Saturday, July 18, 2009

Patent pendency

On July 17, 2009, Patently-O discussed results on a database of 20,000+ randomly selected patent applications and issued patents filed 2000 - 2007. For each of case, [Dennis] tallied the final status as of mid-July 2009. (Status is either patented, abandoned, or pending). Using that data, I created a chart of the percent of utility applications still pending as of July 2009. Presumably, this means that there was a database of 20,000 applications selected from the years 2000 to 2007, and a determination was made of "how many" were NOT issued or abandoned. Although the analysis was stated "To provide some measure of pendency," such an analysis would NOT address the length of pendency, because it includes applications still pending (pendency duration unknown), whose "pendency" would be determined under PRESENT (and future) conditions, which conditions are NOT the same as PAST conditions. Specifically, the analysis is a "snapshot" in time, which may reflect neither the past nor the future. To the extent it includes disposals (issued or abandoned applications, as in 50% of applications disposed of in 3.5 years), SOME information about pendency in the past is within the analysis. An issue arose about pendency (past vs. current), brought up in a later post "Patent Backlog."

Specifically, a later post, titled Patent Backlog involved a comment by "longtime partner", who noted that the data in the 20,000 application database showed 50% of the applications were disposed of within 3.5 years and this was not much different than in 1990.

["longtime partner" had written-->]I suppose I’m something of a lone wolf in thinking that the backlog in applications we have now is not much worse (certainly not DRAMATICALLY) worse than the backlog we’ve had all along, or at least since I’ve been in patent law. Your chart shows that 50% of application were disposed of in about 3.5 years and about 90% in roughly 6 years. When I was an examiner, back in 1990 or thereabouts, there was a lot of verbiage given to getting average pendency to 3 years – so it must have been higher than that at the time.

A first commenter to "longtime partner" challenged the facts about 1990: In 1990 or thereabouts the average pendency was 18.9 months, not somewhere over 36 months.

Ron Katznelson noted:

The “longtime partner at a midsized IP boutique in DC” commenting above is not only wrong by two years as pointed above, but also by comparing apples to oranges. Your “Percent Application Still Pending” distribution cannot provide the average pendency of granted patents. It can only provide a poor LOWER BOUND for average pendency because the contribution to the average by the pending applications is unknown and is GROWING each day they do not issue. Therefore, despite its “predictive” nature, your chart is a rather charitable characterization of the growing problems at the PTO.

Using actual facts for comparing apples to apples to dispel Malcolm Mooney’s assertions about “whining”, it would be useful to look at actual pendency of all patents granted this month, as you have done for patents granted in January-February 2007 at .
If you show these two distributions together, the shift to the right in the recent distribution would paint the real story. It will also show that the increase in average pendency of issued patents over the last 2.5 years far exceeds the pendency increase reported by the PTO. This is because the PTO’s average includes non-grant disposals and because the PTO is pushing to accelerate such disposals (- when have you last seen an “after final” action?), in part, to manipulate downwards the reported pendency.

Ron also commented to earlier post Patent Application Pendency: Percent of Applications Still Pending, which contained the data on the 20,000 applications:

The same argument applies for Continuations that, unlike RCE’s, are not a repeated attempt at the parent’s claims. In contrast, RCE’s are by definition a repeated attempt at earlier-filed claims and therefore their pendency should be (and indeed is) counted from the date of the earlier application. RCE’s do not receive a new application serial number and therefore the charts that Dennis posted correctly reflect pendency based on the proper serialized application date.

IPBiz notes that Quillen and Webster assumed the opposite, that continuations were a repeated attempt at the parent applications claim, one reason that the Quillen and Webster analysis is wrong.


Ron Katznelson and I just went over the latest data from the Official Gazette on the pendency to first office action of ecommerce patent applications.

It looks like the USPTO has substantially increased the rate of first office actions in the past year and pendency to first office action may have leveled off at 4 years.

In all fairness then, I would now expect first office actions in 2013, not 2016, for e commerce applications filed today. If they can keep the high rate of first office actions up, then the delay should come down even more


Post a Comment

<< Home