Patenting used to be a lonely pursuit. Patent applicants would work in isolation, secrecy being their only protection before the patent application was filed. The patent would be granted or rejected in a pas de deux involving just applicant and examiner. Once a patent was granted, licensing battles were also fought mano a mano while other potential infringers watched anxiously from the sidelines.
Attempts to bring patents out into the hubbub of public life sprang from the rise of computers and networks, in tandem with patents that threatened to place a throttle hold on them. Many patent review efforts exist--PatentFizz, EFF's Patent Busting Project, CAMBIA's Patent Lens, and more--but their effectiveness has been constrained by the oddities of the patent system, the difficulties of reading patent language, and the sheer size of the undertaking in an age where some 150,000 patents are granted each year worldwide.
Peer-to-Patent, an academic research project affiliated with several patent offices, and Article One Partners, a commercial venture, are two promising entrants into the space. This article will described and compare these organizations, highlight a new "post-issue" site erected last week by Peer-to-Patent to seek prior art on patents that have already been issued, and try to tease out the social and economic trend represented by the organizations. Contents include:
- Peer-to-Patent and the new post-issue site
- Article One Partners
- Can patents be debated in the public thoroughfare?
In a large field where patent-busting projects jostle each other, Peer-to-Patent stands out as uniquely promising. As I've described before on this site, Peer-to-Patent formed an agreement with the Patent Office and is integrated into the patenting process. This gives it an unmatched advantage for several reasons:
- Participants could crush a patent in the egg--before it was granted--instead of waiting till it reached the nursery and became much harder to kill off (because a granted patent benefits from a presumption of validity in the courts).
- Participants were guaranteed the ear of the patent examiner. The Patent Office promised to consider the top ten instance of prior art found by the Peer-to-Patent community for each patent.
- Participants worked in groups, which was fun and permitted networking for other purposes.
Success, as usual, brings new challenges. The reward for doing good work is usually more work, and Peer-to-Patent is true to type: they're expanding in several ways that create a need for more volunteers.
- They're getting more patent applications. During the initial pilot phase, they had to do their own publicity and beg applicants to offer their patents for review. Now the Patent Office is notifying eligible applicants itself. And because Peer-to-Patent applicants are rewarded by an expedited review (which could shave two or three years off the process), many are taking advantage of the offer.
- They've added a new area--the highly publicized field of business process patents--to the area they covered during the pilot, computer patents.
- They're forming agreements with other patent offices to add new jurisdictions, particularly the United Kingdom.
- They're started a post-issue site that operates more like the other patent-busting sites mentioned earler, seeking prior art for patents that have already been granted.
The last item in that list raised questions for me. It relinquishes the key selling points of Peer-to-Patent (access to applications before they're granted, and access to the patent examiners).
I talked to Mark Webbink, Executive Director of the Center for Patent Innovations that now runs Peer-to-Patent. He explained that the post-issue site grew organically out of the process and was suggested by sponsors who came across existing patents during their research. Perhaps the same valuable network of experts built for the Patent Office can help here. The utility of this site rests on the degree to which the public can help by "grazing," a concept suggested by Mike Schultz, a PR representative for Article One Partners.
Most prior art is intimately known to somebody: the person who invented it, a student who wrote or read a paper about it, a user of the product, etc. If one of these people fortuitously happens upon a patent, he can provide a clincher instance of prior art in a few minutes.
Peer-to-Patent may be most useful for such people, who "graze" by returning often and quickly perusing the summaries of patents. In any case, the success of Peer-to-Patent rests on its positioning in the wide-open public thoroughfare, attracting a wide diversity of visitors from around the world, the polar opposite of a patent applicant working in secrecy.
Most people, though, can find prior art only by strenuous searches: determining the relevant literature, going back over dusty old journals, and so forth. These "delvers" are less likely to participate in a free forum such as Peer-to-Patent, unless they have external motivation such as protecting the interests of their own company or developing a reputation. The patents on the post-issue site appeal to these motivations by highlighting patents that are likely to have a broad impact across a whole industry (currently the computing industry). But delvers may be more attracted to Article One Partners, which I'll describe next.
Article One is highly reminiscent of Bounty Quest, a firm that had a short and unsuccessful run in the early 2000s. (Tim O'Reilly was an investor.) Cheryl Milone, Founder and President of Article One, worked in sales at Bounty Quest and assimilated the lessons from its experiment.
In fact, Article One is interesting for its attempt to turn a very dicey area--prior art for patents--an area that sprawls across many economic, legal, and technical domains, into an actual market with enough regularity and reliability for people to participate. I won't attempt to cover all the details here, but I'll mention enough to show how they've delicately created an arrangement of pulleys and levers to extract revenue from this roiling mix.
Sponsored studies, such as Bounty Quest ran, form only a portion of Article One's revenue model. They prefer to look for patents that seem to have a high economic value, which are not exclusively patents in litigation. In fact, they don't particularly seek out patents that seem vulnerable (a very difficult determination to make); they are equally interested in patents whose novelty is likely to be confirmed by a prior art search that fails to uncover invalidating art.
Overall, Article One plans on three sources of income:
- Sales of promising instances of prior art to patent holders and those at risk of being targeted for infringement. Patent holders can use the prior art to reissue a patent with narrower and stronger claims; those targeted for infringement can try to overturn the patent.
- Investments in companies based on results of searches, which can indicate whether they are likely to profit from patents.
- Sponsored studies (the Bounty Quest model) which generate fees for the initial posting and contingency fees upon a successful conclusion.
All of this depends on a pool or "referring network" of experts willing search for prior art. Article One is engaged in radically fine-grained outsourcing, just like two other companies I recently highlighted, uTest and TopCoder. And like them (as well as Peer-to-Peer) it's network will probably have to develop aspects of community in order to prosper.
Article One attracts and retains people by providing immediate rewards (a member starts earning profit-sharing just for registering) and carefully graded increments in profit-sharing for various activities: posting to forums, recruiting new members, and supplying prior art--even if that prior art is not ultimately chosen. The big rewards go, of course, to the member who finds usable prior art, and a smaller share to the member who brought him or her into the network.
Article One's competitive model recalls TopCoder, Innocentive, and Nine Sigma--the latter contracting to provide its members to Article One. Access to the three million technologists registered with Nine Sigma, or affiliated with it through international networks, gives Article One the authority to tell patent holders that prior art searches were thorough.
Results of searches, naturally, have to be kept secret if Article One is to sell them. This is the big differentiator from Peer-to-Patent and other open prior art sites. Peer-to-Patent leaves all prior art up permanently. But Article One will also publish prior art under some circumstances, particularly when it uses the results to drive investments (the second of the three revenue models described earlier).
If you find prior art, why not just go to a litigant directly? Parties to patent lawsuits often get cold calls from individuals claiming to know prior art. Milone points out that companies will find it much more productive to deal with Article One, which adds the authority of independent verification by recognized experts. In addition, Article One protects the individual who uncovers the prior art, ensuring he profits from the information, as well as from being part of the Article One community and its profit-sharing.
Will Article One take hold? Initial numbers look promising: Milone told me their site has received 650 submissions of potential prior art during their first eight weeks, whereas Peer-to-Patent has received 270 during its first 18 months. And even though Article One has no sales force, several companies initiated contacts with it to sponsor studies. Since the days of Bounty Quest, the power of crowdsourcing has powerfully demonstrated itself.
I should stress that Article One and Peer-to-Patent feel no sense of competition. A long-term comparison of their models will prove enlightening. But what do these projects teach us about the field of patenting?
To return to the theme at the start of this article, the patenting process is dangerously ingrown, and everyone is hoping that a whiff of fresh air will clear the contagion. Even Barack Obama campaigned on the promise of "opening up the patent process to citizen review," a clear reference to Peer-to-Patent. (And the creator of Peer-to-Patent, law professor Beth Simone Noveck, is now on the federal TIGR team.
But the problem isn't just the lack of information. The sums of money involved are too great. The means for resolving patent disputes is too heavy-weight. Our inventors have to worry about lawsuits shaking out to several million--or even several hundred million--dollars. Very few patents go to court or are even licensed, but an atmosphere of anxiety pervades lucrative areas such as pharma, Internet technologies, and genetics.
Perhaps industries could agree on mandatory prior art studies, just as some industries now have mandatory arbitration. The prior art study would not preclude a lawsuit, but it could be an initial stage that could persuade one of the litigants to settle.
Article One would still get plenty of business, because licensing fees for some patents can be very lucrative. And not every dispute can be settled through a prior art search, of course. Many controversies in patenting can be settled only through the courts, such as "What is patentable?" (the Bilski case) or "What is obvious?" (the KSR case). What I hope is that, as more patent holders come out of hiding onto the thoroughfare, they'll find they like the street life.