Sixteen Years of §101: What Actually Moves Examiners

by Dennis Crouch

In Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012), the Supreme Court set out its now familiar two step framework for determining eligibility under 35 U.S.C. § 101.  But the patent system did not really shift until two years later when the Court reiterated the same test. Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014).

Part of the story is that the legal framework only does some of the work.  In any system, we also have to look at how that law is administered.  For me, this means patent examination data. I have been pulling office action data going back to 2010 (several million rejections in total) and using a custom classifier to identify §101 rejections.

The headline finding is that … (I know this is rude, but this data is so interesting that I decided keep it just for subscribers).

When Alice came out in 2014, we had a USPTO Director (Michelle Lee) ready to act and she pushed examiners to immediately adopt the Supreme Court’s new ruling. As the chart shows above, Mayo generated essentially no movement at the examiner level. After Alice, though we had produced an immediate vertical jump in 101 rejections. The January 2019 Patent Eligibility Guidance (PEG) issued under Director Andrei Iancu, and led to a near-vertical drop in the 101 rejection rate, even though no Supreme Court or Federal Circuit decision had changed the underlying law.

What we have seen since then is interesting. Beginning in 2021, 101 rejection rates began drifting the Hirshfeld and Vidal years before the recent transition to Director John Squires. The Squires-era data shows only a small downtick from the Vidal peak, despite the Director’s loud public statements about restoring eligibility.

I wrote previously about the pro-patentee 101 jurisprudence at the PTAB under Dir. Squires.  After writing those posts, I began to hear from patent prosecutors that they were seeing something different, that examiners were not backing down and instead were continuing to reject cases on eligibility grounds.  This data confirms their sense that the rise in 101 rejections over the past 6 years has not gone away.

Methodology: The chart above presents the monthly share of non-final and final rejections issued in published applications that contain at least one §101 rejection. The data run from January 2010 through early 2026. Identifying §101 rejections programmatically is harder than it sounds. The USPTO publishes an internal classification of office action rejection grounds, but I found that field incomplete and inconsistent. I instead built a regular-expression (REGEX) classifier that scans the rejection text itself for canonical §101 language, validated by hand on dozens of randomly drawn cases and supplemented by AI-assisted spot-checking on several hundred more.

The Iancu Correction: The third vertical line marks the January 2019 PEG.  Those examination guidelines restructured Step 2A of the USPTO guidance by adding the Prong Two practical-application inquiry, restricted the categories of qualifying abstract ideas, and instructed examiners to evaluate eligibility without resort to ad hoc analogies to prior judicial decisions. None of those moves was compelled by Supreme Court precedent. They reflected Director Iancu’s policy judgment that the post-Alice examination regime had become unmoored from administrability principles and was producing unjustified rejections, particularly in software and computer-implemented technologies.

The Federal Circuit, for its part, has continued to apply its own §101 case law on a parallel track. The PEG occasionally surfaces in Federal Circuit opinions as a reference point, but the court has been consistent in treating it as deserving no deference. That means that a software claim allowed by an examiner under the PEG framework remains vulnerable in district court litigation where Federal Circuit doctrine controls.

The Hirshfeld and Vidal Drift: From the 2020 low point, examiner-level §101 rejection rates climbed gradually through the tenures of Acting Director Andrew Hirshfeld and Director Kathi Vidal. The drift is more puzzling than the earlier movements because it occurred without any formal change in agency guidance. The 2019 PEG remained on the books and the MPEP §2106 framework was unchanged. Several explanations are plausible, but my general thought is simply the shift away from Dir. Iancu (who was pushing the rate down) and to Dir. Vidal (who supported more extensive 101 inquiry).

I did a tech-specific analysis to consider whether more applications being filed in 101 risky areas explained the trend, but they did not.

Dir. Squires Endpoint: The chart’s right-hand boundary reflects just the first few months of Director Squires’ time in office. Squires was confirmed in September 2025 and has been unusually public about his views on eligibility. His first formal speech as Director articulated what he called the “three pillars of eligibility” framework. He issued the In re Desjardins Appeals Review Panel decision, designated as precedential, vacating a PTAB sua sponte §101 rejection of machine-learning training claims. He issued December 2025 guidance memoranda formalizing the Subject Matter Eligibility Declaration practice under 37 C.F.R. §1.132. My recent analysis of Board-level data shows reversal rates on §101 appeals nearly doubling under Squires, from roughly 10% to 29%. See Dennis Crouch, PTAB Doubles Section 101 Reversal Rate Under Director Squires, Patently-O (Jan. 2, 2026).

But, the chart shows those institutional moves have not led to drops in examiner-level rejection rates. An applicant facing a §101 rejection under Squires has roughly the same likelihood of receiving one as she did six months or even 2 years ago.

Business Methods: My aggregate chart – showing all the patents – smooths over some very important information – that rejection rates vary enormously across technology areas. The overall rates are pulled down by the large body of art units in mechanical and chemical areas where §101 is rarely the primary rejection ground.

The chart for art unit groups 3620 (e-commerce), 3680 (business cryptography and incentive programs), and 3690 (finance, banking, and insurance) shows the most extreme case the other way.  The pre-Alice baseline already ran at roughly 30-40%. The post-Alice jump took the rate to approximately 90%. The 2019 PEG produced a roughly comparable proportional drop to what the aggregate chart shows, but the floor it reached was about 50% – i.e., 50% of office actions included a 101 rejection. Under Hirshfeld and Vidal, the rate drifted back up to the high 50s.

What Moves Examiners

The Iancu-Squires comparison is the cleanest natural experiment available for asking what actually changes examiner behavior. Iancu issued a structured guidance document, accompanied by examiner training, that restructured the Step 2A inquiry and provided examiners with concrete categories and practical-application examples to apply. Rates dropped immediately and stayed down for a substantial period. Squires has issued precedential PTAB decisions, declaration-practice memoranda, and public speeches. Rates have moved only slightly. The difference here I see as institutional channel. Examination behavior responds training and guidance, less often PTAB decisions and Director speeches.

Ultimately, the deeper observation here has been there in the data since 2014. Patent eligibility at the prosecution stage is an administrative phenomenon. The Supreme Court sets its own outer limits of the doctrine and the Federal Circuit develops its application to particular fact patterns, but examiner behavior is governed by a parallel system of agency guidance that is more loosely coupled to the case law.