USPTO Grant Rate 2021

by Dennis Crouch

The following chart provides one look at USPTO historic patent grant rate for patent applications filed over the past 20 years.  The chart groups together patent applications as of their filing-month and then simply reports the percentage patented, abandoned, and still-pending.  The red-line in the chat excludes the still pending applications and thus reports the grant rate of disposed-of applications.

A few notes.

  • There are many ways to calculate a patent grant rate.  This one is perhaps the simplest in that I simply looked at the current status of each published utility application – serial number by serial number.  I treated all utility applications equal, regardless of whether it was part of a larger patent family or whether it was subject to an RCE.
  • Beware of recent grant rate data:  In my model here, there are only two ways that a patent can escape from being still-pending: Either (1) the patent issues or (2) the applicant abandons the application.  And, the former (disposals-by-patenting) typically take less time than the latter (disposals-by-abandoning).  As the PTO begins examining a cohort of patent applications, it typically issues a number of first-action allowances, while most applicants hold on for at least a final rejection before abandoning. What all this means is that more-recent grant rate data can be skewed.

31 thoughts on “USPTO Grant Rate 2021

  1. 6

    I would love to see an expanded chart that brings the timeline back about 50 years to 1972. It would be interesting to see the waxing and waning of allowance rates over that time, with significant IP cases (e.g., KSR, Bilski, etc.) indicated along the way.

    1. 5.1

      ^

      Nice false equivalency.

      You do realize that it was despite the anti-patentists intentions that people still strive for their just protections, right?

      1. 5.1.1

        You do realize that it was despite the anti-patentists intentions that people still strive for their just protections, right?

        That’s true of every regulation that isn’t just automated registration. You could say the same thing about obviousness.

        1. 5.1.1.1

          Except, saying it here and in this context is directly on point, and refutes your false equivalency.

        2. 5.1.1.2

          let’s try again….

          Except saying it here and in this context directly refutes your false equivalency (which is rather the point, eh?)

    2. 5.2

      This graph tells the story of outcomes at the PTO rather than the suffering of patent prosecutors.

      Therefore it is wrong.

        1. 5.2.1.1

          Never stop embarrassing yourself.

          I particularly liked “Not certain how “cohorts” versus any across-the-board applications can erase” in this thread. It suggests a sad-trombone sound when one remembers how you’re supposed to have an engineering degree.

  2. 4

    OT – and just breaking: protecting (certain aspects of) software by way of copyright just took a major hit — Google wins, 6-2.

  3. 3

    link to supremecourt.gov

    JUSTICE BREYER delivered the opinion of the Court.

    Do you need me to say more? Mr. the pharaohs had abacuses that anticipate all computing devices of today delivered the opinion of the court.

    1. 3.1

      What difference does it make who got assigned to draft the Google v. Oracle copyright “fair use” decision when there were only two dissents to the decision?

    2. 3.2

      So APIs are fair game to outright copy. Wow.

      The efficient infringers are drooling.

      1. 3.2.1

        IDK. As a former software engineer, I ‘ll have to think about this more.

        But I suspect this means that any software can now be copied with a little bit of work.

        Particularly as most software is built from levels of APIs with an outer layer being what the user sees but inner layers of APIs access internal functionality. So I guess the question is whether the inner APIs are also open game.

        Pretty much this means that the only protection for software is that you can’t just copy it byte for byte but with a little work you can just copy the recreate the code. No protection. It would be like for a book that you can copy it on a copy machine but you could type it in yourself and then copy and sell it.

  4. 2

    Since this is “the current status of each published utility application” [by serial number] versus filing date, the illustrated abandonments rate is of course somewhat underrated by whatever percentage of the “abandoned” applications did actually issue as continuations, CIPs and divisionals with different serial numbers.
    The inversed-charting yellow “pending” area for older-filed applications in this chart is significant for the same filter reason – it means those same individual applications are long pending, as it is not including any continuations.
    I also would have expected a greater-than-illustrated “Dudas Dip” in grants, but assume we have no way of knowing just how many more rejections of that era were appealed or had other extended prosecutions? I believe the Board backlog was also much longer in those days?

    1. 2.2

      Paul, if you are pointing out that the deliberate “quality measures” of the Dudas Reject Reject Reject era were somewhat ameliorated by the eventual challenge of such improper examination, that point is NOT being clearly presented — nor is the historical fact that such challenge to “Office Quality” was necessitated because of the larger (and still ongoing) propaganda efforts of Efficient Infringers.

  5. 1

    I am immediately struck by the fact that this appears to rewrite history – specifically, somehow the Dudas era drop down to about 30% grant appears to have been retconned.

    That level was a historical fact — as anomalous as it was.

    How is it not shown in this graph?

    1. 1.1

      The chart here uses cohorts based upon applicant date, whereas Dudas’ treatment was applied across-the-board to all applications pending at a particular time. That said, the chart does have some curves.

      1. 1.1.1

        To be more clear, Dudas left in 2009 and so did not have the chance to really act on applications filed in 2008 or 2009. So, that is why you see the rise for those applications.

      2. 1.1.2

        Not certain how “cohorts” versus any across-the-board applications can erase what appears to be a 30% difference.

        Would be interested in the types of error bars that are induced by these differences though – I am sure that you can realize that an error bar tot account for such a large difference pretty much erases any value derivable from the “chart has some curves” take-away….

      3. 1.1.3

        “The chart here uses cohorts based upon applicant date”

        That was obvious both from the clear minimum of mid 2006 being years before the expected drop ( link to patentlyo.com ), as well as your sly reference of “Grouped by Application Filing Date” in the chart title.

        1. 1.1.3.1

          Thanks Ben — even in that linked graph, the “Dudas Drop” as a historical fact is noted — but the “picture there” only shows a drop to about 50%.

          This only exacerbates the apparent rewriting of history. We’ve gone from “showing that history” by way of “cohorting data” to draw attention from near 30% to (the linked) 50% to the present graph of 60%.

          Orwell would be both impressed and very concerned.

          1. 1.1.3.1.1

            Assuming that there was at one point an “allowance rate” of 30% (and you have provided exactly nothing showing that was ever the case), it is likely the “discrepancy” in the metric is due to how RCEs are handled.

            If you’re aggrieved that these statistics do not tell the story you want, then describe how to graph the long-term fate of applications as a function of filing date that also tells your story.

            1. 1.1.3.1.1.1

              Are you really going to try to gaslight the fact that the Dudas allowance stat was down near 30% and that the Office tried to pass this off as an example of its “quality” (in a full Reject Reject Reject mode).

              Really?

              Can you undermine your credibility any more efficiently than with a “response” like that?

              1. 1.1.3.1.1.1.1

                I haven’t said anything about the office’s description of the Dudas drop. Making up arguments for one’s interlocutors is not something which builds credibility in my book.

                Come to think of it, I wonder what audience you think you’re writing to that is impressed by your refusal to address questions or provide evidence.

                For what it’s worth, Quinn reported an allowance rate of 41%/57% for 2009 depending on how RCEs are counted.

                link to ipwatchdog.com

                Off you go. Those goalposts won’t move themselves.

                1. Editing to remove the more direct version of “not succeed” would make even George Carlin blush.

Comments are closed.