USPTO Patent Quality - It's Not As Bad As You Think

Over the past five years, there has been an obsession within the patent community to find ways to increase "patent quality." No matter how you define it, most people agree that there have been periods where the PTO appeared dodgy in its role as a gatekeeper for "high quality patents." Sometimes, the criticism heaped on the PTO was fully justified. Other times, it was not.

Recently, James Malackowski and Jonathan Barney from Ocean Tomo published a fascinating article titled "What Is Patent Quality? A Merchant Banc's Perspective" in the June issue of Les Nouvelles (LES). The article studied many issues related to patent quality and found that many specific problems attributed to the PTO were somewhat overblown. And on the overall issue of patent quality, the article suggests that it may actually be on the rise.

Many critics of the PTO have claimed that the Office is simply overburdened, and cannot handle the flood of incoming applications. The authors reject this position:

Over the past five years U.S. patent filing rates grew at an average annual rate of about 6.4 percent, reaching a peak of about 9.5 percent growth in 2005 and moderating back down to about 7.1 percent growth in 2007.5 The average growth rate in patent filings over the previous five years is actually lower than the average growth rate over the previous ten years (about 7.9 percent) and is only slightly higher than the average growth rate over the previous 20 years (about 6.7 percent).

The most recent peak of 9.5 percent growth in patent filings experienced in 2005 may, at first blush, seem alarmingly high. But it is not unique or even particularly extraordinary when viewed from a historical perspective. Over the past 40 years annual growth in patent filing rates reached similar peaks in 1968 (9.1 percent), 1988-1989 (averaging 9.3 percent), 1995 (11.9 percent) and 1997-2001 (averaging 10.8 percent). Viewed within this proper historical context, the recent surge in the rate of patent filings is nary a blip in the road. Whatever implications can ultimately be drawn from the latest surge of activity, it can hardly be characterized as an overwhelming flood of crisis proportions as frequently described by some in the patent reform camp.

Also, with regard to pendency rates, the study found nothing unusual when viewed from a historical context:

While exceedingly long pendency delays would be a fair cause for concern, the current average pendency of 3.4 years is comparable to those found in other major patent offices around the world. It is also not without precedent in the U.S. For example, in 1952-another period of rapid economic and technological development-average patent pendency reached over 3.6 years. Increasing pendency delays are probably more fairly characterized as an indication of a growing patent system than a broken patent system.

As for prior art citations,

U.S. utility patents issued in 2007 actually cited significantly more prior art, on average, than patents issued five years ago, including 41 percent more U.S. patent documents, 36 percent more foreign patent documents, and 23 percent more non-patent documents.

While this significant growth in prior art citations may not be conclusive evidence of increased search thoroughness or search quality, it is certainly suggestive evidence and is plainly inconsistent with the notion that search thoroughness and search quality have somehow declined in recent years. Even common sense tells us that the most ubiquitously available search tools today (e.g. Google®) can access references that a decade ago would have remained undiscovered. The statistical data appears consistent with the conclusion that examination search thoroughness and search quality is as good or better today than it was five years ago.

Also, regarding the charge that the PTO continues to issue overly broad patents, the study, again, finds that this charge isn't necessarily warranted. Recognizing that claim breadth cannot be precisely measured mechanically or statistically, the study counted the average number of words per independent claim (since each word theoretically introduces an additional legal limitation). Here's what the study found:

[W]e find that patents issued in 2007 had an average word count per independent claim of 160.1. This reflects approximately a 4.4 percent increase over the average per claim word count of 153.3 among patents issued in 2003. As a point of reference, applications published in 2007 contained an average per independent claim word count of 111.1, indicating a substantially broader claim-scope 'ask' relative to what the Patent Office ultimately granted-patent examiners requiring 42.1 additional limiting words, on average. This latter statistic is roughly consistent with previous years' data. Taken together, the data appears consistent with the conclusion that examiners are not granting patents with broader claims, but are granting claims of approximately the same scope or slightly narrower scope (having 6.8 more limiting words on average) than five years ago.

Read/download the study here (courtesy of Joff Wild, IAM Magazine)


- In related news, IBM announced that it will join forces with the University of Tokyo, Columbia University, the University of Munich, and others to establish a technique that can be used worldwide to evaluate patent applications for problems before they are submitted. IBM intends to offer the service free of charge to corporations around the world (see press release here).

- Jon Dudas has been named one of the “Top 50 Most Influential People in IP” by Managing Intellectual Property magazine, noting that "[h]is quality initiatives have been upheld as producing ‘historic improvement’ in the quality of both patent and trademark reviews and issued patents and trademarks” (link)