I’ve been busy working through ways in which universities construe the Bayh-Dole Act and implement practice. You would think after 30 years, universities would have things pretty well packed down by way of compliance. But instead, it appears that somewhere along the way, they veered from the law to a convenient rationalization of the law, and then everyone copied that into their own practice, so now we have a vast convenient ignorance masking as best practice. How does one go up against that? You don’t just trot in and say, gosh, guys, you’ve got it all wrong. No. There are defenses everywhere, and every political trick in the book ready to throw at anyone who might dispute the franchise.
Just to be clear: the concern here is not in criticizing technology transfer operations at universities. Technology transfer is a challenging activity by any measure. More than 50 systems have to be in place for a full service university IP management office. Not just anyone can work that complexity, not to mention working across incompatible organizational boundaries while dealing with competing and conflicted constituencies with just enough know it all pundits around who only get noticed when they pop off at some part of the effort. That said, no one is saying that innovation practices cannot be done better, or that technology transfer offices have or should aspire to having a monopoly on the discussion. This angers some, I’m sure. They would like to be top of the hill, and the last thing they want is their status challenged by a discussion that might open up different ways of addressing innovation from research–ways that do not necessarily run through their offices, or through their management protocols, or their universities. It is beyond them to consider the point made by Geoffrey Moore in Crossing the Chasm, that conservative buyers look for choice rather than unique products, so as to avoid complete dependency on a monopoly control point. When these buyers have choice, they tend to select the most likely to become dominant, and thereafter they work to create and maintain that dominance, since that is the best way to reduce risk of supply, lower costs to acquire, lower costs to manage, protect one’s own upgrade path, and build one’s reputation for good choices. This works everywhere the product to be acquired is infrastructure and not some distinctive thing by which the buyer’s organization looks to create separation from everyone else. That is, most everything a buyer would think to buy.
For innovation, there’s not an obvious “market”. Hi, I’m looking for innovation, got any? No, it’s more like after innovation happens, one can review the course of events and construct a “breakthrough pathway” or “network”. This takes up Andrew Hargadon’s work in How Innovations Happen, where he explores breakthrough networks built out by successful inventors. A breakthrough network gets dipped into by those looking to advance an innovation, and includes the initiators of the innovation (inventors, inventor’s company or assignee, professionals that do patent and licensing work, management and marketing and brokering professionals), potential imitators and service beneficiaries of the innovation (if they are successful, then so also will we riding their coattails), and recipients of the innovation, adopters, the folks that make an innovation more than an assertion or offer, but something actually taken up and used. No use, no innovation. Just business fiction. If you want to see how this lays out, see David Teece’s work on innovation. (h/t Bob Wooldridge at CMU).
The point is: inventors want to work with the best folks in building a breakthrough network, and they don’t mind if this network is unique, so long as it is reliable, can be created in the time frame needed, and preserves a large portion of the inventor’s expected return on effort. That is, if a given network appears primarily to extract most of the value from the innovation, it’s too expensive a channel. If it exists to generate other income for the network participants, as some incubators and invention management services appear to do, then it’s not sufficiently focused or motivated to manage any particular invention. In fact, in these circumstances, the service wants maximal volume. Not only does this give it the greatest chance of getting lucky with some few “hit” transactions, but this also gives it big numbers to show popularity and with that competition for its services. Finally, it is worth mentioning that volume also provides the best excuse to the would-be inventor whose work doesn’t go anywhere, despite the apparent success of the broker’s portfolio overall. “Not everything has commercial value, and we provided your stuff with the same effort as everyone else’s, except that at some point we have to pull resources to the winners. That’s life!”
Now if you are a university inventor and your technology transfer shop sets up as a broker and your university rams through a policy that says that’s the monopoly house for handling inventions, then you see right away that this makes a mess of any breakthrough network you might already have or want to create. Not possible. All breakthrough networks will go through the technology transfer office’s connections. If you want to contribute yours to theirs, they are good with that, even if you aren’t. In a way, taking those assets–the people you know and who would work for you in their own interest, but may have no interest in volunteering for a university licensing operation–amounts to stripping away much of one’s immediate standing with regard to research.
Another way, if you make an invention in your university research and you have a choice of how to create a breakthrough network, then if you are anyone but a star, celebrity Alpha researcher, you likely will be looking for something efficient that is already top of the game or most likely to become dominant. You will choose that network, or parts of it, or brokers who might help you establish it, and you will work to prove your choice well founded. Even if the venture fails, having built out a network means the next bit in the same space may go better. Further, being engaged in the process, you have personal experience to reason from, to shape the next foray into future planning for innovation.
Such choice in early innovation decisions is critical. It would be interesting to surface the literature on early innovator choices and their impact on the formation of successful breakthrough networks. I expect it will be largely indirect and anecdotal. It’s not something easy to study. Thus, any suggestion that goes this way can always be met with: “there is no evidence for that–show us data, preferably quantitative, show us metrics. Bah!” Of course, if there’s no evidence either way, and if we note that many things in our lives do not reduce to numbers or if so, destroy relationships, delight, and initiative, then we might also observe that lack of evidence is not in itself a meaningful reply. It’s rather more a way of saying: “Go away, you disturb our franchise.”
Back to early choice. If the only choice you have is to report the invention to the technology transfer office, and then it is their deal, and anything you might do is interference, conflict of interest, misplaced enthusiasm, and the like, then the early choice set all but goes away, and with it the incentives to build a breakthrough network, and with it a realization that anything you thought you were doing in the past to this end was wasted time. Even where a technology transfer office commits to working “closely” with inventors, at best the claim is interpreted as “we accept the volunteer efforts of inventors any time they are willing to contribute them, so long as they stay within bounds.”
The reality is: the early decisions are ceded to the technology transfer office and its breakthrough network capacity. If it does not have existing breakthrough networks ready to go, then it will probably do something like write a “non-confidential summary” of the invention (or, have the inventor write it, or re-write it), post this on a web site, and send out letters with it to ten or twenty companies pulled from an industry data base or office mailing list. If there are takers, then the next step is a non-disclosure agreement, which itself may take weeks to negotiate, followed by lots of information exchange and rummaging around. Only speculators license without diligence, and the cost of diligence is largely on the backs of the research personnel, not the technology transfer office. Put another way, the technology transfer office works hard at challenging stuff–like signing an NDA with all its protocols on disclosure and use–and then working out just how to commit the researcher inventor’s time and lab resources to potential licensee review. The guys work hard to make extra work for the inventors. This is “working closely”. For all of that, university royalty schedules for sharing of licensing income make no distinction between the inventors who get to “work closely” and those that don’t. One can imagine, then, that the payoff matrix for effort favors not working closely at all–given that odds do not appear to change much regardless of whether one works closely or not. It will still be largely the office’s ability to form a breakthrough network, not the inventors.
The implication of the standing university model of a single tech transfer office is: it suppresses existing and potential breakthrough networks throughout the university in favor of its own capabilities to form breakthrough networks. Its own operating model does not favor creating these networks in advance, but rather does so in response to a report of invention that it chooses to manage. That is, the model claims to make breakthrough networks only after an invention is accepted for management. There is no channel, no set of high level associations that does not quit, no widespread circulation among the captains of industry, finance, and government, no inside knowledge of the hidden gurus and mavens that know how everything works. It’s all on call. We’ll just pull people in as we need them! And we’ll work closely with inventors! Just a few postings and we will know if an invention has “commercial value”. Do you buy this?
Can you see why it might set up poorly for a research inventor. It has nothing to do with the people (though that does play a role in any model and any operation). It has everything to do with the idea that people are working a model that is sucky from the inventor point of view. It is only a model that administrators who have never invented or tried to work towards a defended, competed for industry value chain or government roadmap could think reasonable.
It is “reasonable”. If what one means is “tidy, orderly, easy to explain to people who don’t know anything”. Worse, reasonable, decent, capable people get hired to make this model work. And they work hard, even heroically, and the best results are 1 license earning $50,000 a year–consulting or workshop income–every 1000 inventions. That is, for a school with 150 invention reports a year, about once every 8 years. We are talking once a decade outcomes across an entire research endeavor at a university working with $200m or $300m of extramural funding. No wonder the public doesn’t buy into the idea that university research matters to them. There are other ways to spin the numbers, of course. That universities really intend to give stuff away–“it’s not about the money” is the mantra. But like the T-ball team told that there will be no score kept, every kid on the team knows “who one” at the end of the game, even if the adults don’t. Of course it’s about the money! Why else would the university IP policy demand ownership, suppress other breakthrough networks, put the emphasis on royalty sharing so that it doesn’t matter if inventors are involved, and report a metric that is fundamentally financial–such as licensing income? More specifically: the reason to make a practice a compulsory, policy driven process is to attempt to capture as much “commercially valuable” property as possible.
It is this very attempt to capture value that destroys much of it. One can salvage a case: by destroying weak value the strong inventions rise to the surface, get the majority of attention, and succeed. The success of these few, even if one a decade, is enough to support the office, claim the effort overall is a success, and attract continued participation in the program. That’s the money argument, the success story argument. That’s what’s put out to the public, and within the university as “education” for inventors. It’s intellectually honest only to the point of realism: “if we don’t say something like this, then our program is toast, and since we believe our program is a good thing, and its objectives are a good thing, then we should do what it takes for our program to succeed–so what if we are a bit optimistic in our reporting? We have no mandate to be dour, or we will send everything into a death spiral of bitterness. Optimism can also be self-fulfilling, so we report our efforts the way we want them to be perceived, even if that appearance hardly matches the experience of most everyone involved.” Or something like that.
One might be impressed at this point at just how starkly limited a conventional effort at technology transfer is. And it is reasonable. It is simple to describe. It is orderly. It sounds right to anyone who doesn’t know. People in the program honestly believe in it (especially if they didn’t know anything before they were hired). And there are a few inventors, after a decade or so, willing to trot out and say how wonderful it has been for them. Researchers report their inventions to an office dedicated to invention management. Professionals there triage the reports and working closely with inventors determine which inventions have commercial potential. For these, patent applications are filed and the office works to identify commercial partners willing to license the patents to create new products, or new companies, that will deliver new products to the public. Any money made in this process is shared generously with inventors and the rest goes to support the valiant public efforts of the university. This is really hard to do, but really important, and this is the model everyone uses, and it will work better if we had more cooperation, less criticism, less bad behavior by inventors and industry alike, and especially more money so that research inventions can be developed from “early stage” to “ready for commercial investment”. That’s all that’s needed. All that’s standing between this process and an outpouring of university originated innovation filling the waiting cups of the American public.
For 30 years we have had this story, and we’ve gotten this handful of inventions. And folks embedded in the model still say: there’s no evidence of under-performance. And as we’ve covered it, there isn’t. There isn’t any evidence at all. There are only assertions. But one can ask whether the public was sold on the idea that for a half a trillion dollars in government research investment alone, universities would come up with a bit more to fill the cup with. There is no question in my mind at least that the model chosen is, for the most part, run at about the level of performance it can achieve–it is run with best in world success–and that success is one paying deal a decade. It’s one in 1000, maybe with a huge performance gain it could be 4 in 1000. That’s a lot of unplaced, over-invested, noise interfering with the signal. One might think, the model would improve with greater selectivity. But that means finding a way to induce more inventors to submit more stuff, only to have a greater percentage of them turned away, or told their stuff has no commercial value, or later that it must be that the stuff never had commercial value because, well, see, no one licensed it, though efforts were made. For more stuff, more success stories, more envy, making the model even simpler, so that any uninformed, unthinking, easily gulled by industry university inventor can understand it.
Alternatively, greater selectivity means greater triage–more staffing to get “better information” from inventors about what they know about the markets, their connections, their ideas for possible uses, and what grants they might be going after next. That is, “work closely” with the inventors early on, taking more of their time, to operate the program for what gets under management with greater efficiency. Triage means big check lists–running to over 30 items–and university committees to review intake and patenting decisions. All this also means overhead, delays, and information gathering and database entry to supplement, or supplant, individual judgment. That is, before trying to build a breakthrough network, second guess the ones you’ve already got. Process rules, not the past or present. This isn’t a cynical development of the theme: it’s the reality of the reasoning that underlies the model that’s in place.
What’s the great fear? That university inventors won’t use the system. They will “go out through the back door”. They will “refuse to disclose their inventions”. They will “be deceived by industry and speculative investors”. Yes, these are all big worries. So the system gets loaded with compliance requirements. Penalties are added. Conflict of interest. Loss of future funding. Exclusion from participation in further research. Misuse of university facilities for private gain. Conversion, misappropriation, fraud, audit. Loss of indemnification by the university. Put that stuff in the “education” program. Haul in malfeasors seeking redemption or at least reinstatement to tell their sad stories of how things went bad by resisting or ignoring or gaming the system.
In this context we might return once more to early decisions, breakthrough networks, and Bayh-Dole. Did Bayh-Dole really require this compulsory, simple, shift from individual breakthrough networks to ones built, appropriated, and controlled by university technology transfer offices? Answer: no. Does Bayh-Dole preclude this state of affairs? Also: no. From the point of view of compliance, things appear well formed. From the point of view of performance, the model is working at its maximum capacity. From the point of view of university inventors, the model may be, for a lot of them anyway, the only thing cooking and better therefore than the alternatives, which are mostly to skip reporting altogether or to quit and try to do something on one’s own. But for those that have made the effort to develop breakthrough networks, or portions of these, the conventional model misses so many opportunities that it takes one’s breath away.
If we look at the break points–triggered only on report of invention; suppresses local breakthrough networks in favor of its own; attempts to build these networks on its own responsive to incoming; pushes for increased volume to defend against non-participation and therefore lowers overall outcomes; focuses on commercial investments to make products and start companies to the exclusion of scientific exchange with industry, custom internal commercial implementation, and formation of standards and other open platforms; and in the end survives, even succeeds, on a couple of deals a decade while the remainder is held because it is cheaper to hold than return–we see where alternatives could arise at each point in the model.
We could imagine becoming involved earlier than invention, could develop projects and cultivate in those projects personal and project based breakthrough networks, we could move assets other than invention rights into those networks to test them, maintain them for future need, we could reward those networks for being ready, we could separate volume out to multiple projects, to agents attached to those projects, without a central office at all to receive everything and process it, we could focus on early decisions and early relationships, where technology might be available with minimal marking, limited ownership claims for internal uses, and no licensing formalities to speak of unless requested by recipients. We could build a structured portfolio where assets were related to one another, not just within a university but across a mission directed research effort, so that, say, a Myelin Research Foundation working with multiple universities could coordinate work without interference or overhead of dealing with six or eight university licensing offices all demanding their processes be met.
These alternatives are not merely imagined. We have tested them out, and other schools have, over the past twenty years, and find they work. Do they work “better” than the conventional standard model? Wrong question. There’s no *data* you see. It is clear, however, that they will work “differently” from the standard model. And fully comply with Bayh-Dole. That’s what’s so intolerable to the status quo in university licensing offices. Bayh-Dole is threatened with attack on all sides. Industry wants it down. Activists construe it as enclosing the great commons of anti-capitalistic academics who otherwise would share and share alike. Entrepreneurs and investors complain about overhead and unanticipated consequences. Politicians and pundits look for leverage in a fight to gain access to power and resources. For all this, it’s ill advised, so the status quo advises, to raise anything about Bayh-Dole, interpretations, alternatives, rule-making, best practices–nothing like this can come up because otherwise, political and social disaster for universities. I am not buying it. Times are changing. The effect of the rise of the conventional model is eating away at, and changing, the character of university research. The conventional model is not alone in doing this. In fact, I doubt it is the lead agent in this. Other factors contribute as well–how research is sourced, and how it is managed and justified in universities in terms of training and economic development rather than for results–or rather, that “results” means “training, economic development, and results” where two out of three is good enough.
Technology transfer offices older than 20 years or at universities with at least $100m (50 or so disclosures a year) in extramural research do break even or better. How the break-even bar gets set depends on the number of patents filed, the success in getting this patent work at least reimbursed by licensees, even if there is no further development, the number of staff involved and how they are compensated, the operating costs, including space and travel and the like, and how many disputes arise that require legal attention. An office may not break even for years–these are not losses, but investment. This is fine, even good, even necessary. Some of the big licensing operations at universities essentially started with a big hit–and before Bayh-Dole–and had nearly two decades of significant income to reinvest in operations and more patent work. Offices that start with a vision and no big hit in hand have to build their program with consistent investment over a decade or more. Technology portfolios from university research are more like cottonwoods than carrots. There is a harvest every 10 or 20 years, not every fall. At least, that’s what happens if one remains committed to the conventional model of licensing patent rights to induce private investment in product development motivated by monopoly rights for which one is willing to pay a royalty on sales. If that can’t happen, why it’s because faculty aren’t reporting inventions as they should, there’s a huge funding gap to take inventions from research to market-investment ready, companies lack the “innovation potential” to recognize new value and adopt innovation from the outside, and more money needs to be invested to upgrade patenting and personnel budgets. Or, it’s one model, and it works for some things, but it’s the wrong model for many, many other things–but it depends on holding those things to operate, and it depends on suppressing alternatives to maximize its own performance and build its own status. While it promotes innovation, it relies on crushing a huge swath of personal interest, connection, responsibility, and initiative. Oh, yeah, there’s no evidence that it does that. You know, we’re still crazy after 30 years.
