[co-authors: Mackenna Dunn, and Garrett "Jake" Lee*]

Over the last two years, we have studied the examiner affirmance rates of the Patent Trial and Appeal Board (PTAB) for § 101 rejections. The PTAB is the administrative court of the U.S. Patent and Trademark Office (USPTO) that handles applicant appeals of examiner rejections, as well as inter partes reviews (IPRs) and post-grant reviews (PGRs). Our analyses of 2021 and 2022 data were less than good news for applicants, with the PTAB affirming examiner § 101 rejections 87.1% and 88.4% of the time, respectively.

Given that the PTAB's overall affirmance rate is about 58%,[1] the extreme difficulty of winning a § 101 dispute at the PTAB is shocking. While some might contend that this is a natural process through which the PTAB weeds out weak inventions, it is actually because § 101 jurisprudence has evolved into a hot mess. The Supreme Court's decision in Alice Corp. v. CLS Bank Int'l set forth a notoriously vague test, the Federal Circuit followed this with conflicting case law, and the PTAB does not even follow the USPTO's own 101 examination guidance. This has opened the door to arbitrary and conclusory reasoning often winning the day.

As a result, there have been a number of absurd outcomes, such as the ITC finding a diamond-encrusted drill bit to be an abstract idea and the Federal Circuit coming to the same conclusion about a camera phone. Furthermore, prosecution of software applications before the USPTO has become highly examiner-dependent rather than being based primarily on the actual claim language at issue. In other words, we are living in a timeline in which something went terribly wrong over the last two decades.

In any event, the examiner affirmance rates for 2023 provide little hope for applicants. As noted in the title, it is 91% -- an increase from its already high numbers in 2021 and 2022.

To gather the underlying data, we adopted an approach that was virtually identical to that of last year. From the PTAB's search interface, we specified the following criteria: decision dates between January 1, 2023 and December 31, 2023, a proceeding type of "appeal", a decision type of "decision", and an issue type of "101". These search results were filtered to focus only on decisions in which the applicant appealed an examiner's Alice-based § 101 rejection and the PTAB ruled on this basis of appeal.[2] In other words, we excluded cases in which: (i) the appeal was not of a § 101 issue but the PTAB set forth a new grounds of rejection based on § 101, and (ii) the appeal was of a § 101 issue but the PTAB did not decide the case on § 101 grounds (e.g., the examiner withdrew the § 101 rejections after the appeal was filed).

This took the 633 decisions returned by the search engine down to 495 substantive § 101 decisions.[3] Compared to 2022, the 2023 overall numbers are quite similar. There were 634 decisions and 482 substantive § 101 decisions in 2022.

Figure 1

Figure 1 shows these overall results broken down by technical center (TC) from which the appeals originated.[4] While the affirmance rate per TC is all over the map, the outliers are TCs 3600 and 3700 which are both over 95%. This should surprise no one as so-called "business method" inventions are routed to these TCs. When considered together, the vast majority (over 76%) of all appeals are out of TCs 3600 and 3700, demonstrating how frequently examiners in those TCs issue § 101 rejections.

The affirmance rate for TC2100 (computing technologies) has ticked up from last year, jumping to over 85% from 80%. This demonstrates that current § 101 practice is having a deleterious impact on software patenting in general and not just software-implemented business methods. Also, the number of appeals out of TC2100 has increased from 45 to 70, suggesting that examiners in this TC have been issuing more § 101 rejections starting in late 2021. The affirmance rate for TC1600 is also fairly high, likely due to a number of notoriously tough examiners in the bioinformatics art units.

There are too few data points in any of the other TCs to draw any conclusions. But these results do suggest that a software invention is likely to receive a more favorable examination in TC2100 than TCs 3600 or 3700, and that TCs 2400 and 2600 may be the most favorable TCs for software examination.

Figure 2

Next we considered the type of abstract idea used in the examiners' § 101 rejections. For the abstract idea exception to patentable subject matter, the three main categories are mathematics, mental processes, and methods of organizing human activity. In other words, a claim is deemed ineligible for patenting if it is directed to mathematics, a mental process, or a method of organizing human activity without significantly more. Unfortunately, these categories are ill-defined.

As shown in Figure 2, of all substantive affirmances of § 101 rejections by the PTAB, 17% were based on mathematics, 48% on mental processes, and 68% on methods of organizing human activity. The reason that these three numbers do not sum to 100% is because some examiners and PTAB panels concluded that claims were directed to more than one of the categories (some were placed in all three). Not surprisingly, mental steps was the most prevalent category for TC2100 (likely due to the fact that this category has been interpreted to be broadly applicable to software inventions by the Federal Circuit). Likewise, the high rate of organizing human activity categorizations in TC3600 was expected due to most business methods being conducive to placement therein.

Finally, we looked into situations where the PTAB provided new grounds of rejection based on § 101 (as noted above, most of these decisions were omitted from the analysis leading to Figures 1 and 2). This is a troublesome thought for many applicants considering whether to appeal a § 102 or 103 rejection. The PTAB is permitted to reverse the examiner on those grounds but set forth a new § 101 rejection, even if the examiner found the claims eligible under § 101 and the issue was not a subject of the appeal.

Figure 3

The results are shown in Figure 3. The overall likelihood of receiving a new ground of rejection under § 101 is only 10%. But it is slightly higher for TC2100 and slightly lower for TC3600. This is probably because examiners in TC2100 tend to give fewer § 101 rejections while examiners in TC3600 already give § 101 rejections very frequently. While the numbers for TCs 2400 and 2600 may be too small to be significant, they appear to confirm the supposition that examiners in those TCs give relatively fewer § 101 rejections.

Notably, most of these new grounds of rejection were given for applications where the examiner did not provide a § 101 rejection that was ultimately the subject of the appeal. The results indicate that the concern of receiving a new § 101 rejection from the PTAB in this situation is not unfounded, though it does not happen frequently.

Summing all of this up, the 2023 data is further confirmation that appealing a § 101 rejection to the PTAB is a highly risky endeavor and likely to result in an affirmance. Further, appealing any non-101 rejection to the PTAB could result in the PTAB issuing its own § 101 rejection. Thus, continuing to work with an examiner to find allowable subject matter is the recommended course of action, even if the examiner is difficult. Of course, some examiners -- such as those with very low allowance rates -- are unlikely to consider any amendment to be allowable, and therefore an appeal may be justified. Moreover, the 2023 data also confirms that the use of TC steering tools should be part of every practitioner's workflow. These tools use analytics to predict the TC to which a patent application is likely to be routed, and with them it is easier to prepare applications that are more likely to avoid applicant-hostile TCs 3600 and 3700.

* Garrett "Jake" Lee is a summer clerk with MBHB who is attending George Washington University. Jake graduated from George Mason University with a Masters in Applied and Engineering Physics.

[1] From the USPTO's appeal statistics for its fiscal year 2023. This number jumps to around 67% if affirmances-in-part are considered.

[2] To simplify the analysis, we considered a decision to be an "affirmance" of the examiner's § 101 rejection if the PTAB held at least one claim invalid under § 101.

[3] Many of the 633 decisions were of rejections on other grounds (e.g., §§ 102 or 103). We did not consider anything but the § 101 determinations. But we omitted decisions with § 101 rejections based on laws of nature or natural phenomena, non-statutory double patenting, utility, and claims that did not fall into one of the four statutory categories (which accounted for only a handful of the decisions).

[4] RD00 is a new, experimental TC that supposedly is an attempt to make examination more efficient and consistent.

[View source.]