I got a Quora request for an answer. Just posted this.
Let’s start with some basics regarding policy. Since the early 15th century, policy has been a creature of government, and of organizations with the idea of behaving as if a government. The idea of policy is to set out a pattern of practice that serves the interests of the organization. Thus, whatever IP policy one wants to draft will depend on the interests of the organization that the policy targets—or, more concretely, the interests of those having influence over the organization—a CEO, or president, or director, or (worse, possibly) a few people on a committee, or (even worse, probably) a lawyer serving as counsel for the organization, without much experience in actual IP management. It is easy for an IP policy to go bad, while not actually looking bad to the naive eye. That’s an admonition, not a principle.
As a practical matter, a policy identifies practices to be followed by an organization’s personnel, identifies organizational resources to be used to engage in such practices, and delegates authority to individuals to make decisions with regard to the policy, practices, and resources.
In the context of intellectual property, an organization’s policy operates in an arena of IP laws—patent, copyright, and trademark laws. In the United States, there are federal statutes and regulations, state law, and common law addressing various aspects of IP. A bland, but still useful IP policy might attempt to restate the defaults established by these various IP laws as they pertain to the organization. That might provide helpful reading, along the lines of “follow the law.” It’s good to follow the law, one might think, at least in matters of how an organization deals with IP. But often folks get the idea that (for clarity, say, urk) it is in the interest of the organization to change the defaults established in IP laws, and use policy for this purpose—to dictate what should be in employment agreements, or patent agreements, or consulting agreements, or what should be considered “unethical” (yes, the University of California has had a provision in its ethics policy that no one is allowed to do something for a “higher purpose”—noodle on that for a minute). So policy is often used to change defaults in the interest of the organization—or at least that is the purpose that the folks drafting policy will claim, even if all they write is to make it sound like the policy is in the interest of the organization.
At the very start, folks will make the assumption that having an IP policy is in the interest of the organization, and well this may indeed not be the case. The University of Wisconsin went for over 120 years without an IP policy, and did some great work anyway, inventing the idea of vitamins, figuring out why a farmer’s cows were dying and turning that into warfarin, and the like. The National Research Council, however, could not handle the idea that American universities didn’t have IP policies and went on a multi-decade campaign, led by Archie Palmer, to get universities to adopt IP policies. Now every university has an IP policy and they are so far down the line on them that when in one case a university’s IP policy was questioned, rather than accepting that the administration had plainly misread the policy (and misrepresented it to the court), the university’s lawyers demanded that I read other universities’ IP policies into the record of the deposition as if those other policies applied to that university. So, er, f’d.
So: first principle. Do you really need an IP policy for your organization? Maybe the defaults in law are good enough. Maybe the defaults in law are better than anything your folk can possibly draft for the organization. I once had a vice president of research at a very big university tell me privately that she knew their IP policy was terrible and the practices it endorsed didn’t work, but she was unwilling to change until there was a better policy somewhere that did work, that the university could copy. I told her that “no policy” was way better than what she had, so why not at least start there. But no, there just had to be a policy other than “no policy” even though it worked fine for Wisconsin for a century.
Now let’s consider what defaults an IP policy might seek to alter. In IP relationships, five elements tend to dominate: ownership, control, money, attribution, and risk. Typically, folks look at only two of these at a time—someone wants to own because they want money, or they want to control because they want attribution (even if they assign copyright away to a publisher, say). American university IP policies tend to focus on ownership and money, as if these are the things that matter. Usually lost in the first paragraph’s statement of purpose will be something about public purpose or innovation or making inventors happy, but that stuff is nonsense fluttery once you get to what the policy actually does. It is not always the case that an organization is ahead by owning IP or expecting to exploit IP for money. (One might consider that the time-proven most valuable bit of IP is the goodwill attached to trademarks—and I don’t know a single university IP policy that considers goodwill as the core of its IP policy—money, not even reputation (though the thing administrators truly hate is bad press).
And of these five elements, perhaps the most important to consider is risk. The liability that comes with ownership, or control, or attribution. Or the bitterness and dispute and audit that often comes with money. So, if you must have an IP policy—make sure you cover at least ownership, control, money, attribution, and risk. Ask—why should the organization want any of these things? Why does the organization want the liability of having these things?
A clever but poorly choosing administrator, or attorney, often will propose that the IP policy should claim everything that can be claimed (as to ownership), and give back (to inventors and authors) only what the organization is forced to give back (such as, through litigation, or pressure from a major donor). This way lies badness. Organizational ownership is pretty much death to invention and works of authorship. American universities have taken out over 100,000 U.S. patents (at, say, $10K per patent—do that math some time—not to mention failed patent applications (mostly same math but no chewy nougat) and foreign patent applications (both failed and issued). Most of those inventions are held behind the patent ownership wall, never licensed. The patent serves to make companies avoid the university’s research—design around the patent, make the university’s patent irrelevant. How does such a thing serve the university’s interest? Got me there. Seems it makes all those millions of dollars of research per year even more useless than otherwise.
The idea of claiming everything goes to absurd, if not cruel, lengths. Everything that can be patented or not patented. I get the patented part—at least there is an ownership theory. But how about the not patentable—that would be everything else, whether it can be owned or not owned. How does that work, even.
So another principle: don’t claim everything; instead claim only what the organization can actually use, and even then, it is better to claim less than one could than to claim more. To claim less means that the organization has to run an IP management program that people really do want to use, that actually works—whether that is to encourage innovation (as if innovation needs encouragement) or to make money for everyone involved (as if that’s a good organizational purpose for IP—works, for for-profits (sometimes, well, rarely), but sucks for universities (even in the rare instance where money does get made, still sucks). Furthermore, by claiming less (or even not claiming anything at all, but authorizing resources to be used if someone voluntarily offers to assign, and the organization’s personnel authorized for IP things agrees), the organization will see better IP, or at least IP better suited to its management, with people—inventors and the like—who have chosen to work with the organization. People who want to evade an organization’s IP claims have a ton of ways of doing it, and an organization that aims to stamp all these out builds an IP policy of tyranny, a Moloch policy that would rather destroy IP opportunity than see anyone else benefit. A bureaucrat’s thumb in every innovation pie is another way of explaining the idea to take everything and choke back only what one is forced to.
The folks talking everything also conflate patent and copyright matters. They love the abstractness of “owning” stuff so they draft that way. Typically, they take a patent policy, abstract it, and that makes copyright works of authorship just a special form of unpatentable “invention” (where invention means everything patentable and not patentable). This is a horrible conflation and leads to the stupidity of requiring all those “inventions” to be “disclosed” (as if patentable inventions) when the policy applies to every email, every note on a notepad or on a dry erase board). Sure, maybe a for-profit wants all of that held as trade secret (but trade secrets are not statutory IP—and that’s a whole nother thing—consider the intersection of US export control law and trade secret in an organization employing non-US nationals—hoo-haw).
So don’t conflate patent and copyright, inventions and original works of authorship. Better to have two policies, each focused on a form of IP and how it operates, in the interest of the organization, than to try for a happy general abstraction with bits and pieces later trying to correct for the bad choice on policy architecture to start. And don’t fixate on patent and copyright, when trademark and its goodwill are day in and out the most valuable of the IP assets.
In all of this IP policy talk, we should consider non-IP intangible assets, or NIPIA. A poorly conceived and drafted IP policy can destroy NIPIA. What is NIPIA? NIPIA may be all sorts of things—a social network, qualified leads, people who exhibit “strength of weak ties,” articulate inventors, skilled prototype builders, an ad hoc standard, development of a shared research platform. One can use a patent to try to make money (and generally fail) or dedicate that patent to an emerging standards (and have a seat at the table for decisions about a future roadmap for development). Which ought to matter more to an organization? For the organization’s IP policy. NIPIA generally cannot be owned. Even trade secret cannot get at an employee who has some public visibility, has design skills, makes improvements, knows a bunch of people in the big wide world, and can move on as soon as the organization goes out of its way to make things bitter with a stupid IP policy and management unit. At PARC, in the Bay Area, engineers have told me that they were required to disclose their inventions. The organization then would decide whether the inventions were “strategic” or not. If “strategic,” then the organization would take over, create a new division, staff it up, and run the invention into the ground. If non-strategic, then the engineers could run with their work outside the company. Hence we have ethernet, graphical user interfaces, mice, and lots of other whatnot. So, PARC’s IP genius was not claiming everything and its loss was claiming strategic things too hard. Here, the principle comes from the lyric of an otherwise mostly forgettable 1981 rock song—”hold on loosely but don’t let go, if you cling too tightly you’re gonna lose control.” Probably one could build a great IP policy consisting only of lyrics from popular music. It would work, it would be great, but it would lack bureaucratic gravitas, which we all know is essential, essential I tell you, to innovation, money, and all that.
That’s a start, anyway. IP policy is like an onion. Lots of layers. Peel one back. Eyes tear up.