles, e.g., as you define the "technical character of 2." This leaves open the question "Is technical novelty sufficient" sufficient per se or must the invention provide an advance over the prior art"?

[Michael Borella]

@Skeptical,

Love ya man, but we'll continue to agree to disagree on this one.

[Michael Borella]

@Moondog,

In a previous article, I've adapted the five specificities of Vanda to the abstract idea exception:

"Ideally for § 101 purposes, a claim recites a specific function of a specific element that does something in a specific way to obtain a specific result and thus solve a specific problem. Not all five specificities need be present in each claim, but the more of these specificities there are, the more likely the claim will survive an eligibility challenge. Conversely, the fewer specificities in the claim, the more likely it will be found abstract."

[Michael Borella]

@Max,

Under 101, the US courts have routinely been picking the claim apart despite stating that they review the claim as a whole.

[Michael Borella]

@Warren,

Good point. As far as I know, this has not been addressed in the case law (at least by the CAFC). One could certainly argue that a non-trivial, novel technical feature is an "improvement". But the requisite extent of any such improvement remains undefined.

[MaxDrei]

Thanks for that confirmation, Michael. Perhaps "as a whole" and "picking apart" are compatible though.

Thus: before you start picking the claim part you must carefully consider it "as a whole" because only by first carefully considering it "as a whole" can you go on then to pick it apart in a way that is fair to the inventor.

[Easwaran]

Thought provoking post, as usual. I think this is a good framework for drafting applications to address 101. And to address other things too, because if you think about it, you're also covering your 112 and art concerns with the specificity and novelty elements. And perhaps the "technical" lens is a good one for attaining a non-3600 classification.

That said, I would propose taking a more objective look at the cases to really determine the predictive power of this framework. For instance, rather than use the court characterizations on each metric to assign a score, explicitly do not do so and just review the courts representative claim. For instance, when a court says words to the effect of "this claim is specific," those statements are likely a run-up to an eligibility determination. But the claim might not actually be more specific than the representative claim of another case going the other way. Do if you think the right score is a 2, give the 2, then see what the court says. Similarly, it may be good to add an "overlap" score that reflects whether there are particular features contributing the favorable character in the various categories.

[skeptical]

Easwaran,

Your comment indicates one of the more insidious "boil a frog by slow degree" problems with the post hoc attempt to 'fit' the judicial muckery into some semblance of reason (and NOT to note that such Common Law law writing in the first instance is NOT the providence of the judicial branch):

EVERY SINGLE branch of the government from CAFC judges, to Director Iancu, to key senators have remarked that the judicial re-writing of 35 USC 101 has created a Gordian Knot of CONFLICTING case law.

This (ongoing) effort by the otherwise respectable Dr. Borella is NOT respectable.

It is itself a blight on patent law, and we will NOT rectify the mess by this type of post hoc 'rationalization.'

We should NOT be purposefully obfuscating 101 with the likes of 112, 102, and foreign law. We should NOT be supporting any such attempts.