My study of non-practicing patent plaintiffs was sparked by a discussion with a colleague about where patent trolls come from. My theory was that patent trolls tended to enforce patents that startups obtained, but that lay fallow when the startups lost funding. Unfortunately, I had no data to back up my intuition, nor did anyone else. So, I thought I would gather that data.
Thus began two years of data gathering. I started with the top 10 most litigious NPE’s (since 2003), as reported by Patent Freedom. With some help from Patent Freedom, I then found every case I could involving these NPEs (1011 in total) and then found every patent asserted in those cases (400 total). I then drilled down, recording information not only the patents, but who obtained them. Finally, I gathered information about the organizations that obtained patents (121 total) .
Below is a summary of some interesting things I found. There’s a lot more in the draft, of course.
First, contrary to popular belief, not all NPE patents are business methods patents. Indeed, only 32 of the 400 patents (8%) included class 705 (the patent class most associated with business methods). Another 88 patents are in patent classes usually associated with software, for a total of 31% for both business method and other software patent classes. This probably overstates the number, because patents may be assigned to more than one class.
Of course, there may be others that are not classified as software, but certainly not the other 69%. Classifcations included telephonic communications, television, video distribution, and computer hardware. Many of the patents, unsurprisingly, related to communications — a field of growing importance.
Second, the study finds that NPE patents do not overwhelmingly come from non-productive companies. Of the 400 patents, 286 were initially assigned to a company; there were a total of 121 unique companies listed as initial assignees on these 286 patents. More than 75% were assigned to corporations while the remainder were assigned to LLC’s and limited partnerships. Another five patents were initially assigned to three other entities: the U.S. government, a hospital, and a university.
Of the 121 entities classified, only 21, or 17%, were solely licensing entities at the time of patenting, accounting for 30 of the patents. In other words, a small portion of the patents were obtained by the purest of patent trolls. Another 50 patents were issued to inventor-owned companies. The remaining assignees (including some of the inventor-owned companies) either had or were attempting to build product or service based businesses at the time of the patent. There were many familiar names among these companies, including IBM, Xerox, AT&T, Motorola, Proctor & Gamble, Siemens, Harris Corporation, California Technical Institute, and Cedars-Sinai Hospital. Sales, employment, and venture capital data all imply that at least half of the companies contributing NPE patents had significant operations. Also contrary to conventional wisdom, the patents were not castoffs from failed startups. Only 21 of the 121 entities are out of business. Third, the NPEs do appear to be “trollish,” at least by the definition that they wait for an industry to develop. The average number of days between patent issuance and the filing of a complaint was 3124 (more than 8.5 years) with a standard deviation of 1976 days (5.4 years). Of course, this could mean that the patentees were ahead of their time, but even so, there appears to be significant delay before these patents were asserted in court. Even so, there were still plenty of patents that were rushed to the courthouse after issuance. Fourth, contrary to the conventional wisdom that justifies patent trolls, I did not find clear evidence that NPEs enhance venture capital investments by providing an additional liquidity option. To be sure, firms with patents were 10 times more likely to obtain venture funding than firms without patents, but it is not clear that NPEs create this differential. There was no statisitically significant difference in investment percentage between a random population of patent holding firms and firms that contributed patents to NPEs. Another consideration on this point is how many of the patents were contributed by failed startups. As noted above, only 21 of the companies who obtained NPE patents are defunct. More telling, however, is the fact that only 3 of the 21 failed companies received venture funding, while one additional company was publicly traded before going out of business. In other words, if NPEs are supposed to be a source of post-failure liquidity in order to encourage venture funding, they are doing a very poor job of it in practice. On a related note, it does not appear that many of the NPE patents come from small companies crushed by their competitors. Fifth, I did find that NPEs provide a valuable enforcement option for individual inventors who obtain patents. In short, NPEs enforce around twice the percentage of inventor plaintiffs than appears in a random population of litigated patents.
Finally, the article makes some general observations about patent quality – namely that patents asserted by NPEs look a lot like patents asserted by productive companies according to objective measures. The next phase will look at the litigation outcomes to determine the patent quality of the studied patents. I hope to learn much more about quality over time and by technology category, among other things.
In the meantime, I look forward to any comments readers have about the study or its conclusions. I was surprised about how wrong my own intuition was, which is why I focus on the myths about patent trolls. Just about everything we thought we knew – good or bad – does not appear to be true. The article may not change too many minds about patent trolls. Those who believe NPEs are bad for society won’t care much about where they came from. However, I think that NPEs are a reflection of inventive society — their patents come from all sorts of sources, and how we feel about NPEs should depend on how we feel about the people who invested in the research that create the patents and the role patent law played in innovation.
Posted by Michael Risch on April 21, 2011 at 03:16 PM
Comments
Thanks very much. I will take a look.
Posted by: 8 ball pool | Apr 9, 2020 2:40:32 AM
Well, in each of those there is an implicit benchmark. If you are rushing to file in the ED Tex., you are trying to beat the other side DJing somewhere else. If you are trying to file before patent reform, you are trying to avoid the patent reform law. Because it is a different benchmark from the problem of an NPE waiting until an industry gets started, it is not as if the NPE “can’t win” in the sense that the critics are self-contradictory. I took your response to be that I am saying that NPEs should sue (or provide some other type of notice) before the industry sinks lots and lots of fixed investments, and other people are saying that NPEs should wait. That argument has not been made, I don’t think.
And as for the point that waiting-and-trolling is consistent with a plain bad faith waiting strategy (e.g. Rambus) and a failed startup strategy (e.g. Mercexchange), I think that is right. Not that I think that matters in the least. If all the Mercexchange patent did is incentivized a failed startup and then exact a tax on the successful independent inventor, I’m still not seeing what social good the patent did at the end of the day.
Posted by: TJ | Apr 22, 2011 9:54:11 PM
I should add that the delay is consistent with waiting for an industry to develop, but it’s also consistent with the initial owners of the patents attempting to commercialize the invention before resorting to enforcement of the patent through litigation.
Posted by: Michael Risch | Apr 22, 2011 8:12:33 PM
Well, one strand is essentially the same as your view that folks hide in the patent office, and then get their patents and sue the next day. Of course, you are right that that may be preferable to hiding out, and then waiting to sue. Though given the statute of limitations, there is an argument that waiting is better because there is less time before the patent expires.
That’s not the only argument though. In general, while I was in practice and in various press accounts, people have commented negatively about those who file suits right after their patents issue. I’m not saying that the criticism is based on any logical argument – only that it exists. Here’s discussion about rush to file pre-reform: http://thepriorart.typepad.com/the_prior_art/2008/12/stanford-iplc-release.html
Here’s one about the rush to file in ED Tx (though I suppose they might have been filed elsewhere): http://www.nytimes.com/2006/09/24/business/24ward.html?_r=1
Here’s a law.com article characterizing the myth as patent trolls in a “gold rush” (but debunking it) http://www.law.com/jsp/lawtechnologynews/PubArticleLTN.jsp?id=900005494579&slreturn=1&hbxlogin=1
Here’s another one saying that trolls are a phenomenon of business methods in 1998, so people are rushing to get patents and then running to sue: http://infringement.blogs.com/philip_brooks_patent_infr/2005/09/face_the_patent.html
etc.
Posted by: Michael Risch | Apr 22, 2011 7:49:44 PM
I guess I have never heard of this version of the “rush to court” argument, and I cannot see how it would make any sense. Saying the problem is that NPEs file too quickly and should negotiate first instead — I get that. Saying that NPEs file too quickly and should never file at all — I get that, too. In neither case am I endorsing the argument, but I can at least see how it works.
Saying that NPEs file too quickly and what they really should do is file the exact same suit a year or two later — I don’t get that at all. Our entire legal system is predicted on the opposite. We have statutes of limitations and laches doctrine to counteract this in every other context except patent law (where there is no statute of limitations and laches is weak). And we have DJ actions in theory to prevent patentees from waiting to file while accruing damages. So who makes this version of the “rush to court” argument, exactly?
Posted by: TJ | Apr 22, 2011 7:34:04 PM
Some say they rush by filing ASAP (with a long continuation chain, of course). That’s what I’m talking about – not the avoidance of DJ actions. That would be harder to measure, obviously.
Posted by: Michael Risch | Apr 22, 2011 4:51:57 PM
Michael,
We are talking different benchmarks between “delay” and “rush to court”. The people who criticize NPEs for rushing to court–as far as I am aware since I’m not really one them–seem to do it on the basis that many NPEs file before negotiating to prevent a DJ that chooses a different forum from the EDTX. The delay I am talking about is whether they wait for the industry to develop. An NPE can certainly wait for 10 or 15 years for the industry to develop and then “rush to court” without negotiating. In fact, it can rush to court the day after its patent issues, since the best place to hide (as long as one is willing to give up running damages in favor of injunction leverage) is actually the PTO.
Posted by: TJ | Apr 22, 2011 4:17:25 PM
Thanks for the comments, TJ. You are right that if delay is your key criteria, then your view is definitely supported. I will note that there is a strand of criticism out there that says that trolls rush to court, so they can’t win either way on this one.
As to your questions: You later say that “only 21 of the companies who obtained NPE patents are defunct.” Are these the same 21, or is the number just a coincidence? **coincidence – different companies. The paper has more detail, but licensing companies where we could not find information were not considered “defunct” because such companies are often private with their information. Other companies that had no data available were considered defunct.
Second, is there any reason to believe that the sample is not representative of all NPE patents? **It’s possible, and I address that concern in the article. First, I doubt that’s true, and I’ve not seen any evidence to support it. Second, even if it is true, the volume of litigation by these 10 NPEs is large enough to be worth studying separately.
Posted by: Michael Risch | Apr 22, 2011 8:13:03 AM
So I’ll be the first to confirm that your data does not change at least one person’s view of things, but it is interesting. Your second point is surprising, because I largely had your same intuition about where NPEs would get their patents from. Point 3 seems to confirm to me everything that is wrong with trolls, when they act like trolls. So the only point that you concede the conventional wisdom to be mostly right on is also the only one that I think matters.
Two questions. The first is a request for clarification. You say that “21, or 17%, were solely licensing entities at the time of patenting.” You later say that “only 21 of the companies who obtained NPE patents are defunct.” Are these the same 21, or is the number just a coincidence?
Second, is there any reason to believe that the sample is not representative of all NPE patents? For example, is it possible that many failed startups would, instead of selling its patent to a large NPE, choose to become a small NPE itself? Might smaller NPEs be disproportionately holding the failed-startup-castoff patents that you are not seeing (but which we had both expected to find)?
Posted by: TJ | Apr 21, 2011 9:56:20 PM