Probing 10 Patent Troll Myths – A Factured Fairytale Part 2

Editor’s Note: This is part 2 of a 5 part series written by Steven J. Moore and with the assistance of Marvin Wachs and Timothy Moore, also of the Kelley Drye & Warren Patent Department. Part 1 was A Fractured Fairy Tale: Separating Fact & Fiction on Patent Trolls.

For all of us mystery buffs, what makes a good story is learning at the end who is ulti-mately responsible for the skullduggery set forth at the beginning of the yarn. Were all the prob-lems due to Colonel Mustard and his ineptitude with the candlestick, or do we have another case of “the butler did it?” And so, we set up our response to the many fables swirling around the so-called “Patent Troll” to lead us to whom we now believe is ultimately responsible for the rise in litigation by such entities. In this regard, we will look at numerous statements that have been asserted in the past with respect to the “patent troll,” and see how such statements stack up against the data. By following our responses to the various “myths” of the patent troll, we be-lieve most will themselves detect the “power behind the troll” long before we specifically un-cloak the same (no, it is not the Romulans! ) But just in case, we do the unmasking at the end of the paper.

There are many myths that are attached to the fable of the so-called “patent troll.”  Acting like the MythBusters, we probed some of them, and set forth our findings below.

Myth 1: NPEs assert poor patents.

  • Truth:The success rate of NPEs overall across 267 random cases indicates that the litigation outcome for NPE suits looks very similar to litigation outcome for Producer suits.  However, when individual inventor suits and individual inventor company suits were removed from the mix of overall NPEs, we found that non-independent inventor NPEs had an outcome profile that looked significantly  better than the Producers, both in very likely favorable and likely unfavorable outcomes in litigation.

Congress and the White House have apparently found common ground, and each agrees that Trolls assert poor patents “that leverage and hijack somebody else’s idea.”  See, Fitzpatrick, Alex, Obama:  Patent Trolls Hijack Others’ Ideas to ‘Extort Some Money, Mashable, see also Press Release, Senator Charles E. Schumer, Patent Troll Companies Make Money By Purchasing Broad Patents & Suing Legitimate Tech Companies With Similar Products (May 2, 2013), available at.  Genesis for the proposition that NPEs assert poor patents come from a number of sources including members of, and former members of, Mark Lemley’s Stanford group.  See, e.g, John R. Allison, Mark A. Lemley & Joshua Walker, Patent Quality and Settlement Among Repeat Patent Litigants, 99 Geo. L.J. 677-694 (2011) wherein it is maintained that victories for patent trolls were rare based on their own internal coding.

We used the Lexis Total Patent database to determine the outcomes for each of the 267 cases filed between 2004 and 2006 (pre-AIA) that we reviewed.  While we thought such inputs were very likely to be unbiased (as there surely was no impetus for Lexis employees to shade any result), we checked many of these codings against Pacer dockets as an integrity check of the Lexis input to assure that there were no overt anomalies.  We also checked the integrity of Lexis’ Reexamination summary database, and consulted with the USPTO reexamination records to the extent needed to understand what went on during a reexamination.


We added to the Lexis characterizations of favorable and unfavorable results certain dismissals that occurred shortly after a favorable or unfavorable final reexamination finding in respect to the patent asserted in the case.  We characterized as a favorable reexamination finding one where the USPTO found one or more (or all) claims asserted in the case against the defendant(s) as being valid, while we characterized all other such dismissals settling after an adverse finding for the patentee (i.e. no claim asserted in the complaint having been maintained as valid as drafted) as likely unfavorable.  Any dismissal wherein it did not seem clear what had happened in the case, we characterized as a dismissal wherein the final resolution of the case was unclear. See Fig. 1: Litigation Profile of Producers v. NPE and Fig. 2: Litigation Profile of Global 2000+ Producers.

In terms of the profile of Producers vs. NPEs overall, as shown in the pie charts at Fig. 1 we saw little difference between the outcome profile of Producers and the NPEs (independent inventors and independent inventor companies included).  That is, NPEs showed nearly the same profile of outcomes as Producers.  In terms of overall NPEs, we noted 18% of cases brought by this group to be favorable or very likely favorable for plaintiffs, 4% of all cases brought by this group to be unfavorable to or very likely unfavorable to them, and the remaining 78% of cases being dismissed in circumstances wherein the resolution was unclear.  In terms of Producers,  we noted 20% of all cases having a clearly defined favorable or very likely favorable outcome for plaintiffs, 9% of all case outcomes being unfavorable or very likely unfavorable to them, and the remaining 71% of cases being dismissals wherein the ultimate outcome of the case was unclear.  Of the 34 cases which we identified as being brought by Global 2000+ producer companies on patents held by the same, we also saw a similar profile, with an increase in the percent of cases that were identifiable as likely unfavorable to the plaintiff (Fig. 2).

However, when we looked at cases brought by NPE plaintiffs which were neither independent inventors nor independent-inventor companies, we saw a very different profile, which is contrary to common belief.  We found nearly 33% of the time the outcome of the case was favorable, or very likely favorable for plaintiffs, with only 7% of the outcomes for the same group being unfavorable or very likely unfavorable, with 60% of the remaining cases being dismissals wherein the ultimate outcome of the case was not clear (Fig. 3).  The latter suggests that this subgroup of NPE Plaintiffs actually was doing better than Producer plaintiffs were doing when asserting their own patents. See Fig 3: Litigation Outcome Profile of NPEs Who are Neither Independent Inventors or Independent Inventor Companies.

We note that our findings are limited by the fact that we do not know the ultimate resolution of cases based on simple dismissals.  Given that we could find no other basis for dismissal, we surmise that that in terms of favorability outcome was likely the same for each group although this is not certain.

Interestingly, we later discovered that our findings complement closely those of Professor Risch of Villanova law school who looked extensively at 350 patents of well-known NPE patent acquisition companies, and Timo Fisher and Joachim Henkel who compared 565 of such patents against 1,130 patents acquired by practicing firms, both of which parties weighted a patents’ worth based on parameters well accepted for the valuation of patents.  See, Michael Risch, Patent Troll Myths, 42 Seton Hall L. Rev. 457 (2012); see also Timo Fisher and Joachim Henkel, Patent Trolls on Markets for Technology – An Empirical Analysis of Trolls’ Patent Acquisition, (working paper, 2010).  Each of these groups found the patent quality of NPE patents to be at least of the same quality of Producers, if not better.  Furthermore in looking at those patents that were actually litigated to completion, Professor Risch noted that he also did not find the patents of these well-known NPEs to fare demonstrably worse than other litigated patents.

Myth 2: The majority of patent litigation cases filed after the enactment of the AIA are filed by NPEs.

  • Truth: A review of 425 random cases filed after the enactment of AIA (and thus under the new joinder rules) do not confirm that the majority of the patent litigation cases are filed by entities that do not sell product or services to the public (NPEs) even interpreting the data in accord with the new joinder rules.  However, if one includes all independent inventor/independent inventor companies into the mix of NPEs about 45% are filed by NPEs.

The press is replete with statements to the effect that the lawsuits filed by “so-called patent trolls now make up the majority of all patent litigation in the United States, marking a sharp rise” from the past.  See April 9, 2013 Law 360: “Trolls File the Majority of Patent Litigation, Study Finds.”   These “factoids” seem to have derived from two different studies, each of which used entirely different criteria for coding the so-called “bad-guy.”

The Jeruss study (Robin Feldman, Tom Ewing & Sara Jeruss, The AIA 500 Expanded: The Effects of Patent Monetization Entities, (UCLA Journal of Law & Technology, Research Paper No. 45), available at defines the “bad guy” not as a “troll,” or NPE, but rather as a “Patent Monetizing Entity” which is defined as “one whose primary focus is deriving income from licensing and litigation as opposed to making products.”  The Jeruss study suggests that such “PMEs” now account for 51 to 56% of all patent litigations filed.  A number of authors of the Jeruss study have positions at Lex Machina, a spin-off from Mark Lemley’s Stanford IP Litigation Clearinghouse project.  Professor Lemley, of course, is one of the most outspoken opponents of the “patent troll,” and the “large number” of purportedly “bad patents” being issued by the USPTO.


The other entity reporting such high numbers is RPX Corporation, a company which sells protection to other companies against the so-called Trolls.  This study defines the “bad guy” as the “Patent Assertion Entity,” which is defined as an entity that asserts patents on existing products “as a business model.”  The RPX study states that, in 2012, 62% of patent litigations were filed by “Patent Assertion Entities.”  See, Colleen V. Chien, Patent Trolls by the Numbers (Santa Clara University Law School, Research Paper, 2013) available at =2233041 (Professor Chien, another outspoken critic of trolls, also worked with the Lemley’s Stanford Litigation Database group in the past).   

We believe our coding protocol allowed very little in terms of judgment calls.  Unlike some studies, we did not limit the Producer to only companies that produce actual product to the public, but also included as Producers those entities that actually sell services to the public.  Given that it is now reported that most corporate entities in the United States today are service companies, this seemed to be entirely rationale.  We classified any entity which was truly selling either a product OR a service to the public as a Producer, as well as universities, research institutes, and emerging companies.  All other entities were classified as non-Producers.

No matter how hard we tried, we could not figure a way to obtain a majority of cases being filed by entities that neither sold products or services to the public, which we reference herein as NPEs.  In fact, we believe that our  assignment to the NPE category is actually inflated, as we did not count certain entities as Producers, even if they were selling products and services to the public, if it appeared from searches that such commercial activity might be solely a by-product of a litigation strategy (we based our decision on the number of cases they filed, the number of employees working for the entity, and when the producing activities suddenly sprung up).

Our study of 425 random cases found no more than 46% of all litigations are currently filed by NPEs when counting each case against each defendant as a distinct case, as required under the AIA (Fig. 4).  The difference between our results and the results of Jeruss et al. and the RPX studies may relate to a number of factors.  First, we did not include suits by universities or emerging companies (such as most biotechnology companies) as NPE suits.   Rather we grouped them into the “Producing” category where we believed they belong.  Second, the Jeruss and RPX studies did not look at the distinction between NPEs and Producers per se, but were based on completely new “boogey men” the so-called “patent monetization entity” and the “patent assertion entity”.  See Robin Feldman, Tom Ewing & Sara Jeruss, The AIA 500 Expanded: The Effects of Patent Monetization Entities. In our view these new “bad guy” definitions allowed for more subjective categorization permitting greater latitude in assignment to the “bad guy” group.   Our study shows, even with what may be called our inflated numbers for the NPE category, that less than a majority of cases filed after the AIA passage are being filed by NPEs, even giving credit to the inflation of law suit cases caused by the new joinder rules. See Fig. 4: Distribution of Producer and NPE Named Plaintiffs in 428 Post-AIA Joinder-Limited Litigations.

We note that in respect of the NPE figure which we determined for cases filed between 2006 and 2008 (Fig. 5). our results match closely that of Professor Chien’s 17% figure reported in 2009.  See, Colleen V. Chien, Of Trolls, Davids, Goliaths, and Kings:  Narratives and Evidence in the Litigation of High-Tech Patents, 87N. C. L.  Rev., 1571 – 1651 (2009).  The latter suggests that our coding was at least in conformity with this group at the time. See Fig. 5: Distribution of Producer and NPE Named Plaintiffs in 267 Pre-AIA Litigations.



Warning & Disclaimer: The pages, articles and comments on do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of

Join the Discussion

6 comments so far.

  • [Avatar for Michael Risch]
    Michael Risch
    August 9, 2013 09:42 am

    I agree that John Smith’s comments don’t really change the study much.If you have to think hard about whether Walmart or a web developer is an NPE or not, you’re just not doing the same study that everyone else is doing. That’s OK, of course, but the purpose of this study was to compare with others.

    I don’t think you’re classifying plaintiffs much differently than Feldman, Ewing, & Jeruss. That said, it is unclear how each of you would classify R&D design shops like InterDigital or product company IP arms.

    That said, I have a more fundamental concern. You have a random sample, and they look at all the cases. Thus, you can have sampling error and they won’t. What’s your standard error with that sample size? You look post AIA, they look at 2011 and 2012. Why does that matter? There was a serious spike just prior to the AIA, and then a lull right after. If you look only post AIA, your percentages will obviously be lower. Even more important, you are combining two years, so a low 2011 will drag down a high 2012. What’s your 2012 number? That’s where the 58%+ number comes from – not all post AIA, but just 2012, and there’s no real dispute that the share is growing. Just some thoughts to help do an apples to apples comparison.

  • [Avatar for Steve Moore]
    Steve Moore
    August 2, 2013 10:11 pm

    There simply weren’t many close calls and the samples were random. It is easy to hypothesize problems. I suggest you try it yourself. Nothing is stopping you but time and money. I note every study deals with this issue. The fact that our finding in regard to the percent of NPEs pre-AIA matches very closely with that of prior studies looking at the percent of NPEs in the same time frame suggests that are classifications are very much in line with the authors of these prior studies. As a former PhD researcher I am used to criticisms in regard to research studies. However I would appreciate if you laid off ad hominem sarcasm which lacks any factual basis.

  • [Avatar for John Smith]
    John Smith
    August 1, 2013 01:34 pm

    “Would you want us to call Wal-Mart a NPE? ”

    A great question, and I would suppose that it depends on what the patent being asserted was, or on Walmart’s overall patent assertion behavior (which I haven’t the foggiest notion about). That determination would require facts beyond those which are at hand. And without this determination being made with a more reliable, more transparent, method I’m sorry but I don’t see your article gaining traction with the people who it needs to.

    “Website design companies, are selling a service to the public not an physical product. Telecommunication service providers provide a service, not an tangible product to a consumer. ”

    See, and there again, website design companies are questionable, or would be to many people. Many folks just don’t see website design as a field with much patentable/eligible (thus leading to the questionability of quality prong of the “troll” determination for many, as opposed to the NPE determination) from the get go and then they see many firms with many more patents than they practice, or with patents with a scope way beyond what the company itself is doing. Once they see one or more of those things they may thus be led to believe that the company is not truly a practicing entitiy for the purpose of running these numbers. Though I perfectly understand why you lump them in.

    Again, I understand why this was a tough call for you to make, and why you had to make it. At least we are agreed on whether or not this characterization holds the power to change the results of your study enormously. Finally I must say it is surprising to me that you didn’t have any “close calls” in your making the determination in your random sample. That might also raise some suspicions to some folks as to just how random, or how wide the group to pick samples from, truly was. Not as to malicious intent on your part, just problems in the method utilized to randomly select.

    In any event, at least you tried, so that is something. And don’t let these difficulties in doing a study get you down, keep up the good work.

  • [Avatar for Anon]
    July 31, 2013 02:43 pm


    From past exchanges, your views on what is ‘appropriate’ are entirely without merit.

  • [Avatar for Steve Moore]
    Steve Moore
    July 31, 2013 02:16 pm

    You might have a point if we were faced with such a situation. We were not. The services we are talking about are all from actual companies whose primary business is not patent licensing. There were no law firms in our random sample. For example, retailers who sell products are providing a service to the public, not actually producing a product. Would you want us to call Wal-Mart a NPE? Website design companies, are selling a service to the public not an physical product. Telecommunication service providers provide a service, not an tangible product to a consumer. These were all very legitimate calls and we did not have a single instance in our sample of the examples you cite — that is, there was no shading to call Intellectual Ventures a service company, or Wall Street financial firms as service companies.

  • [Avatar for John Smith]
    John Smith
    July 31, 2013 01:03 pm

    “Unlike some studies, we did not limit the Producer to only companies that produce actual product to the public, but also included as Producers those entities that actually sell services to the public”

    I understand the need to make this decision as to whether or not service providers will be counted along with the producers as a producer or not, however…

    “Given that it is now reported that most corporate entities in the United States today are service companies, this seemed to be entirely rationale. ”

    Simply lumping them in regardless of the “service” being “sold” does not seem appropriate. I’m not really sure that your decision to lump them in in total actually is “rational”, and this decision it will taint all of your results compared to what normal people would conclude. For instance, does a law firm that decides to go trolling count as a “producer”? How about intellectual ventures? How about wall street financial firms? You seem to have skewed the results of your entire study by making this one huge decision to simply lump many of the entities engaging in the behavior people find objectionable to actually be on the side of the white knights. And by having done so you seem to have done the rest of the study for nothing.