Right Sized Patents

Dennis Crouch

Many progressive policies focus on reducing disparities (income, wealth, education, and opportunities) that reflect some social injustice between those at the top and those at the bottom of our social spectrum. Conservatives often recognize the gaps but disagree about whether the result qualifies as injustice as well as about government’s role in redistribution.

Patent policy is often easier to implement than social policy (especially compared with other property law changes) because a new generation of patents emerges every twenty years and the old generation does not hang-around protecting and directing wealth but instead melds into the Soylent of the public domain.

In some ways though, patents are bucking the social trend and becoming more standardized and less diverse – at least by some outward measurements such as document size, claims per patent, and prosecution pendency.

Standard deviation is the most common measure of spread or dispersion of a statistic across a population. A population with a wider spread and more outliers will have a greater standard deviation.  For patent claims, this measure of spread has been dropping steadily for the past decade.  The first chart below shows the standard deviation of patent claim count for issued US utility patents grouped by year of issuance.


The reduction in variance has been accompanied by a drop in the average number of claims per utility patent as shown in the next chart (note the axis break).


My working hypothesis is that there are three primary drivers for the drop in claims-per-patent and all three also serve to (partially) explain the drop in spread.  These drivers: (1) a substantial increase in fees for filing more than 20 claims-per-patent that was put into place in 2004 and has continued to rise; (2) a suggested (but not confirmed) rise in restriction requirements for patents with more than 20 claims; and (3) a decrease in patent valuation leaving patentees less willing to support the higher cost of a large number of claims.  (As an aside, reasons (1) and (2) may also helping to dive the continued rise in re-filing.)

At the same time, however, another interesting trend is reducing the spread: there are fewer “small” patents with only 1-5 claims. Thus, the trend is a reduction of both small and mega patents (as measured by claim count) with a proportional increase in the middle with ‘average sized patents’.   The final chart below provides some further context to these statements.


The time series chart groups each year’s patents into a set of bands according to the number of issued claims.  As a savvy chart reader could surmise, the percentage of patents with 1-5 claims has dropped since 2002 as has the percentage of patents with >30 claims.  Those reductions were accompanied by a proportional rise in patents with 16-20 claims.  Time will tell whether this shift also leads to a more just society.

42 thoughts on “Right Sized Patents

  1. 15

    By “patents are bucking the social trend” you mean that income and wealth inequality (among humans) is growing while claim inequality (among patents) is shrinking? How does average claims per patent have anything to do with income inequality or a “more just society?” Do you actually see a connection? I don’t.

    1. 15.1


      I concur – hence the poke at 6.

      W A A A Y too much “guarding the fields of rye” mentality in play….

    2. 15.3

      Another thing about this blog post is that it fails to ask the question if the higher fees on claims has resulted in more patent applications being split or the same disclosure being filed with different sets of claims (because it has gotten so expensive to add a few extra claims.)

  2. 14

    From the point of view of social justice, I think your results are indicative of what is likely a disturbing shutting out of the little guy from the patent process for the following reasons (although clearly there is nothing in your data that conclusively shows that the little guy has been shut out). If there are fewer applications with more than 20 claims, then even the big corporations are finding those applications too expensive, and so certainly applications with more than 20 claims will be too expensive for the little guy. The smaller number of applications with only a few claims is likely indicative of the uncertainties in our patent law, and that having only a few claims is likely considered not enough claims by most patent practitioners, which raises the cost of patents, and the higher cost is likely significant in a negative way to the little guy.

    Most of the changes is patent law in the past decade are likely not favorable to the little guy (e.g., the filing fees alone likely make IPRs and similar post issuance proceedings unaffordable for small poorly funded entites). I am not sure what a good metric is for social justice in a patent system, but possibly one metric is the number/percentage of patents filed by individuals and/or legal entities that were poorly funded at the time of filing. Possible another metric is number/percentage of entites/individuals that were poorly funded at the time of filing, but that were able to successfully monetize their patents/inventions as compared to the extent to which well funded entities were able to successfully monetize their inventions (after factoring out various other effects, such as biases against nonpracticing entities).

    Best Regards,
    David Lewis

  3. 12

    Yes, but what about the size of the disclosure? From what I’ve seen because of the cost of the claims, some clients are still stuffing massive disclosures into the application but only claiming a portion of it. The reason is that you need to keep these things going forever because of the IPRs, so keep it going and claim different portions as needed.

    1. 12.1

      This is what I see as well. IPRs aren’t necessarily why it’s good to keep an application family alive. The ability to draft narrow claims that map directly to a competitor’s product with that product sitting in front of you is the most likely driver of this continuation strategy.

  4. 11

    ” Time will tell whether this shift also leads to a more just society.”

    You should ask the SJW’s that are busy trying to wreck your fine institution. I’m sure they know already whether it will or not lol.

  5. 10

    16 does seem like a fairly decent number of claims in an issued patent. If you file 20 and a few dependent claims are incorporated into independent claims, and you filed a comprehensive claim set to start with. Once you get allowable subject matter anyone with a brain can pile on cumulative claims, but why?

    Filing more than 20 is silly anyways, you will not get a quality examination of all the claims. You would likely get restricted if the examiner is in certain art units.

    If you have multiple (legitimate) inventions and want a restriction so you don’t get a double patenting rejection, then you should just clearly put them into separate independent claims then build out the first application and the divisional.

    1. 10.1

      Filing more than 20 is silly anyways

      True dat.

      And then there’s this:

      link to patentlyo.com

      [T]he 80 pending applications at issue represent roughly one-fifth of 399 applications [Gil Hyatt] has pending before the USPTO, which contain an estimated total of 115,000 claims

    2. 10.2

      Re: divisional / double patenting – I don’t think it matters either way. Terminal disclaimers aren’t a big deal.

  6. 9

    Reason #4: Valuation has become more closely tied to the significance of the claims than the sheer number.

    I’ve lost count of the number of cases that inflate claims with limitations that are substantively meaningless, like:

    > 2. The method of claim 1, wherein the computer further comprises a server.

    > 3. The method of claim 2, wherein the server further comprises a webserver.

    > 4. The method of claim 3, wherein the webserver is configured to transmit a web page in response to a request.

    > 5. The method of claim 4, wherein the web page is formatted according to HTML.

    > 6. The method of claim 4, wherein the web page is transmitted using HTTP.

    > 7. The method of claim 2, wherein the server comprises an email server…

    On and on and on. Lots of bulk – no meaningful improvement in claim scope, clarity, or enforceability.

    The message has clearly been received that not all patents are alike, and with it have come refinements to drafting and prosecution practices that reduce these inefficiencies.

    1. 9.1

      David, people who add dependent claims that only add prior art context, not invention, either do not know what they are doing, or do not have their client’s best interests in mind.

      1. 9.1.1

        I agree with you Ned.

        I was taught that you drafted the application without regard to claim count on the first pass.

        Any “fitting” of claims into certain numbers was then a valuation exercise based on a multitude of factors, including, but certainly not limited to, a budget for fees.

        Sadly, I think that it became a “1azy” standard practice to just “take the deal” of a standard 3/20 claim set.

  7. 8

    OT but there was interesting exchange starting around the 26:00 point in today’s oral argument in Software Rights Archive, LLC v. Facebook, Inc. (2015-1648). Link to audio here: link to cafc.uscourts.gov

    Judges: Newman, Mayer, & Chen

    Specifically, the issue being discussed at that point in the proceedings was whether the CAFC can sua sponte raise a 101 rejection that wasn’t briefed, at least in the form of an advisory opinion before sending a case back down to the Board. The “innovation” in question was using “indirect relationships” in a search in combination with “direct relationships” … on a computer (Mayer: “This is clearly a 101.”)

    Not sure how it’ll turn out but if that’s the kind of thing that would make your head explode …

    1. 8.1

      MM, this is a IPR. 101 is not a basis for invalidating claims in a IPR.

      35 U.S.C. 311(b) (b)
      SCOPE.—A petitioner in an inter partes review may request to cancel as unpatentable 1 or more claims of a patent only on a ground that could be raised under section 102 or 103 and only on the basis of prior art consisting of patents or printed publications.

      1. 8.1.2

        Ned: MM, this is a IPR. 101 is not a basis for invalidating claims in a IPR.

        It’s true that the petitioner can’t raise 101 as a basis for invalidation. But that does mean that the PTO is forbidden from raising the issue on its own? Even when the CAFC is advising the PTO that the claim is ineligible?

        The PTO is entitled already — presumably — to consider subject matter eligibility issues when those issues are presented as falling within the so-called “printed matter doctrine.” And it certainly screens proposed amendments for compliance with 112 so that’s in the mix already as well.

        I guess the more important lesson is that IPRs definitely need to be expanded to include 101 issues. As observed correctly in the oral argument, it’s a huge waste of resources for the CAFC and everyone else to dance around the gaping eligibility problems that plague a lot of the claims subjected to IPR.

        Everyone should listen sometime to these judges trying to figure out — based on conflicting testimony — what a claimed algorithm is “actually” describing and what’s the “inventive feature” of that claimed algorithm (i.e., what’s the “new” logical step). It’s a farce.


          MM – in an adversarial setting, isn’t there always going to be conflicting testimony? I don’t think that this conflict necessarily means there is a farce. That’s the nature of weighing information.


            I don’t think that this conflict necessarily means there is a farce.

            I agree with you on that count.

            The farcical aspect arises from the abstract nature of the subject matter being discussed (logic — to be construed by applying logic) and the inevitable desire of one of the two sides to introduce confusion and complication, rather than eliminate it. Typically it’s the patentee engaging in the latter activity, and it usually begins with the unnecessarily jargon-laden claims themselves.


          MM, there is no doubt that some in the Federal Circuit believe the PTAB can find the patent before it invalid on any basis it wants to. But this is not an examination where the PTAB can suggest to the examiner a new grounds of rejection and remand back. They can only hold a claim invalid if they find that the petitioner proved invalidity by a preponderance of evidence. They simply cannot assume a petition was filed with different evidence and different arguments, and then decide the case upon the hypothetical case not presented – at least that is the issue for Supreme Court right now Cuozzo.

          Think about this. The petitioner presents his case. The patent owner responds. And then the PTAB issues a final written decision deciding the case not on the evidence presented, not on the arguments presented, not on the grounds asserted in the petition, not on the grounds even that the statue says are the only grounds upon which a decision can be made, but goes off and holds the claims invalid under 101 without giving the patent owner even an opportunity to present a case in opposition.

          Beyond being nonstatutory, such a situation would deny the patent owner essential aspects of due process. But trust me, given Cuozzo, and given some of the things that the Federal Circuit has said in cases like this, the Federal Circuit does not see anything wrong at all about ignoring the rights of the patent owner, or the statute, or simple decency.


            Ned: goes off and holds the claims invalid under 101 without giving the patent owner even an opportunity to present a case in opposition.

            I don’t think it needs to happen that way. The patent owner could be given a chance to response.

            Note that the CAFC wouldn’t consider doing this if it felt it was a great reach. My impression is that there was no reaching at all. Like a lot of junky claims, the claims are drafted broadly to cover the use of basic logic applied to identify the best “hit” or “hits” as applied to a handful of subject documents. They aren’t limited to searching a massive database or library.

            So the applicant will be given a “chance” to respond. The problem is, based on the admitted prior art, there is no chance of winning.

            Why waste everyone’s time?


              MM, what you are thinking about here is a procedure that works in court or in ordinary ex parte prosecution, but not in IPRs because of its rigid, statutorily imposed structure.

              As well, consider the consequences if the PTAB could hold the patent invalid under any statutory basis, and perhaps on evidence or argument not even submitted in the petition. The petitioner who loses, having presented just a couple of references and a single theory, would be totally estopped in court to raise 101, 112, or use other art, or the same art in a way not presented in the IPR.

  8. 6

    Talk about chasing windmills…

    To me will tell whether this shift also leads to a more just society.

    How the H is rewarding and promoting innovation supposed to be tied to this vague and fuzzy notion of a more just society?

    This reminds me of “not keeping score” and “participation ribbons” and the houy and nonsense of political correctness.

  9. 5

    Putting my Jean-Jacques Rousseau hat on, even if every citizen had the same wealth/income, there still would be corporations with greater power. Crony capitalism would continue to exist, would it not? The best solution is therefor to create powers in opposition, competition.

    Putting my Orwell hat on, might I observe also that some animals are always more equal than others.

    Putting my Teddy R./Lord Coke hat on, might I say that monopolies are bad from any point of view. That is why patents that take from the public the privileges they previously had and put them into the hands of the one must be illegal.

    And, speaking of abusive monopolies, consider the Federal Circuit.

    1. 5.1

      Putting on another hat you might say:

      Socialists often recognize that government force to reduce disparity gaps turns everyone into either a thief or a victim, but disagree about what euphemism to use and whether the result reflects the natural state of men towards one another.


          Interesting thought Ned.

          Here’s a follow-on project:

          Chart claims per original submitted specification (fold into the parent all child Con and Div claims).

          This would eliminate much (how much is uncertain) of the lower bands that make up to 30% of the graph – and possibly up to 50%, depending on how the Div’s shake out.

          When up to 50% of the data displayed is possibly “hiding” some other mechanism, I tend to think of the graph as not all that compelling.

  10. 4

    Higher percentage of foreign origin applications and foreign origin applications have fewer claims?

  11. 3

    Is it April 1st? That was my first thought on reading (mid evening, after one or two pre-prandial tinctures) the last sentence from Dennis.

    I mean, honestly. People (including patent attorneys) think what they are taught to think. What they are taught to think, these days, is that 20 claims is enough.

    Besides, which of you doesn’t begrudge paying even one penny in PTO claims fees? Regardless of the legal merits of the case, those from the USA who instruct me to file at the EPO want the claims brought down to 15, to deprive the EPO of any claim fee income whatsoever. So, I guess, when filing at the USPTO, the first priority is to ensure that not one penny of client’s money gets paid to the USPTO in claims fees.

    I mean, amongst patent attorneys, which one of you wants to be thought eccentric, to be obtuse, to be weird? Better to be a sheep, eh?

    Dennis, when you wrote your last sentence, what substance had you just finished absorbing, I wonder.

  12. 2

    Off topic: Anyone have their PKI certificate revoked by the PTO? We’ve had two of them revoked even though we use them all the time. Apparently, the PTO is revoking them after 27 months even if they’re being used. And, they give you no warning whatsoever — no email, no call, nothing. They just revoke the certificate. You have to reapply, but that takes a week or so.

    Who at the PTO thought this was a good idea?

    1. 2.1

      Hi PB – you may not have to reapply for a revoked PKI. I think that they may simply revoke a given instance of a certificate. S0me form of revocation has happened to me, and I tried other instances of the certificate to find one that worked. It’s also possible that we are experiencing different forms of revocation.

  13. 1

    You might investigate whether there is a correlation between whether the US case claims foreign priority. E.g., whether cases with less than 10 claims have a Japanese filed parent that the US claims priority to. I have seen it be common practice for JP companies that file in the US to file less than 20 claims (despite getting 20 for free) likely because they did not want to go through the hassle or expense of beefing up their claim set coming from JP. If there was a precipitous drop off in these companies filing in the US, then that may explain the overall trend of a reduction in the number of claims.

    1. 1.1

      The same could be said for the EPO, where >15 claims entails a (much) higher fee.

    2. 1.2

      These are good points and I’ll look at the data. Interesting thing with US cases is that the growth over the past decade is in patents with 16-20 claims. The Japan/Europe theories involve <16 claims and so wouldn't explain the data shown in the third chart. One reality is that the limits/costs on claim numbers has been in place in Europe and Japan for this entire period and thus are largely a constant except insofar as the proportion of US cases originating from those locals has increased.

      1. 1.2.1

        Agreed DC. I think 16-20 can be explained by looking at the top 10 – 20 filers, and noting that their internal policies require a 20 claim set – 3/17. Then, during prosecution, a limitation from a dependent claim is incorporated into each independent claim as a compromise – typically reducing the quantity by 2 – 3 claims. Often, the big corps require a strategically narrow independent claim at the time of filing – this means that perhaps 2 independent claims lack the narrow limitation, requiring the cancellation of 2 dependent claims – leaving 18 claims.

Comments are closed.