University Patent Practice: Share

Here are five ways to use a patent:

  1. Nonuse     Don’t practice the claimed invention and exclude all others
  2. Troll          Don’t practice, demand payment if others practice the claimed invention
  3. Flip           Don’t practice, and exclusively license, assign, or sell the patent
  4. Practice   Practice the claimed invention while excluding all others
  5. Share       Practice and allow others to practice the claimed invention

Each of these methods has its own variations–some with significantly varied effects. Let’s look at each of these approaches and consider university patent management practices. The first three patent uses involve non-practice. The last two involve use–practice. Let’s look at Sharing.


We turn now to “share,” the last of the five strategies for patent management. In the share approach, one practices an invention and allows others to practice the invention. While the word “share” might give the impression of fluffy happiness, share isn’t necessarily all sweetness. In industry settings, sharing patents might take the form of a patent pool, which can create barriers to entry to an industry and allow a few players to dictate players, markets, products, and innovation. Or patents might be cross-licensed, so companies agree not to compete based on legal positions but rather on other bases, such as efficiencies. Then again, along with cross-licensing patents, companies might agree not to compete and divide up an area, so that they have a mutual advantage over other competitors. These sorts of arrangements can be, in the US at least, viewed as anti-competitive and perhaps illegal, but stuff happens, and patents get drawn in along with anything else.

A legacy of anti-trust concerns around patent pools and cross-licensing has put a bit of darkness around sharing of patent rights. However, most everyone in industry (and government) also recognizes that it is futile to have endless legal battles over bits of technology that must be combined in order for there to be products at all. Thus, patents can also be contributed to standards managed by independent standards organizations–not necessarily independent of all influence from industry players, of course, but typically not subject to the outright control of any one player. Standards are generally based on the idea of FRAND patent licensing–fair, reasonable, and non-discriminatory. Anyone who desires access to the standard should be granted a license to the patents required to practice the standard, without regard for competitive position (other than, perhaps, claiming rights against the practice of the standard itself).

When a patent is contributed to a standard, its claims are evaluated for their participation in the standard. Claims of the patent that are required to practice the standard are designated as “essential” and become part of the standard’s FRAND package of rights. Any other claims of the patent are called “non-essential” and the patent owner is free to exploit those rights in whatever way she, he, or it chooses.

As with nonuse and flipping, we find a distinction between the existence of a patent and its claims. It is the claimed invention that matters–not merely that there is a patent. An exclusive license that grants rights in all claims to a licensee that practices only a few of those claims means that portions of the claimed invention have been licensed for nonuse. The unused claims are unavailable for use by others unless the licensee is motivated (or compelled) to grant sublicenses for unused claims, or accepts a downgrade of the license to nonexclusive regarding those claims, so that the original patent owner can grant licenses directly.

Bayh-Dole’s march-in provisions also track the idea that nonuse can lead to a federal agency determination that can force licensing of a patent on a subject invention, but does so without reference to claims. Apparently, the assumption made by Bayh-Dole is that a patent = an invention (which is of course true in a sense) but ignores that an invention, as far as patenting is concerned, is defined by the claim structure of the patent document, and may include all sorts of variations, applications, and equivalents. For Bayh-Dole, apparently, practicing any one of the variations or applications is sufficient to suppress the practice of all the others. Of course, under Bayh-Dole there has never been any consequence whatsoever of never practicing within any claims of a patent on a subject invention, as there has never been a successful march-in proceeding.

In the case of sharing a patent to a standard, only those claims necessary to the standard come within the control of the standard. Again, it is not the patent that matters, but rather the specific claims.

The idea of sharing a patented invention runs against present, established university practice. Why go to all the work to patent an invention only to share it with everyone? One can share without a patent–clearly, the point of patenting must be to be able to flip the invention exclusively. The rationalization is that every invention must be “developed” and “development” takes a huge investment, and an exclusive license is the only way to induce that investment, as the licensee then can enjoy a monopoly as a means to recover that investment (and more). Here’s the outgoing director of MIT’s technology licensing office:

Patents are needed because the whole idea is if you’re going to get somebody to invest a lot of time and a lot of money, if you succeed you don’t want the other guy, the bigger guy, saying, “Well, thank you very much. Now that you’ve shown the way, get out of the way.”

We are primarily using patents as an incentive for investment. We’ll give you an exclusive license if you’ll commit to development. And then if you succeed, you can recoup your risky investment because you will have patent protection in the market.

There’s no interest evident here in the idea that development costs might be shared across an industry, that common standards might allow for competitive development based on interoperable components, or that university discoveries are not “early” proprietary products that need to be “ripened” in startups so much as they are contributions to the general stock of resources available to developers.

Many university licensing officers also believe that they will make more money from exclusive licenses and will have less work to do, as they have only to license each patent once. This perception exists despite the early success of biotech inventions such as the Cohen-Boyer gene-splicing patents that were licensed non-exclusively. But exclusive licensing turns out to be difficult. One doesn’t license to established companies because a technology is “too early” (which may be true, too) but because most companies do not want to undertake the full cost of developing a product when they don’t have to, and don’t want to pay the higher rates and deal with the invasive requirements that often come with an offer of an exclusive license. The form of the proposed exclusive relationship carries with it such deterrents that many companies refuse. University licensing officers then make a virtue of starting companies–often just shells–and licensing exclusively to them in the hopes that a round of investment will lead to a valuable equity position, quite apart from whether a product ever gets developed based on the exclusively licensed patent.

Non-exclusive licenses do get lip service within the university licensing community, but typically it is a matter of defense of practices, a”see, we do grant non-exclusive licenses, too” kind of rhetoric. Non-exclusive licenses are routinely granted in sponsored research agreements. The US government, of course, requires a non-exclusive license to the patent rights in each subject invention acquired by a university. Most industry sponsors of research require an even broader non-exclusive license than the government receives, including the right to sell and typically also involving an option to negotiate an exclusive license. (Typically such licensing terms run to one or two sentences in a research agreement, but require ten or twenty pages if placed in in a “license agreement.” I once dealt with university counsel that insisted that for every license, we must use a many-paged license template. I pointed out that the university routinely granted licenses a sentence long in research agreement. It became apparent that the university attorney thought of a license as a document, not as an enforceable promise.)

These sorts of upfront, “embedded” licenses in research agreements do happen–all the time–but rarely are they part of a licensing office’s default strategy. If there’s one rule with non-exclusive licensing, it is if you have just granted your first non-exclusive license, then grant another! The primary problem university licensing officers see in embedded non-exclusive licenses is that they are typically royalty free. There’s no money to be made (or, the university licensing officers don’t know how to make money) and therefore such licenses are not interesting. If one followed up such licenses with offers of services, technical information, and improvements, perhaps a relationship would form that would become paying, though the payments would not be viewed as in exchange for a license.

This is not just some hypothetical idea. We ran such programs for a decade at the University of Washington, and demonstrated that they worked–technology was transferred, licenses were granted, but the primary exchange was services, not relief from the threat of being sued.

Despite the idea of building relationships that involve non-exclusive licensing, a non-exclusive license “is just a tax” according to another often repeated slogan made popular by well placed university licensing officials (for instance, here, slide 5). One might reply that an exclusive license “is just a flip.” A more substantive reply, however, is that non-exclusive licenses come in all sorts of varieties, and while some might be the result of trolling behavior (which extracts a share of the proceeds from companies caught practicing a patented invention unlicensed), other non-exclusive licenses derive their value from sharing.

Consider the nature of transactions by which a university might receive money–tuition, donation, research services, licensing, litigation. If a university invites industry reps to a workshop briefing on a new technology, it might charge a conference fee of $500 and no one would object. With 50 attendees, a university would gross $25,000, each company attendee making a $500 decision that it was worth it to learn about a new discovery. It would merely be icing on the cake if in addition each attendee was assured of a royalty-free non-exclusive license to the new technology.

But consider what happens when a university leads with the patent right–with the idea of an exclusive license. Even if the university asks only $25,000 up front, the matter goes to legal counsel and the vice president for product development. Do we want to make this product? Do we have to take a license? Do we want to pay this amount? It’s an entirely different decision environment. And universities rarely make exclusive offers at $25,000–it’s more like $100,000 plus a 5% royalty on sales plus milestone payments of $50K, payment of all patenting costs, sublicensing fees, and perhaps some company stock as well. It’s good money when you get it, but universities tend not to get it very often–maybe once a decade or longer, if ever. By contrast, a sharing approach might get a university $25K a year for a decade, plus build working relationships with scores of industry representatives who then may contribute to research and instructional programs in all sorts of ways–it’s just that their contributions won’t be booked as licensing fees.

The odd thing is, sharing is at the historical roots of university interest in patents. Starting with Cottrell’s formation of the Research Corporation as an agent to manage the interface of university inventors with industry, the basic gesture has been to make patented inventions arising from university research available on a non-discriminatory, non-exclusive basis. Such an approach is consistent with scholarly publication, with a mandate to serve the public without choosing favorites or compromising institutional standing, and with research collaboration, especially with industry. If companies know that whatever gets supported at a university will find its way via a licensing agent to a FRAND program, then they are assured of access on reasonable terms. Universities that took an early interest in patents did so not to make money (despite MIT’s efforts) but rather to prevent the formation of monopolies around university research. This concern is 180 degrees from the present university licensing mindset that only by actively promoting monopoly positions–even at the expense of publication and collaboration–will new products ever get made.

There are deeper problems with the non-sharing position. First, many university inventions can be practiced without the need for a product implementation. Methods, for instance, often are just that, methods. Research devices that have been invented and constructed in a university shop can be constructed in anyone else’s shop, too. Software that’s been written to perform a function in a research setting can be distributed and with the source code also available, adapted to various uses–again, without a “commercial” product version made first. Laboratory medicine assays–part method and part recipe–can often be implemented based on a published specification without the need to purchase a commodified version of each assay. (And even if purchase of a commodified version was the desired route, purchase at monopoly pricing is less desirable than pricing based on competition.)

Thus, the majority of university inventions do not have to pass through a commercial version in order to be used. They may be used for research (research on the invention, research using the invention) and they may be locally implemented for commercial purposes (to produce product, to test product, to design product)–all without the need for “commercialization.” The federal standard (in Bayh-Dole) is use of inventions, not commercialization and not even use by for-profit companies. Most certainly, the federal standard is not use of the patent at the expense of the invention, nor an insistence on “commercialization” (commodity production of product for sale at a profit) that suppresses research, collaboration, and local use of an invention.

Share strategies then may break up a patent’s rights. Sharing may permit anyone to make, have made, or use a patented invention, while reserving the sell right for more focused treatment. We call this a “make-use” commons approach. By doing this, one creates immediately easy access to an invention for research collaboration and local implementation. When an invention is new, the opportunities for providing support services are the greatest. A university-based consulting program (or instructional program) offered at the time an invention is announced might carry the primary financial interest. Promoting use of an invention may itself be the greatest contributor to financial returns for a university later. If many organizations are using an invention, then there will be opportunities for companies to make commodity versions.

We saw this with a software package developed at the University of Washington, made available for local use for a modest fee that included source code, updates, access to the programmers for assistance, and a standing option to sell if one wanted. For a time, we had about 200 industry sites paying about $4,000 a year to participate. When things started out, the lab gave out copies of the software to industry associates with an understanding “if we ever get around to licensing this code, you agree to take the license.” When we got involved, we went back to those companies to find out what they thought should be in a license, went through a few drafts with them, and found a common set of principles and price points.

What these early adopter companies realized is that whatever they were willing to pay, we were going to charge all the other industry users of the software, and they would get the benefit of that development work. When there were only a handful of companies involved, they were still getting a five to one return on their payments–for each $4K they paid, we did $20K of work to which they had access. The next round of companies asked to participate in part because they heard there was software being used and wanted to see what it was. Participants expanded from about five to thirty. Now the return for each was 30:1. The next group in wanted to use the software because that’s what everyone was using. Things expanded again to 100 or more.

It was then that we started getting organizations that wanted the software in a commodity format–they didn’t want to compile the source, they didn’t know how to install the software, and they wanted support for a product. And that’s what created incentives for some companies to start selling a commercial version of the software. We never had to suppress use in favor of an exclusive license, and we’d never have got to an exclusive license, I’m sure, if we had started that way. Of course, there were no patents involved–which leads technology transfer officers fixated on the glories of exclusive patent licenses saying that “software is different.” But software is not different from most university research inventions. It’s the university research invention that only can be practiced in a commodified, commercial form that is the rare thing.

Invention-sharing strategies in a university, then, may take the form of workshops–leading with instruction and assistance and of make-use commons–making use rights broadly available as a pre-condition of the possibility of commodity sales later. What about patented inventions that haven’t been taught or licensed for use at the outset? This is where university patent administrators have to deal with their demons. An invention covered by a patent that sits idle may well end up being used–and worse, if people feel they must find workarounds to avoid using the university work. What then, when a university discovers use of an “unlicensed patent” and equates it with infringement rather than success?

The demonic approach–let’s call it–is to shake down infringers for a license–perhaps nicely at first, but then more stridently, leading to litigation and trolling behavior. But one doesn’t have to play Satan in this situation, much as the folks at Caltech must disagree, given their current trolling activity. One could, instead, offer royalty-free licenses to those using or appearing to use the invention, with forgiveness for any past infringement, a free campus parking pass, and an invitation to visit campus and work with the inventors and associated labs on improvements and applications. One gives up the opportunity to create a dispute and the attendant costs and bitternesses, but one gains the change to create new relationships, reduce potential tensions, and position the university as a player rather than an adversary. Recognition rather than threats leading to payments (or invalidation of the patents) may be the greater value to the university. Recognition creates opportunity, where threats lead to wariness. The neighbor who invites you to the barbecue is arguably way better than the one that tries to poison your cat if it happens to pass through his (or her) yard.

There are other modes of sharing–creating standards (ad hoc or formal), dedicating work to the public domain (so there is no ownership and no licensing involved), open innovation licensing (using a license that restricts only attempts at restriction), and joint research (creating an environment for the pooling of research funds). Any of these approaches are highly consistent with university values and practices, may result in financial support for the university (and private income opportunities for inventors and others, if that’s what matters), and use the patent system in ways that are distinctive to university practice rather than importing corporate or troll practices and expecting such practices to be accepted with all their costs, fuss, and damage if, once a decade, there’s a big cash payout or settlement.


This entry was posted in Agreements, IP, Policy, Technology Transfer. Bookmark the permalink.