5

Emerging Best Practices in Translating University Research into Innovation

The symposium’s final panel session focused on technology transfer and featured three speakers. Frederic Farina, chief innovation and corporate partnerships officer at the California Institute of Technology (Caltech), and Orin Herskowitz, director of Columbia University’s Columbia Technology Ventures, together provided some context about university technology transfer and described some of the emerging trends and trade-offs in technology transfer that might be relevant to the ERCs. Dean Chang, associate vice president for innovation and entrepreneurship at the University of Maryland, discussed lessons learned from NSF’s Innovation Corps (I-Corps) program. An open discussion moderated by Maxine Savitz followed the panel presentations.

TRENDS IN TECHNOLOGY TRANSFER

One unusual feature of Caltech’s technology transfer office is that it combines all activities related to technology transfer, corporate partnerships, intellectual property, and entrepreneurship under one roof, said Frederic Farina. The reason for creating a one-stop shop was to make it easy for industry to work with the university and meet whatever a company’s needs would be—whether it is licensing a patent or forming a research collaboration. In this arrangement, one person serves as the contact to manage a company’s entire relationship with the university, similar to the way an account manager functions. In this way, a company does not have to navigate a path from the Office of the General Counsel to the Office of Sponsored Research and then to the various deans and department heads, explained Farina.

Caltech, he continued, has a very aggressive patenting strategy. Rather than evaluating inventions, his office errs on the side of caution and applies for patents. “We believe it is difficult to predict success, particularly for the type of fundamental and oftentimes early stage research we do at Caltech,” said Farina. The university, with 300 faculty members and 2,000 students, created 228 invention disclosures in 2015, he noted, and creates an average of some 12 startup companies annually. From Farina’s perspective, the role of his office is to service the faculty, an aspect that he believes many universities overlook. “I think a key aspect of what we do is engage the faculty in the process and support them in the activity,” said Farina.

Orin Herskowitz said Columbia University has a similar mission and approach, which is to help faculty get their technologies from the laboratory to the marketplace and to do so in a way that generates funds to support research and education at the university. Two unusual missions that his office had taken on are to help procure funding to support specific research projects and to educate faculty and students on matters of entrepreneurship. He noted that while his office, too, serves as the single point of contact for industry, it works within what he characterized as a broad and loosely federated, anarchic system of entrepreneurship. “There are many initiatives around entrepreneurship, but not run from the same office,” said Herskowitz.

Columbia, he explained, has approximately 400 faculty members who produce some 400 invention disclosures a year. In 2015, his office issued 117 licenses and options, half of them exclusive and half of those were to startups. In 2015, the university realized $195 million in licensing revenues. He also noted that his office has 40 full-time staff members and 35 graduate student trainees, and that annual patent expenses average between $9 million and $10 million.

As far as how universities fit into the nation’s innovation space, the approximately $800 billion in research funds that went to universities over the past 23 years generated some 320,000 invention disclosures that resulted in approximately 175,000 patent applications and 70,000 awarded patents. As of 2014, said Herskowitz, there are 37,349 active licenses and options for university-generated patents, and 9,261 startups, more than 130 drugs and devices, and more than 300, 000 new jobs have been created from this licensing activity (Figure 5.1). These data, he explained, come from 23 years of Association of University Technology Managers (AUTM) annual licensing surveys. He also noted that the end of the university technology funnel is the beginning of the industry and venture capital funnel, which sees roughly 1 of every 100 pharmaceutical compounds becoming approved drugs and roughly 1 of every 10 venture investments becoming a significant hit.

What this means for universities is that commercial success that generates significant licensing revenues is a rare outcome for most universities. In fact, said Herskowitz, 85 percent of U.S. universities’ technology transfer activities lose money, with the bulk of licensing revenues accruing to a small subset of universities from a small number of “blockbuster” products. In fact, one AUTM Licensing Survey found that less than 40 percent of licenses generated any revenue, and less than 1 percent of licenses generate more than $1 million in annual licensing revenues. Columbia University, for example, has generated more than $3 billion in licensing revenues, but some 90 percent come from four patents, said Herskowitz. One factor that plays a critical role in determining if a licensed invention becomes a successful product, said Farina, is having the inventors involved in the commercialization process. “I don’t know of a single success in which an invention was just handed over to a company without any additional interactions,” he said.

One reason why so few inventions become successful, revenue-generating products is the oft-mentioned “valley of death,” the gap that exists between government and foundation grants that take an invention to the point of early feasibility studies and the latter stages of product development when

industry and venture capital feels comfortable with the risk of taking a product to the market (Figure 5.2). Herskowitz noted that many universities are increasingly focusing on how to fund early feasibility studies, technical validation and prototyping, and early market testing activities that would bridge the valley of death. Farina suggested that the ERCs should start devoting 10 percent of their budgets to these bridging activities.

Another challenge to the successful commercialization of inventions is that it often takes years to license an invention. Herskowitz, citing 32 years of data from Columbia University and similar data from the National Cancer Institute, the Universities of California, and Cornell University, said that only 55 percent of deals are completed within 3 years of first disclosure, and only 85 percent are finalized within 6 years, and even then it can take many more years for a license to generate significant revenues. Farina added that maintaining patents is expensive and that universities are having to make decisions about which inventions they will continue to support. Small universities may not even have the resources to patent inventions, he noted.

With regard to emerging trends in technology transfer, Herskowitz and Farina said that universities, cities, states, and the federal government are showing increased enthusiasm and support for technology commercialization, which is leading to better collaborations, benchmarking, and sharing of best practices. Campus entrepreneurship is showing explosive growth, and the number of discovery outcomes from multidisciplinary and multi-university collaborations are increasing, which increases the complexity of licensing activities. The changing patent landscape will have an undetermined impact on technology transfer, as will the increasing importance of, and experimentation with, so-called alternative licensing approaches such as patent pools and “click-licensing.”

Looking at the future of the ERCs, Herskowitz and Farina listed several trade-offs that need to be considered. For example, should the ERCs focus on engaging established industry partners or on startups? Should they be developing new, strong innovation and commercialization ecosystems or leverage existing strong ecosystems? Should they focus on “blue-sky” research or applied research? Should the focus be on funding basic research or supporting commercialization activities? Should they employ simple metrics or complex metrics for success? Each ERC might chose different combinations of these, said Herskowitz, depending on the ERC’s research focus.

AN EXPERIENCE WITH NSF’S INNOVATION CORPS

The purpose of NSF’s I-Corps is to increase the economic impact of the more than $7 billion of research it funds annually, explained Dean Chang. I-Corps was developed by entrepreneurs and is taught by entrepreneurs, he said, noting that there are seven I-Corps lead universities—University of California, Berkeley; University of Southern California, University of Michigan; University of Texas, Austin; Georgia Tech; University of Maryland; and City University of New York—and a total of 17 I-Corps universities in the national innovation network. According to Chang, virtually every faculty member who has gone through the I-Corps program says afterwards that every university research program should have some level of familiarity with the concepts I-Corps teaches, which prompted him to suggest that every ERC should be an ERIC—an Engineering Research and I-Corps Center.

One of the fundamental parts of the I-Corps program requires participants to interview 100 potential customers for their technology to get a better sense of what customers want from a product. “The number one reason startups fail is that they build something nobody wants,” said Chang. “Since 90 percent of startups fail, you have to wonder why so many of them can build something nobody wants.” Too often, he said, a startup will deliver a product that does not meet 100 percent of potential customers’ needs because the entrepreneur does not fully understand what customers truly need in a new product. In Chang’s case, a company he helped start was within 2 months of going bankrupt because it built a product that met only 75 percent of what its customers wanted.

That does not have to be case, said Chang, because industry will happily reveal what it needs and whether or not a given technology has the potential to deliver a needed solution. “Go out there and talk to customers. Ask them what they want,” said Chang. “The key is to not go out and sell. The goal is to learn.” Herskowitz noted that most of the value that Columbia University gets from its similar Lean LanchPad program is that it gets its inventors out of their laboratories and into the offices of potential customers.

As an example of the power of talking to customers, Chang recounted the story of an experience that Doug Dietz, a GE engineer who was the principal designer of the magnetic resonance imaging (MRI) scanner, had when he went to a hospital to see how his machine was working. There, he witnessed a child screaming as she was being dragged into the MRI suite and realized how terrifying the typical MRI machine was to a child. He responded by meeting with children, parents, technicians, and receptionists and used the information he gained from those interviews to create a child-friendly MRI suite that completely changed the MRI experience for the better for children. However, to get GE managers to buy into the $40,000 in renovation costs associated with creating a child-friendly MRI suite, he also had to develop a compelling business case, which proved to be simple in the end, given that 80 percent of the children who had a traditional MRI experience ultimately had to be sedated by an anesthesiologist, an expensive and time-consuming process. In contrast, the child-friendly MRI suite increased throughput, dramatically shortening the payback period not only for the child-friendly design but the MRI itself.

Chang ended his presentation by posing several questions. “Should we and can we foster more empathetic and impactful engineers? Where does an engineer’s job end—at the end of the technology or during the design-thinking phase of the experience? Does it delve into the business model side of things, where you get the buy-in?” he asked. In his opinion, every innovator on campus should get the training to answer all of those questions. At the University of Maryland, for example, I-Corps is now a campus-wide

initiative with the Office of the President, the provost, his office, the technology transfer office, and the business school. He noted that while the ERCs bring in researchers from multiple scientific and engineering disciplines, they should also be engaging faculty from the social sciences, arts, and business schools.

DISCUSSION

In the ensuing discussion, a symposium participant commented that companies such as Google, Facebook, and Yahoo do not operate on a patent model, and so success should not be measured solely on the basis of patents and licensing activity. Instead, commercialization depends more on customer pull and networking. His hope was that if ERCs become ERICs that they work hard at developing a networking ecosystem that fosters technology transfer. This participant also said that in his opinion, universities are too often reluctant to let the market kill off startups and technology that will not succeed. Herskowitz agreed with both of these comments, adding that while Google actually did begin with a patent license from Stanford University, that license has not driven most of the company’s success. He also noted that technology transfer offices are increasingly providing other services to support entrepreneurship and commercialization activities that go beyond patents. Farina also agreed as well, although he said that a startup’s patent portfolio can serve as a “security blanket” for venture capitalists.