The Innovation Incubator: Technology Transfer at Stanford University
(originally published by Booz & Company)The successful transfer of new technologies from the research laboratory to the commercial sector has many benefits: the creation of wealth, new jobs and new solutions to society's problems. For nearly three decades, Stanford University has been a leader in technology transfer, fostering the growth of northern California's Silicon Valley and the biotechnology industry and providing a model for other research and educational institutions across the country and the globe.
Today, Stanford continues to show the way, providing creative solutions to new challenges as the need for university research becomes even more urgent. Few organizations can afford to finance basic research facilities on the order of a Bell Labs or a Sarnoff center, and with an ever greater percentage of economic growth coming from innovative, but capital-hungry, small companies, access to new technologies is critical. And for the great research universities, income from patent licenses can offset the shrinkage of federal funding.
If the birth of Silicon Valley dates from the meeting of William Hewlett and David Packard in a Stanford classroom in the late 1930's, the modern era of technology transfer begins with the founding of Stanford's Office of Technology Licensing by Niels Reimers in 1970. Of course, inventions had been made and licensed to industry before, but Stanford's process had been ad hoc, alternately managed by patent attorneys and university administrators. It was Mr. Reimers' vision, now widely emulated, to create a dedicated organization that would actively market Stanford's intellectual property.
One has to think back to that mid-Vietnam, pre-Watergate era to imagine how controversial this proposal was; in fact, it passed the university's regulatory processes by only the slimmest of margins. The idea that Stanford and private industry should profit from publicly funded research was cause for concern among faculty and staff. The potential for confiict of interest -- in which scholarly research might be skewed for the possibility of financial reward -- seemed all too real.
Actually, university technology managers, whether at Stanford, Harvard or the University of California, still grapple with these issues. But 28 years of success provide a strong argument for the benefits of proactive technology transfer, and, if potential confiicts cannot be eliminated, they can be managed. The total of Stanford's cumulative income from patent licenses since the creation of the technology licensing office is more than $300 million, but the annual revenues of the companies born at the university total more than $100 billion; their market capitalizations are in the tens of billions, and the jobs created in the hundreds of thousands.
None of that potential was immediately visible when Mr. Reimers, an aerospace engineer by background, arrived at Stanford in May of 1968. Patent license income from 1954 to 1967 "was something like $45,000," said Mr. Reimers, who left Stanford in 1991 to consult to other universities. "I thought more could be done and decided to go in a different direction, to go to a market-based program rather than a legal base," he said. "Working part time, I brought in $55,000 in the first year and was paid $13,500, so Stanford never had to reach into its pocket," he said during a recent interview at his California home.
No one at Stanford said "first we kill all the lawyers," but keeping attorneys out of the licensing practice is the gospel there, as well as at schools like Harvard and the Massachusetts Institute of Technology that have modeled their own technology transfer programs on Stanford's practices. Stanford's Office of Technology Licensing employs no attorneys, and, while outside counsel is available, such oversight is not required if an agreement does not deviate from the university's standard practices of granting no warranty on an invention and total indemnification by the licensee. None do.
"We feel we are a marketing office, not a legal office," said Katherine Ku, who has been director of Stanford's Office of Technology Licensing since 1992. Ms. Ku came to Stanford from a small biotechnology company, and most of the office's associates have similar technical and scientific backgrounds. This kind of training is a stronger tool in negotiating license agreements than a law degree, she said. Successful negotiations require an understanding of the technology.
Jeffrey Labovitz, acting director of technology licensing at the University of California at San Francisco, another program set up by Mr. Reimers, agrees. "A lot of technology-transfer offices are built around patent attorneys, and they lead with the agreement as opposed to the deal, so it's very hard for them to negotiate," he said. "Deal-making is very much a creative endeavor, so the more you know about the variables, the more creative you can be. You need the law, but you don't want to beat companies over the head with policy before you figure out what is the best deal," he said.
Before the deal, however, is the disclosure. Since the passage of the Bayh/Dole Act in 1990, which allowed universities to take title to inventions, Stanford researchers are required to disclose to university management any patentable inventions and turn them over to the Office of Technology Licensing.
Within the office, each invention becomes a case, which is managed from beginning to end by one associate. It is the associate's task to determine whether the invention is indeed patentable and, if so, whether it is licensable and likely to generate sufficient royalties to offset the $10,000 cost of filing a patent. The "magic number," associates say, is $100,000 over the 20-year life of the patent. The same associate devises a marketing strategy, shops the invention to potential licensees and negotiates a deal.
"We believe a licensing agreement is the beginning of a relationship that lasts many years, so we believe in having one person handle it cradle to grave," said Ms. Ku. "You can't make a good licensing deal unless you know the market and the technology, its strengths and weaknesses. If you hand off in the midst of negotiations, there's not that same connection between what the technology is and how much you can ask for it," she said.
Stanford professors were not always obliged to turn over their inventions to the university; indeed, some today seem unaware that the policy changed several years ago. Technology licensing associates say they are often startled when a new invention is published in a trade or academic journal before they have had a chance to file for a patent and seek licensees, but they also say they do not see their role as policing campus research.
"We're facilitators, here to get the technology out there. We're not charged with enforcing policy," said Mary Watanabe, an associate who, like Ms. Ku, came to the Office of Technology Licensing (O.T.L.) from the biotech industry. "I don't think anybody at the university has the idea that we have to clamp down on everything to make sure Stanford makes a buck out of it. There's enough intellectual property to go around," she said. "On the other hand, we don't want things going out the back door; that's a missed opportunity."
In reality, there has long existed a parallel, informal path from the university's biochemistry labs and computer science classes to commercial markets, say the venture capitalists on Sand Hill Road along Stanford's northwest border. Indeed, for every Cisco Systems Inc., whose original products were built with computer networking technology invented at Stanford and used under license, there is a Sun Microsystems Inc. Although Sun stands for Stanford University Network, and three of its four founders were alumni, the company actually used no patentable -- and therefore licensable -- technology. Unlike Cisco, Sun pays Stanford no royalties (although it has been generous in grants to the university).
John L. Hennessy, dean of Stanford's school of engineering, said this bifurcated approach simply refiects the nature of technological invention. "There's two kinds of technologies in the world: stuff that is patentable and broadly applicable and the right thing to do is to give it to O.T.L. Then there's stuff that is more a preliminary proof of a concept. It's not patentable, and the real value is in the people and their understanding of that technology and how it can develop into a useful product," he said. "O.T.L.'s role there is not to get in the way. That's when the right thing to do is to say, 'Godspeed, go do it,'" he said.
For all of Stanford's reputation as the birthplace of startups -- not only Sun and Cisco, but Silicon Graphics Inc., MIPS Computer Systems Inc., Yahoo!, Synteni Inc. and Pangea Systems Inc. -- there is no university policy favoring the creation of new companies as licensees over established companies. University inventors who want to start a company must first turn over their invention to the Office of Technology Licensing, which tries to find the licensee most likely to succeed with the product. Many times, that is not a startup.
"From an institutional standpoint, we have a public service mission," said Ms. Ku. "At Stanford, we feel we need to make sure we pick the best licensee. We ask the companies, including the potential startup, to give us a business plan; then we pick," she said.
Many times, no big companies recognize the invention's value, so it goes to the startup by default. But Ms. Ku said there is still a value to the competitive process. "You have to have that fire in your belly to make a company go, so maybe it's a good thing to put a few hurdles in the way of the entrepreneur," she said. "But in the Valley, everyone wants to do a startup."
For Dari Shalon, who came to Stanford with a grant from the National Science Foundation and the express intent of starting a company, having his invention shopped around was devastating. "They did not get any takers, which was a lucky break for me, so I said I would take it back and pay the patent expenses," he said. But after an article about the invention, a gene chip, appeared in the prestigious Science magazine, Stanford was besieged with licensing requests from pharmaceuticals and biotech companies. "To their credit, they stood behind me," said Dr. Shalon, who subsequently sold his company, Synteni, to Incyte Pharmaceuticals Inc.
Other universities now give preference to the inventor if that individual wants to start a company. These include M.I.T., which enlisted Mr. Reimers to revamp its technology transfer program after the Stanford model in 1986, when Mr. Reimers took a leave from Stanford.
"I think Stanford is more meticulous than we are," said Lita Nelsen, director of M.I.T.'s Technology Licensing Office. "If the inventor has the wish, and it makes even a little bit of sense, we will favor licensing to the startup and giving them a push for two reasons: the vision, the passion, the drive of the person who conceived the technology is not to be discounted in terms of pushing something forward, particularly undeveloped new technology; and, two, the concept that you grab somebody's baby and shop it around [is wrong] -- perhaps because it was Federally funded you should do that, but there is no legal obligation to do that," she said.
Even when Stanford does shop an invention around, the proprietary ownership of something developed with Federal funds can remain a contentious issue. When Thomas Brennan, a Stanford professor, developed a method for automating DNA synthesis, the Office of Technology Licensing first shopped it to major companies working in that field. Finding no takers, it licensed the method to Protogene Laboratories Inc., Dr. Brennan's startup, which was subsequently acquired by Life Sciences Inc. Life Sciences sells DNA, but not the machines to make it.
"That became quite rancorous because the human genome centers all wanted these machines," said Robert Molinari, chief executive officer of the Protogene unit of Life Sciences. "We were willing to sell them DNA, at about a fourth the going price, but the centers wanted the machines in-house so they could get DNA at cost," he said. "But if we had given it to them in that form, we could never have done the corporate deal that drove down the price of DNA."
One reason that inventions wind up in the hands of startups is that larger companies are reluctant to pay for licenses, Ms. Ku said. Before 1990, they often obtained university research free, and they are accustomed to cross-licensing patents among themselves at no cost. "A lot of the big companies aren't licensable," she said. "There's the not-invented-here syndrome. They don't want to pay. They want ownership and control. Or they don't see a product within two years, so it's out of their range," she said.
So Stanford does more and more deals with startups, whether it wants to or not. Working with startups requires some revision of fee strategies. Typically, Stanford negotiates for an upfront licensing fee, which varies from very little to more than $100,000, and a royalty on sales, which can also vary from the low single digits to more than 10 percent. Everything is fiexible, depending on the state of the development of the technology, the size of the market, whether a license is exclusive or non-exclusive.
"In general, Stanford tends to be pretty fiexible and creative and tries to find a win-win," said Brian Kissel, chief executive officer of Paraform Inc., a software startup using technology licensed from the university. "They're enlightened in that they know the research technology is only one part of a success, so they're willing to work with the licensing company," he said. A former Stanford Business School student, Mr. Kissel was an intern in the Office of Technology Licensing before starting Paraform.
Startups are short on cash, so often they want to offer the university equity rather than an up-front payment. But taking equity, particularly in faculty-sponsored companies, has long been a touchy subject for universities, and Stanford is no exception. Before 1981, Stanford had a blanket policy forbidding the taking of equity, and for years afterward it simply was not done. Indeed, Mr. Reimers resigned in 1991 over the insistence by Stanford's president at the time, Donald Kennedy, that he submit all equity deals to administrative scrutiny for potential confiicts of interest. Even today, Stanford has equity in only a couple of dozen of the hundreds of companies it has licensed.
"We try to get equity, but the feeling is equity should be the icing on the cake," said Ms. Ku. In one major biotech deal, with Ariad Pharmaceuticals Inc., the company created a subsidiary focused on the licensed technology and gave Stanford equity in that. "We saw that as perfectly reasonable," said Harvey Berger, Ariad's president and chief executive officer. "We've renegotiated the deal three or four times as the industry has changed," he said. "It's not adversarial. It's truly supportive."
Some other research universities are more conservative than Stanford on this issue, among them the University of California at San Francisco, which, as a public institution, rarely ever takes equity. "Part of the problem of equity is having the staff to manage it," said Mr. Labovitz, the acting director of that institution's technology licensing program. "If you are going to be a shareholder, you need to know how to buy and sell stocks. Early shareholders get diluted out quickly unless they ante up and reinvest, and most universities are not prepared to do that," he said.
Authors
Lawrence M. Fisher, lafish3@attbi.com, covered technology for the New York Times for 15 years and has written for dozens of other publications. Mr. Fisher, who is based in San Francisco, is a recipient of the Hearst Award for investigative journalism. |