Dear Vice Provost for Research,
An insightful vice provost for research once told me that the director of technology transfer had the second most difficult job in the university, after the dean of medicine. Having served as a director of a campus technology transfer office, I believe he was right.
A director of technology transfer has to be at least a five-skills player, and play each of those skills every day at game speed. She (or he) has to know intellectual property law and practice–and even patent practice alone has enough detail to keep one going for years learning the fine points. She also has to know research administration, contracting, and policy; know how to work collaboratively with research personnel working across a broad spectrum of areas of inquiry, from biomedical to computer networks to language learning; be adept at marketing and market strategies pertaining to new technology and ideas; know finances and handle large budgets, often with larger risk exposure than is immediately evident; be able to negotiate with large companies, startups, foundations, other universities, government officials, and venture investors; know and contribute to university and public policy on innovation and economic development; and be capable of making articulate presentations to the university, to industry, to the venture community, and to the public–and dealing with a multitude of competing interests. I guess that’s seven skills, but you get the point.
Dealing with new technology is no easy matter. At one point I put together a list of the administrative subsystems needed to operate a typical technology transfer office–there are more than 50, ranging from managing appointments of outside counsel for patenting matters to dealing with new disclosures of inventions to keeping track of the information necessary to pay royalties to inventors.
That said, it’s easy to understand why a technology transfer office doesn’t produce as much money on as many inventions as you might expect. It’s all so complicated. But while this is a rational-sounding explanation for the difficulty, it turns out to be wrong. The pitch that you may have heard–and even may have come to believe–is that with proper management and proper organizational structure, a university can take in all the inventions created by its faculty, students, and staff, sort through those inventions to find the ones with commercial value, file patent applications on those inventions, and then market the patent rights to industry, to startups, and to investors, generating tons of money for the institution. Every so often, a university “hits it big” with a patent or cluster of patents–the Axel patents at Columbia, the Cohen-Boyer patents at Stanford and UC, the nifty deal Emory pulled off, selling its future royalty stream to the HIV drug emitricitabine for $525 million. You perhaps want that same sort of deal for your school. Who wouldn’t?
But these sorts of deals are few and far between. Stanford has had three such deals in 40 years or so. UC estimates that 0.5% of its invention portfolio ever sees a commercial license, let alone a “big hit.” Half of UC’s licensing revenues come from 5 technologies in portfolio of 5,000 patents–and most of UC’s top 25 revenue-producing technologies date from the mid-1990s or earlier. Those big hit deals don’t come so often. In the meantime, universities have accumulated huge portfolios of unlicensed patents. Many university licensing offices do not even “break even” unless they have one “big hit” license to cover their losses on everything else for the two decades of a patent’s term.
One recent strategy to try to show “progress” in building a university licensing program has been to create shell companies and license patents to those companies for equity and a hope for future royalties. The licensing then is relatively easy because the university basically sits on both sides of the deal. That makes licensing numbers go up, gives the appearance of productivity, and also adds the benefit of appearing to favor small businesses–just not the small businesses that already exist and are selling actual products. Then those shell companies with licenses sit, maybe with a Phase I SBIR grant or some state economic development seed funding; sometimes they sit for years, waiting for their “early stage” technology to cross paths with a market. Most often, this just does not happen. But when a useful technology is developed, it is typically not because the startup did fantastic research and development for five or ten years and produced a commercial product that sells like hotcakes. No, what happens is that other companies develop something that arguably incorporates some method or system within the broad scope of claims of a patent licensed to a university startup, and then everyone gets geared up to claim infringement and demand a proper share of revenue from whoever did the heavy lifting to figure out how to build a commercially viable product.
The public goal of practical use gets turned, by the patent and exclusive license, into corporate theft. There’s money in threatening to sue, or actually suing, but when you start suing the companies that otherwise support research on your campus or hire your graduates, funny things start to happen. Companies close research centers near your campus (as Roche did after Stanford sued them) and companies don’t list you as a strategic research partner (as Boeing did after years of IP bickering with the University of Washington) or companies don’t go to your job fairs anymore (as Micron did after the University of Illinois sued them). So you have to decide if you are going to go down this road, and set up a situation in which money from a patent infringement suit is worth more than the goodwill and interactions that come with company involvement in your school’s research and instructional programs.
There’s a broader issue, however. The fundamental concept of aggregating all IP, sorting it for stuff that can be sold off, and raking in the money just does not work. It’s a failed approach. It’s snake oil. Any clever person can learn the sales technique of flattering the mark, disparaging the current efforts, presenting impressive credentials, and proposing a vision that leads to vast sums of money arising from a system, yes a system, of managing IP assets for value. And vice provosts for research, along with some few (but often vocal) faculty, are especially vulnerable to the snake-oil pitch.
The snake-oil model depends on taking ownership of everything and releasing (or leaving to die) the stuff that’s not worth anything (or doesn’t appear to be worth anything). The need to take everything is to prevent anyone from demonstrating a different approach to innovation. That’s called a “back door”–and doesn’t give the university the same opportunity to profit from the transfer of technology. Anyone not participating in the university’s effort to collect everything of value that might be licensed for profit is, then, cast as defrauding the public, having a conflict of interest, and giving away what should be the university’s property for private gain, or career enhancement, or worse, out of a defective understanding what constitutes a proper public mission.
But gathering all possible technology into one office staffed by five or ten or thirty or even fifty individuals is daunting. Per year: hundreds of invention disclosures–at the big research schools, one a day or more–thousands of software and digital media and database developments. And all of this across the areas of research in which university faculty are active. It’s not possible to staff an office with ten people with “area expertise” and expect them to know how to position some technical finding with a patent so that it will be licensed by a company and included with ten or fifty or three hundred other patentable bits to make a new product. Maybe ten or twenty inventions a year will be in the expertise sweet spot for a typical university office. Everything else is a distraction–whether valuable or not–and reduces the chances that meaningful licensing work will get done. Which do you think is going to be the more productive licensing office–one that gets 200 invention disclosures a year, but has the expertise to work with 20 of them, or an office that gets 20 disclosures a year, and each is matched to the expertise in the office?
The snake-oil model for university IP management is an impossible approach, but it looks like something an administrator would love–control everything, create a process for sorting stuff, and then making a system by which product is produced as planned. Get all inventions, choose the best, file patent applications, license rights to industry or investors, cash those royalty checks. But the system does not work. I repeat: This *system* does not work. The system is not tuned to the practices of research-led innovation. The system instead creates opportunities that favor schemers, second- and third-tier investors, and technology trolls and vultures. It doesn’t build strong institutional collaborations, enhance a university’s reputation as a place where new things happen, or a develop a surrounding creative-class culture. It also generally does not break even financially. It’s snake oil that will (so it claims) magically (through a complex process only a select few could possibly understand) provide a new source of significant revenue to save university budgets.
And once there are a bunch of folks who have taken the bait, changed their universities’ policies, and brought in more clever folks selling the snake oil (usually at substantial cost), the pitch becomes prophecy, and it takes a conversion experience–if not a confession–to back out of it all. “Best practice” becomes what everyone else appears to be doing, not what actually works. As one university administrator told me, “You may be right, but if I implemented a program that wasn’t mainstream, then if anything went wrong, I’d be responsible.” It’s attractive, then, to do what most everyone else is doing. Another administrator (I paraphrase): “There’s no reason to reinvent the wheel here–look at what other universities are doing–they have already spent the time to figure it all out.” Best practice is doing what everyone else is doing, just doing it somehow better, for at least bragging rights.
Again, the vision of taking all IP, patenting it, and making money on licensing that IP has failed. The model never has worked systematically, has never delivered as promised. Prior to Bayh-Dole, the successes in licensing mostly started with voluntary decisions by faculty inventors about who to work with, and why, and with what goals. The university administration stayed out of it, other than to keep “commercialization” activities outside of the campus environment, refer faculty to vetted innovation agents, and ask for a financial share based on the “equity” of its contributions to the effort, such as providing extra funding or allowing labs or equipment to be used for the commercial work. Teach and do research, or go make products and money, but don’t mix the academic work and the commercial work. Good advice, still.
No amount of reorganizing a licensing office, renaming it, or hiring new people with “better” credentials is going to turn the model into something that “works.” It also is not a matter of somehow “changing the culture” on campus to “be more entrepreneurial.” Nor is it a matter of making businesses cooperate with the model. They won’t. They think the aggregate-patent-license model sucks, at least when they see its muzzle pointed at them. Nor is it a matter of needing more money from the state for research or technology commercialization or economic development. The “if only we had another million dollars” argument is a good sign that the model isn’t working. That’s the common pitch of a dying venture. But the aggregate-patent-license model doesn’t work with even ten million more dollars, or a hundred million more dollars. It appears the minimum funding to start to have a chance to attract serious attention for the model is somewhere north of a half a billion dollars. At that scale, there’s enough money floating around that almost any model will appear to “work”–for as long as the money is there.
Anything less is, essentially, a non-therapeutic dose and is a matter of flushing resources through a system without achieving anything other than putting on a show of good intentions and covering failures with the usual rationalizations–a new organization, people who appear more “venture-savvy,” people who know rich people, changing the culture, more money, do away with conflict of interest rules, let the university speculate with its own money on new ventures.
Some of the offices claiming the biggest successes with this model have turned out to be frauds–notably the University of Utah and the University of Washington. Utah created 20 companies a year for five years–100 startups–and turned that into pitches for hundreds of millions of state investment on the idea that the economic development arising from research discoveries would be worth it. Just, there wasn’t any. Of the 100 companies, more than 90 were paper companies, shells fabricated by the University. Most had no funding, no employees, no operations, no products, no nothing. When the state audited the program later, they could find only a handful of companies and enough employees to count on one’s fingers and toes. A waste of state money, and a culture of deception.
Even when the model gets lucky and a university sees some multi-million dollar check, often that money comes at the expense of hundreds–if not thousands–of opportunities that over one or two or three decades have been held hostage, abandoned, or destroyed in the effort to find one big deal on which to create a poster of success.
No university technology transfer office using the aggregate-patent-license model is willing to document the path of destruction it has created. You will have to dig to discover the story at your own institution, and likely the people who control that information will fight you to prevent you from seeing the data. They will see it as a matter of their livelihoods, their reputations, their careers. You will have to talk with faculty inventors and non-inventors, with entrepreneurs, with technical leads at companies that made the mistake of reading an academic paper and thinking they could get some help. You will have to examine the situations of the scores–hundreds–of inventions reported and assigned that didn’t go anywhere, not even to the public domain. I’ve been there. I’ve seen the carnage. Trust me, it’s there. Not all technology transfer offices are set up on the aggregate-patent-license model, but most are. Maybe you are lucky.
Despite its failures, the model has benefited from a great deal of positive spin. As a result, many vice provosts for research have been envious of what Utah had claimed to have done–why can’t we start so many companies? What is wrong with our technology transfer office? At the University of Washington, they set out to emulate Utah, and claimed to have started 17 and then 18 companies in a year. It turns out, however, that Washington was only too good at emulating Utah and did it one better, counting companies that weren’t startups, or weren’t founded on university technology, or were moribund. But Washington managed over six years to blow off over $100 million in university and state money advertising itself as a leader in startups. It even won international recognition for its startup programs. But it was all a sham–and worse, because the money spent went into the pockets of people who didn’t produce anything that benefited the state or the university.
Still, you may want all those startups, that support, the hope for a fat royalty check for your school. It’s a nice thought–and the advertised approach sounds rational–but it’s not, and it doesn’t work. The aggregate-patent-license model will deliver on appearances, but it won’t produce new products based on university research and worse, it won’t build public or industry good will and support. You need to break the mental habit of thinking that the model will work, somehow, someday. The model is worse than flawed. It’s an addiction that leads to institutionally self-destructive behaviors, and at a scale that encourages denial rather than constructive change of direction.
But you can change the social narrative, and you can create programs to manage inventions that are productive, that do build good will and collaboration, and which aren’t the sales pitches of clever folks willing to make things look good while wasting your university’s treasure and opportunities. You have been cornered into a prophecy about public benefit that is every bit as silly as hoping to be rescued by a UFO from the planet Clarion. Now it’s time to extract yourself and your institution and rebuild your IP management program on sound principles and with better expectations. There are ways to reduce the complexity of IP management, reduce the cost associated with your program, build strong collaborations, and even benefit financially. I will explain how in my next letter.
Thanks for taking the time to read this letter. I wish you the best as you consider how to build a strong IP management program.
Best,
Gerald