Broken Symmetry

What's the New Exit for Patents?

I'm intrigued by the way crowdsourced funding is upsetting the balance of power in financing new films, and companies. The basic trend is toward less and less friction (or what Coase calls "transaction costs") in obtaining financing. At the limit, we have perfectly efficient financing in which ideas receive funding up to the opportunity cost of capital (i.e., up to the next best startup or film that could have been funded). Five years ago, things looked grim thanks to the costs of SOX and the resulting shift of new listings overseas. Technology to the rescue! Gives warm fuzzies to those who tend to be more pessimistic.

But is there less friction for inventors and patentees? Maybe a little, but not the dramatic shift we've seen in private equity financing. And why not? Here things get murky. Looking at equity financing, we can see that at least part of the reason crowdsourced financing has taken off is because of the new "exit" opportunities it offers. At least in relative terms, people prefer what they get from contributing on kickstarter to watching their money go down the tubes in a Facebook IPO. Are there new exit opportunities for inventors and patentees? Nope. It's the same exit opportunities that have always been there. So is there less friction? Maybe a little less. But without the new exit opportunities, it's unclear whether what appears to be reduced friction for inventors and patentees is, in fact, improved allocative efficiency, or whether instead it's an opportunistic redistribution of wealth. We shall see.

Regardless, it's worth asking: if we could dramatically reduce the friction for inventors and patentees, would the public at large view the patent system as more effective or less effective at its goal of promoting innovation? The answer, I think, depends not on whether friction is reduced, but how. It seems to me that there is simply no (lasting) benefit to any stakeholder in the patent system to reducing the friction in reallocations of ownership until the friction in exits can be reduced even further. Universities avoid the problem by limiting the intake to their pipeline, which is wiser than clogging it up with inventory that goes stale. What's the new exit? That's the billion dollar question to be answered by anybody with an interest in seeing the patent system evolve.

Chicago Innovating in Technology Transfer

From the UChicagoTech blog:

One of the biggest challenges to early stage innovation is that it most needs experienced advice when it is least able to attract it, says Alan Thomas, director of UChicagoTech, the University of Chicago’s Office of Technology and Intellectual Property. “A structural problem in Chicago is that at this embryonic point there have been very few resources to turn to,” he says.

...

Start-up ventures are more likely to thrive when advised by multiple mentors with proven skills and experience. “Collectively we are stewarding discoveries flowing from between a billion and a billion and half dollars per year of basic research,” Thomas says.

Thomas and his team have a more clear-eyed view of the landscape than some of the more established tech transfer offices. The potential weak points I see in the plan are (1) that mentors won't have enough skin in the game to do more than provide very early and general advice; and related: (2) adverse selection resulting from more experienced inventors and mentors never opting into the program. Both might be alleviated if they can get a handful of high-profile anchor tenants signed up immediately. Still, it looks promising, and I look forward to hearing more as the plan is implemented.

Square Cofounder an Omitted Coinventor of Patent Apps Owned by Wash. U. Professor?

Techcrunch reports:

In or about February 2009, Mr. McKelvey contacted his friend of 20 years, Mr. Robert Morley, Jr., to help build a prototype magnetic card reader that could be plugged into a cell phone input jack. Mr. Morley, a professor at Washington University, had previous experience working with magnetic card readers.

Over a period of several days, working in Mr. McKelvey’s glass studio and Mr. Morley’s laboratory at Washington University, Mr. McKelvey and Mr. Morley exchanged ideas regarding the design of the prototype card reader.

Square is one of the hottest startups in Silicon Valley, backed by Khosla Ventures GP (and former Facebook and YouTube CFO) Gideon Yu, and cofounded by Twitter cofounder Jack Dorsey. A recent hire was Keith Rabois, an early PayPal exec who had a successful exit at Slide most recently. Keith was also an IP litigator back in the day.

Based only on the allegations in the complaint, it looks to me like Square has a potentially viable claim for co-inventorship and co-ownership here. I got curious to see just how deep the rabbit goes by looking up to see whether Morley had any federal funding for his research at Wash U. His CV doesn't show any recent federal funding, which is lucky for him, since whatever he might get from Square in resolving this dispute could escheat to Wash. U. depending on how Stanford v. Roche gets decided.

Robert Martin on University-Industry Relations

From Inside Higher Ed:

Rather than a commercialization model to explain higher education’s woes, a better analogy is a political campaign, where the focus is to raise all the money you can and spend all the money you raise to create an image that may or may not accurately reflect reality. The conflict of values is between public relations expediency and the academy’s traditional regard for quiet substance. As faculty, we acquiesced in this takeover because it was in our financial interest to comply; more fund-raising meant more money for salaries and perks.

...

An individual faculty member’s ability to influence governance is dependent on his or her academic reputation and alternative employment opportunities. Established scholars have alternatives; they can vote with their feet. Established teachers have little opportunity to vote with their feet.

Very interesting and candid take. [This Robert Martin is of no relation. -MFM]

Ron Conway on the Patent Licensing Business Model

RC: I would never invest in a company that was a basket of patents and then made a business of litigating with people. I think that is ... I want to say "slimy," but I don't want to be that bad. I just think that's a really difficult way to think you're going to add value and make a return on investment.

Via O'Reilly Radar

Unsustainable Growth, Startup Edition

From a post by Ron on things Paul Graham gets wrong:

It is true that VC returns have historically been bimodal, but this is not inherent in the nature of startups, it's because VC's engineer the environment to force the distribution to be bimodal. They do this because they are constrained by the structure of their funds to move money in and out on a particular schedule. They are usually not free to reinvest the proceeds from an early small win into a new startup that might produce another fast small win. Because of this, if a VC cashes out early, that capital can no longer work for them. So they use their influence to force companies to continue to grow past the point where they otherwise would have. This in turn forces the companies to take on additional risks that they otherwise might not have, which causes them in many instances to fail.

This argument is related to an earlier post on e the fundamental speed limit for organic growth. VCs are trying to force second-order phase transitions in markets. Sometimes that works, sometimes not. Organic growth is less risky, but also produces lower internal rates of return for its investors.

But I don't give VCs as much credit (or blame) as Ron for the outcomes. It's a transaction. Entrepreneurs have to take the money that VCs offer. Ultimately, there is too much money in VC relative to the number of scalable startup models that exist at any given moment in time because investors derive more satisfaction from seeing startups grow than they do from arbitraging currency or bond spreads (for example). We don't want to change that, that's a feature, not a bug. But it does suggest that taking other peoples' money to do VC investing is not a good idea for most managers.

These problems are working themselves out organically in the low-capital intensive consumer Internet markets. Paul Graham is partly responsible for that! Once the viability of new funding pipelines has been demonstrated, more capital-intensive industries will probably follow suit. But VCs in the more capital-intensive markets are already more patient, with lower expecations for exits, than the VCs in the lower-capital intensity industries.

PTO Commissioner David Kappos on University Tech Transfer

Available here:

The methodologies universities use to facilitate IP transfer are currently in flux, as traditional models evolve and new best practices emerge.

Universities on the cutting edge of this field realize that the long accepted choice between maximum diffusion of innovation and maximum revenue capture is a false choice. Put simply, universities need not sacrifice their societal obligation to disseminate knowledge in exchange for research funding. In fact, when executed well, diffusion-focused technology transfer generates more net aggregate funding, not less.

Now, royalty generation from university innovation will continue to play a role in the commercialization life cycle. But it will synergize with diffusion-focused tech transfer. And we will see less conflict between the models, because astute university innovators realize that when royalty generation comes into conflict with diffusion, it inhibits the fundamental mission of educating researchers and conducting research to produce breakthroughs that create American jobs and economic opportunity. Astute university innovators are already working to a better end.

Via Tech Transfer Tactics. I couldn't agree more at this level of generality. Now let's see whether the theory gets put into practice with more inventor and industry-friendly institutions.

Zoho Founder on Friedman vs. Hayek and Venture Capital

From Sridhar Vembu:

Now that brings up a philosophical point: what is the purpose of a business? Milton Friedman is unambiguous: The Social Responsibility of Business is to Increase its Profits . From the perspective of a financial investor, by which term I mean anyone investing other people's money, Professor Friedman would be absolutely right; profit maximization is the dharma of the financial investor. Indeed, Professor Friedman was mainly talking about publicly traded companies with broadly diversified ownership. But from a broader capitalistic, libertarian perspective, it is not for Professor Friedman or for that matter, the government, to suggest what the motivations and goals of a capitalist - I am using this word in the sense of a person risking his own money - ought to be.

Maximizing profit can be one goal, but who is to say that ought to be the goal of every capitalist? True libertarians like Mises and Hayek acknowledge that the ultimate ends of each person, certainly including men and women who create companies, are no one's business but their own.

Via peHUB

AUTM Letter to OSTP on Commercialization of University Research

The AUTM letter is as good as the AAU letter was bad. Although the proof-of-concept centers cited in the letter have not been the unmitigated success that the letter suggests, the AUTM at least has acknowledged the problem, its scope, and identified the categories of solutions that seem to be providing limited success. It will be interesting to see what the OSTP does with the feedback it received from this RFI. I'm skeptical that the AUTM will get the additional funding that it wants from the federal government. But perhaps a structural change that would open up new sources of private funding might be in the offing.

Recommended reading.

AAU Responds to OSTP on Commercialization of University Research

The OSTP extended the deadline for responses to its RFI, and at least part of the reason may have been because of this letter, posted by the Tech Transfer Tactics blog.

The introductory remarks, sadly, play right into the obsolete cultural norm I excoriated in my own letter. Specifically, the AAU says "increased commercialization must not come at the expense of our universities' primary education, research, and public service missions." Why does the AAU think that increased commercialization would come at the expense of these things when there is overwhelming evidence that increased interaction with industry promotes all three?

The AAU then states mysteriously that "It is, in fact, these [public service?] missions that have led to universities' most important contributions to economic development, job creation, and technology commercialization." Huh? Unless I misunderstand the argument, the AAU is saying that we should not promote commercialization at the expense of teaching, research, and public service because teaching, research, and public service are the best way to promote commercialization. Do you want more commercialization or not?

The AAU is abundantly clear, however, on the importance of maintaining unencumbered ownership of its IP. The only sentence that appears underlined and in boldface in their entire letter states: "The current legal framework for university technology commercialization, as set forth by the Bayh-Dole Act of 1980 and implementing regulations, is effective and needs to be maintained."

A flat denial that any improvements could be made to the overall Bayh-Dole framework is an embarrassment to their cause. (Note: the letter actually does say that improvements could be made, but without explaining what improvements, by whom, how, or why. I consider this equivalent to denial.) The only reasoning they offer here (namely, that a one-size-fits-all solution isn't likely to work) seems to cut against their stance since Bayh-Dole does apply to every university the same way!

Let's turn to the numbers they cite from the AUTM, of which they seem proud. They're proud, of course, because their benchmark was tech transfer before Bayh-Dole. According to the AAU (citing the AUTM), fewer than 250 patents issued to universities annually before Bayh-Dole, whereas 3,280 issued in 2008.

Number of patents is a poor measure of innovation, but even by that measure this is nothing to boast about. Over a decade ending in the late 1990s, IBM invested about $2 billion per year into R&D and produced approximately 3,000 patents per year. The AUTM, by contrast, had over ten times that much (over $20 billion per year). Nobody expects AUTM members to be as efficient as a public corporation. (Not that that is such a high bar either.) But couldn't the AUTM have at least done 20% of what IBM did?

Part of the explanation for the discrepancy is right in the intro to the letter: "[I]t is important to recognize that the transfer of new knowledge from universities to the public domain and marketplace occur through a wide range of mechanisms, including publication and the movement of people both into and out of our institutions. (My italics.) Yup, that's right, your patents are headed out the back door.

The rest of the letter is a mixed bag. The AAU seems to have felt obligated to offer affirmative suggestions since it rejected the need for a change to Bayh-Dole. Thus, it blames a "focus disproportionately on revenue generation as opposed to moving new knowledge into the public domain for the public good." (What public good does knowledge do that nobody cares about?! Again, the premise is that making money is mutually exclusive with research and teaching.) This argument is incoherent.

The AAU do make a good point about state policies, however. Some state universities face a huge disadvantage to their local private counterparts (what's up, California?). But other states seem to give public universities a leg up on their private competition, if for no other reason than local alumni don't provide a reliable source of revenue for private endowments.

Then, of course, there is the bogeyman of conflicts trotted out again. I have commented on this point many times before, and so won't repeat what I have to say on this again here. Suffice it to say that if you don't see commercialization as mutually exclusive with teaching and research, than conflicts no longer seem fatal. Transparency and accountability are the tools for handling conflicts, not isolation and (feigned) independence.

Still I would be remisss if I didn't praise the AAU for this modicum of sense that creeps in at page 5:

The human capital aspect of technology transfer is often undervalued, but it is a fact that the students our universities educate and the faculty who conduct our research are usually the most effective means through which translation and technology transfer occurs. For example, many of the most successful university-industry interactions are based on the education and training of students who have the specific skills to meet industry needs, or on relationships that top-notch faculty members have with particular companies. These interactions do not center on intellectual property (IP) at all. Therefore, as new models for commercialization are developed, an emphasis should be placed upon fostering new interactions, relationships, and linkages, both formal and informal, that help to facilitate these efforts. Education and training must also be considered to be major components of such models.

Right. So how again would amending Bayh-Dole to give faculty and students more input into how and when IP is commercialized make tech transfer less effective?

UC Davis Chancellor Katehi on the Future of University and Industry Relationships

From an excellent report and interview by Gene Quinn:

That’s where I’m going to when I speak about public/private at the early stage where people who have the knowledge and interest come and make an assessment. They assess and they look at the landscape intellectually speaking. Considering there are 10 institutions they can even do it in a way that gives them an economy of scale....

And they could bundle together, that’s what we’re trying to do with the Department of Commerce. To tell them what’s the point of going for one patent individually and then cutting the cost of one patent instead of having many in the same area together as a bundle. Because then the research the discovery that they’re doing in the department of commerce and the PTO can be the same. You don’t need to do 5 different independent discoveries for 5 similar things. A lot of patents are in areas and they are interrelated and every time there is a whole new wave of patents there is new discovery with a different examiner and that’s why the cost is very high. So somebody could come and bundle them and then send them.

There are many good reasons for universities to start bundling IP by subject matter. I wish I'd been there to hear her talk. Something I didn't see mentioned is that most (at least research) universities already have organizations setup to manage resources over the time horizons necessary for successful industry collaboration. Endowment offices routinely sign 10-year or longer contracts with private equity funds. Universities already have the personnel and resources they need to transform the way they manage industry relationships. With people like Katehi leading, it's just a matter of time.

Welcome Technology Transfer Tactics Readers

You can find the post referenced on Technology Transfer Tactics here. Please take the time to read the whole thing. Parts of the summary were taken out of context.

Commercialization of University Research

[Submitted today to the OSTP. -MFM]

The Office of Science and Technology Policy (OSTP) explains in its Request for Information (RFI) that “the Administration is interested in working with all stakeholders (including universities, companies, Federal research labs, entrepreneurs, investors, and non-profits).” I write to share my experience as a person who has been in the shoes of five out of six of these stakeholders. In addition to having worked as paid research intern at Pacific Northwest National Laboratory, I have worked as a research and teaching assistant at the University of Chicago, a CFO for a consumer Internet venture capital fund in San Francisco, and as an entrepreneur. At present, I practice patent law at the law firm of Drinker Biddle & Reath LLP, representing a mix of small and large clients, from individuals to large multinational companies, in both transactions and litigation relating to intellectual property. Notwithstanding all that, the views expressed here are mine alone, and should not be attributed to any of institutions with which I have been or am associated.

The most important and difficult problem for commercialization of university research is the financing gap between curious exploration and venture capital startups. Because new businesses provide the only sustainable mechanism for the United States to create jobs and wealth, this is also the most important and difficult problem for the United States to solve in its long-term bid to compete in the global economy.

If we want to increase the flow of inventions from our universities and laboratories into commerce, then academics must have more contact with — and inevitably, more dependence upon — people in commerce. No intangible or inanimate thing can create. Although intellectual property rights can and do facilitate multilateral agreements among universities, researchers, and industry, we must never forget this basic fact of nature.

Given that fact, one might ask the following: What is the biggest obstacle to increasing communication between universities and industry? Although there are many practical answers to this question, on my view the ultimate answer is cultural. An obsolete cultural norm is at the core of our difficulty in promoting technology transfer.

This obsolete cultural norm holds that academics must be independent from and disinterested in the consequences of their research. This norm is an ideal, which has never been realized in practice. But even as an ideal it is obsolete.

Louis Menand, in his recent treatise, The Marketplace of Ideas, explains how this cultural norm coevolved with the specialization of academics into departments and professional schools, each of which was self-regulated and, to an approximation, autonomous from the rest of the university. As a result, academics had been sequestered, both from each other and from the rest of society.

More recently, a norm of cross-diciplinarity, or interdisciplinarity, has emerged within the research university. Many academics have come to appreciate the benefits of communicating with and learning from colleagues in other departments and schools. Whereas specialization and sequestration led to stable consensus and uniform doctrine, cross-disciplinarity has brought about a Cambrian explosion in new theory. Although, like anything new at a university, cross-disciplinarity has its critics, they are few.

Transcendent University Culture

My primary purpose in writing is to point out that the next logical step in the evolution of research universities is for this cross-disciplinarity to extend through the equally artificial walls that separate research universities from the rest of our economy. If we can accept that disciplinary boundaries are artificial, and that research can have both disciplinary prejudices and credibility, then we can accept researchers as a part of commerce, and researchers that have both economic interests and academic reputations.

For our system to change, the plurality of stakeholders must recognize and embrace this new conception of what it means to be an academic. To preempt the most common objection, we must not assume that more and more frequent communication with industry entails irreconcilable conflicts of interest. Conflicts of interest can be grave. But prohibiting close relations between academics and entrepreneurs is the clumsiest and costliest way that conflicts can be addressed. More importantly, research universities already have highly evolved mechanisms for dealing with conflicts of interests, which fellow faculty have a strong interest in regulating in order to preserve their own reputations.

I therefore earnestly encourage the OSTP to bear in mind the extensive negative influence of this obsolete cultural norm in evaluating the merits of the various proposals for reform of the technology transfer system. In my remaining remarks, I give a brief answer to Part I of the RFI.

I have nothing to say about Part II of the RFI except that prior to reading this RFI, I had never heard of the NSF Engineering Research Centers Program. Upon looking them up on the Internet, I realize this is because none are located in Silicon Valley, where I have spent the bulk of my professional life. Silicon Valley is also the startup and patent capital of the United States, and perhaps the world. One might justly wonder why, besides a need to appease special interest groups, a POCC had not been located here. Private POCCs, such as Paul Graham’s Y Combinator, seem to have flourished only in the technological fields where intellectual property rights are weakest.

Answers to Questions from Part I: Promising Practices and Successful Models

What do all promising practices and successful models of technology transfer have in common? High-bandwidth feedback loops between the university and industry, promoted by fast, easy negotiations with technology transfer offices over intellectual property rights. To be more specific, the most successful models seem to be subscription-based, nonexclusive licenses of intellectual property, sometimes sweetened for the university with backended payments, such as warrants or reach-through royalties.

If this model works, then why is it not the rule at most universities? The ultimate answer is that given above — namely, that university culture often forces faculty and students to choose between the university and industry. The proximate answer is that most universities handle technology transfers on a case by case basis. With limited administrative resources, inventors avoid the hassle either by ignoring their intellectual property rights or leaving to pursue them without interference by the university.

The latter claim might sound incredible, but even rough estimates show that universities now realize only a fraction of the potential value of their research through licensing or spinoffs of their intellectual property. Top research universities that employ thousands of researchers who publish thousands of papers file only a few hundred invention disclosures with their technology transfer offices. The same universities spend billions on buildings and equipment, pay only modest salaries, and see only about 2 or 3 percent come back through licensing revenue. Comparison against for-profit benchmarks suggests that this percentage could be doubled, tripled, or quadrupled with no additional capital expenditures. Public universities ought to pay attention to these numbers, given the impending credit crisis for many state and local governments.

Aside from an obsolete cultural norm of autonomy, plain envy prevents many universities from implementing subscription-based, nonexclusive licensing of university intellectual property. Whereas the most obvious compensation structure for technology transfer offices would be commissions based on the number of nonexclusive licenses closed, the total compensation paid successful employees would be too high a multiple of the average university employee salary to be tolerated by other employees. Many university endowments have effectively sidestepped this envy problem by physically and legally dissociating themselves from the rest of campus, and by hiring private investment managers. These are viable options, albeit politically unpalatable given the people already employed within most universities. Perhaps only the federal government has the authority necessary to bring about the large-scale radical changes necessary.

We have a host of spectacular successes, both old and new, to strengthen our resolve. The Stanford Electronics Research Laboratory spawned a host of semiconductor companies — so many that the area in which they were (and are) located is now known as Silicon Valley. Nonexclusive licensing of the Cohen-Boyer patent by Stanford’s then-nascent Office of Technology Licensing gave birth to Genentech and a biotechnology industry in both Northern and Southern California. More recently, the MIT Media Lab has given birth to a host of digital technologies, including the electronic ink in display technology used in the Amazon Kindle, and predictive analytics drawn from massive amounts of real-time location-specific data.[1]

What these successes have taught us is that when universities are easy to deal with, investors will risk money, and entrepreneurs time, to complete the customer discovery work necessary to identify commercial opportunities. There is no one-size-fits-all answer to the question of how research can be commercialized. Rather, commercializing research is an iterative process of asking and answering that question over and over again. Nobody bothers to ask when it’s none of their business.

Freedom to Discover and Invent

The lasting legacy of the cultural and institutional reforms prescribed here will be freedom. More freedom for faculty and students to do what they do best, which is discover and invent. Should universities open to industry, faculty and students would remain free to leave the university whenever their passion for business overtakes their passion for research. Streamlining licensing of intellectual property rights will only add to their freedom an ability to diversify into multiple sources of information and potential applications by partnering with a diversity of entrepreneurs and investors.

---

[1] See "Need a Cab? Analysis Shows Where to Find One," New York Times, April 2, 2010 available at http://www.nytimes.com/2010/04/03/nyregion/03icab.html (describing products from MIT Media Lab spinoff Sense Networks).

Why Research Universities Should Start Offering their IP on a Subscription Basis

Over the past several years, I have heard many complaints (from graduate students and faculty on the one hand, and venture capitalists or licensing excecutives for large companies on the other) about the difficulties associated with negotiating licenses or sponsored research agreements. One might naively imagine that such agreements are routine, at least at research universities. Many have a staff of technology transfer employees whose job it is to facilitate these agreements.

But the agreements are not usually so easy to close. In fact, the agreements are so difficult that (at least in places like the Bay Area) the easiest solution for faculty or students intent on seeing their ideas comercialized is to leave the university, either temporarily or permanently, to launch a startup or consult with an established company on a new product or service.

The biggest loser is the university. Or at least that's true if you believe that there is no need to worry that greater university-industry relationships will somehow corrupt university culture. Therein lies the rub. The fact is that many people both within and without the university believe that academics should be independent from the allegedly corrupting influences of commerce.

Nevermind the hypocrisy inherent in holding that view but, at the same time, encouraging faculty to compete for federal grant money that is distributed based on political values that are directly or indirectly influenced by the same corrupting commerce. Focus on the key premise: Does an increased flow of information back and forth between university and industry hinder the growth of knowledge.

The answer to that question must be no. Facts are gathered proactively in response to practical problems. It is these facts that form the basis for academic theory. Academics can collect them on their own, and should be encouraged to do so. But there is no way that academic institutions can keep up with the scale of production of data already ongoing outside the academic system. University-industry collaborations are the future of our global economy, and the university must adapt to the new reality in which the polite fiction of academic independence can no longer be respected. Academic independence -- to the extent that it ever existed in research universities -- has been replaced with dynamic balance, detente.

Which brings us back to the original problem. Suppose the top-tier research universities, whose endowments (not coincidentally!) are often significant portions of the top-tier venture capital funds were to demand that each of those venture capital funds pay an annual subscription for (say all non-biotech or pharma) university intellectual property on a non-exclusive basis. Exclusive deals would remain subject to negotiation. The funds would go primarily to preparation and prosecution of patent applications.

Imagine the collaboration that might result from the lowered transactions costs of collaboration. The alignment of interests between the university and its industry subscribers now permits seamless communication without concern of loss of patent rights or trade secrets. Inventors have better feedback on what, if any, applications are worth filing from industry, rather than from other faculty or from tech transfer office employees alone.

The endowments see reduced expenses -- no more patent prep and pros fees -- and additional income (in whatever is leftover from the subscription fees). Inventors are more likely to stay at the University rather than leave. Only the venture capital funds and companies that aren't paying for the IP that now goes out the back door have something to complain about, and then only initially. In a new equilibrium, a portion of the University becomes a kind of Bell Labs for the economy. Renegotiated employment agreements with star inventors bring billions of capital gains to the University directly, rather than indirectly through the venture capital funds, which charge at least 20% in carried interest plus fees.

But you have to believe that it would be good for faculty and students to talk more with engineers and scientists in industry. For the time being, most people don't.

How do you measure innovation without reference to patents?

Or to put the question differently, how should patents be valued if not by the cost of litigation?

This is the question posed today by venerable Silicon Valley sage Tim O'Reilly:

It seems to me that there ought to be a way to measure the introduction of new products, and rank them by novelty and by widespread acceptance, in some way that reflects a more substantial measure of innovation and its impact on the economy.

There is no single right answer to this question. For earlier ruminations, see Reigniting the Engine of Growth, which was cross-posted to PatentlyO (where some of the comments were very good). But the best answers will share at least this in common: Simplicity

Complex rules are too easy to game. To discourage fraud and bad behavior, rules have to be simple (not just valuation rules, by the way, this is true for any rules).

But innovation is complex. Doesn't that preclude any simple measure? Yes and no. Yes in that any measure will be an approximation. This is true for any complex phenomenon, however. That doesn't stop us from measuring temperature, humidity, barometric pressure, and then forecasting the weather. Weather forecasts are not perfect, but at certain times and in certain places they are useful. Why else would we bother? The small talk alone is not sufficient.

The ultimate answer must be that there are ways to measure innovation. Indeed, everything that for-profit organizations do is complex, even putting aside innovation. Yet we have a highly developed set of rules we follow and measurements we take in tracking the performance of these for-profit organizations. They're called accounting rules and financial statements.

Some people think that you can go even one more step removed and look at the public market price of for-profit organizations, compare that against the number of patents pending or issued, and make predictions about how a company's stock price will move in response to the number of patents. That may work in a few isolated cases, but in general public stock prices are driven up and down by too many other influences. If you stick with the financial statements, you're only increasing your signal to noise ratio

To recap what is obvious for anybody exposed to accounting but unfamiliar to the rest, financial statements include (1) a snapshot in time of the assets, liabilities, and equity of a company (this is called the balance because it has to balance assets with liabilities and equity) and (2) a time-averaged summary of revenues and expenses (this is the income statement, or just P&L for "Profits and Losses"). Actually most also include a cash-flow statement, but we can ignore that for purposes of explaining how innovation might be measured. The cash-flow statement is there so that people can figure out how much of the revenue and expenses were "accounting adjustments," which smear revenue and expenses out over time, rather than actual cash changing hands.

So how can a financial statement give you insight into how innovative a company is? The answer is that whenever a company makes an actual improvement (whether or not that improvement is patented!) in a product or process for making a product, that actual improvement will be reflected in its stream of revenues and expenses.

For example, say PayPal introduces a new feature on April 1. Before April 1, PayPal was spending $0.30 per customer to generate $0.50 in revenue per customer. After April 1, PayPal spends $0.25 per customer and generates the same $0.50 in revenue. The value of the feature is $0.05 per customer. In this example, the PayPal increased its profit margin. That's the most obvious way that innovations add value to companies, and it's the one that most people think of, but it's not the only one!

In some cases, the company may not actually have lower costs or be able to charge higher prices as the result of an innovation. Yet the company may still be more profitable as the result of the innovation. This seems to be the more common case for most incremental innovations, which in fact are the majority of those patented in my experience.

Some of you are scratching your heads, wondering: "How can a company be more profitable if it doesn't charge higher prices or pay less in costs?" The answer is turnover. Consider Company X, which sells 10 widgets a day for $100 after paying $60 to make each them. Total profit per day = 10 x ($100 - $60) = $400.

Now consider Company Y, which sells 100 widgets a day for $100 after paying $60 to make each of them. Total profit per day = 100 ($100 - $60) = $4000.

What's the difference between these companies? Turnover rate. Company Y is selling widgets for the same price at the same cost much faster than Company X. You might think this is easy to do, but in fact it is very difficult.

Now of course if Company X and Company Y have otherwise identical accounting entries, the difference in turnover rates will show up as a difference in earnings on the next financial statement. And if the differential in turnover is sustainable over many accounting periods, it will almost certainly show up as increased earnings.

The problem that readers of public financial statements have in backing out these inventory turnover rates is that many products (both old and new) are combined into a single earnings figure. Moreover, adjusting entries often washout the signal reflected in increased (or decreased!) inventory turnover rates. In practice, at least retail investors have to do a little extra research to figure out whether inventory turnover ratios are increasing or decreasing for particular products, and by how much.

So there you have it, Tim. That's my suggestion for a simple way to measure the value of innovation: Read it off the time-series of revenue and expenses. If you can convince the SEC to get public companies to report this info, we will all be indebted to you.

Not coincidentally, inventory turnover ratios are also the number one way Warren Buffett identifies for companies to raise the return on equity.

Thanks to Paul Kedrosky for the link.

An Obesity Epidemic in Venture Capital?

I'm glad Fred Wilson wrote this so that I don't have to:

In short, since I started investing in the web in '93/'94, I have invested in about 100 software-based web companies. And the success rate of fat companies versus lean companies is stark. I have never, not once, been successful with an investment in a company that raised a boatload of money before it found traction and product market fit with its primary product.

Fred's post is in response to a provocative piece by Ben Horowitz, The Case for the Fat Startup.

Fred does a better job than I could have taking this apart because of his superior experience. But should any entrepreneur be fooled about whether it's in their best interest to take more money from a VC? Of course folks like Ben want you to take more of their (limited partners') money.

I was going to pick apart Ben's post and show how his evidence and reasoning were incomplete. Most importantly, Ben talks about "lean" in a very loose sense that doesn't really get at what folks like Steve Blank mean by the concept. Running lean doesn't mean not spending money. Also, Ben's sample of N=1 doesn't mean anything. Anytime somebody uses a stock price chart as evidence for an argument, be skeptical.

But I hesitated because there are actually points Ben makes that I agree with. First, contrarian thinking is a good start for anybody making capital allocation decisions, including startup CEOs. So I agree with Ben's point that now might be a better time to take money than three years ago, for example. That doesn't mean it's a good idea in an absolute sense though.

I also agree with Ben's emphasis that it is the intellectual property rather than the business that most startups are building:

If you are a high-tech start-up, your value is in your intellectual property. Don’t stare at your spreadsheets so long that you get confused about that.

But a suppressed premise here is that the ultimate exit is acquisition by a larger company. If the ultimate goal were indefinite independence post-IPO, then developing IP without worrying about the IRR might still be the wrong strategy for some startups.

Fred's credibility with me again rises. It's nice to see a VC explaining to entrepreneurs how the game works from the inside. Aside from the guys at Venture Hacks, who are a kind of benchmark for this, I haven't seen that (at least not consistently) from any other VC bloggers.

UPDATE: Ben replies to Fred's comments here. Read Ben's reply for yourself. His reply did more to convince me that Fred has the right side of this one. Ben's reply basically reduces to pointing out that achieving product/market fit is difficult and uncertain (okay, but how does having more VC money early make it less difficult and uncertain?), and that his N=1 sample is still representative of the population. The latter point is ridiculous. Even if he hadn't raised a bunch of money already (making the earlier stage VCs allies in raising more money), Marc had brand name at that point from Netscape that almost nobody has. It didn't matter that Marc was at Ning. Ben's failure to acknowledge this is disappointing.

How do you turn R&D into jobs?

From remarks today by Secretary Locke:

How do universities using federal R&D dollars balance the sometimes competing demands of:

• The pursuit of knowledge for its own sake; and
• Focusing on discoveries that have real potential to spawn new industries, new businesses and new jobs.

What is the best way to integrate universities into broader regional strategies for economic development?

And finally, how do we make it easier to connect entrepreneurs and other business builders with ideas coming out of university research labs?

Our romantic ideal of the university professor as distinterested gadfly -- as Socrates in spectacles -- must give way to a new ideal of ideas in action -- of the university professor as intellectual entrepreneur -- as Feynman at NASA.

Surowiecki on What's Wrong with Venture Capital?

Don't miss his piece:

But it won't be enough for VCs simply to change what they invest in and where they invest. The real problem is not complex: there's too much venture capital, and there are too many venture capitalists, for the industry to be really profitable. The industry as a whole now has about $200 billion under management, more than twice what it did in 1998, and venture funds invested $20 billion to $30 billion a year for most of the past decade. And on the level of individual funds, huge amounts of capital combined with falling startup costs have, in Anderson's words, made funds "muscle bound": a $500 million fund can't make too many small investments, even if that's what would make economic sense, because the partners don't have the time to supervise hundreds of companies. (This is one reason, along with the desire to limit risk, that many VCs have started to wait until later rounds to invest.) In the absence of another bubble, there's no way for new companies to generate profits big enough to provide a reasonable return on $20 billion to $30 billion a year. Kedrosky, for one, argues that for the industry to consistently generate competitive returns, annual investment and money under management need to fall by more than half. And while Wilson describes himself as "very optimistic" about the coming decade, he says that the industry "needs to return to the size and shape it was in the late '80s and early '90s."

Via PE Hub

Too bad Surowiecki didn't elaborate on possible ways to increase the diversity in strategies among VC funds.

Myhrvold on Invention Capital

From HBR:

Outside the pharmaceutical and biotech industries, few companies consider inventing, or producing patented intellectual property, to be their primary mission. Corporate R&D has become mostly “D”: the development of products. Hardly any large corporations have “inventing” as a job category—even though it requires a different mind-set, has different goals, and must be managed differently than research and development positions. At universities and government agencies that fund academic research, patents typically don’t enter into tenure or grant decisions. Published research is rewarded, but invention usually is not. Indeed, these organizations primarily fund blue-sky programs aimed at expanding scientific knowledge. That’s a worthy thing to do. But it’s quite different from invention, which applies scientific knowledge in novel ways to create something useful, something that has economic value.

Via Zura. I have more to say about Menand's arguments in The Marketplace of Ideas, Myhrvold's "invention capital," and Kauffman's proposal for "free agents." But I'm buried in "discovery" -- a classic legal misnomer.

Secretary Locke is Right on the Money

See his speech to PCAST on January 7, 2010 here: Download 25305935-Commerce-Secretary-Locke-s-speech-on-technology-transfer

Thanks to Dan Ballard for the pointer. I disagree with Dan about the necessity of funding. Funding basic science is one of the few legitimate roles of government -- a classic case of dealing with market failure.

Locke is asking exactly the right questions. What happened to Bell Labs? Why aren't new technologies making it out of our universities and into startup companies? As much as I respect venture capitalists and entrepreneurs today, the kinds of ideas getting funding today are not the same caliber as the kinds of ideas that were getting funding during the Sputnik era, when we were the envy of the world.

What do I think we should do? I think IP and inventors need to stick closer together. The incentive to commercialize an idea is strongest for its inventors. Sometimes, but not always, the inventors have the best information about who, what, when, where, and how the work might best be commercialized. The difference between the old Bell Labs and the university and government labs of today is in the comparative difficulty in solving what Hayek calls the "information problem" of matching inventors with entrepreneurs interested in developing work into commercializable form. What used to be done within vertically integrated firms is now coordinated by a decentralized marketplace. Government regulators need to take steps to improve liquidity, either by requiring public disclosure of licensing or assignment terms (reducing the uncertainty of valuations) or by reallocating some of the rights and obligations most likely to incur transactions costs that cannot be covered within the value of the deal.

Why would a VC want an independent inventor defense to patent infringement?

Fred Wilson writes:

But anyone who has spent a significant time in technology based businesses will understand that two groups working completely independently from each other will often solve a problem similarly. One group is not copying or ripping off the other group. They are simply coming to similar conclusions about how to get something done.

In these cases, it makes no sense to protect one group from the other. Nobody has taken anyone's "intellectual property." Both groups should own their inventions outright without having to license technology from the other.

That's not how it works today and as a result, our portfolio companies and entrepreneurs and startups all over this country are paying a very high tax on innovation.

VCs, of course, would benefit handsomely if their portfolio companies never had to pay for intellectual property. But that's not the main reason why a VC would want an independent inventor defense.

The main reason is because the more groups working on the same idea simultaneously, the better the terms for an investor. In effect, weaker intellectual property rights lead to more vigorous competition for VC dollars, and better terms for VCs.

Give you a new perspective on that old saw you hear from VCs: "Ideas are cheap; execution is everything."?

Intellectual Property, Venture Capital, and Startups

To borrow a phrase from Prof. Bainbridge, here are a few things that I think I think about the topics above:

1. Talented engineers and scientists want to earn as much money and recognition for their work as possible.
2. All companies, large or small, require talented engineers and scientists to assign their intellectual property to the company in exchange for their pay.
3. Having assigned away intellectual property rights, talented engineers and scientists must rely on contract rights to earn a slice of the profit from their work.
4. Startup companies write contracts that give talented engineers a slice of equity, which is derived from the value of their work.
5. Larger companies with established products or services write contracts that give talented engineers a salary for so long as the established products or services continue to sell.
6. Larger companies will be most interested in buying or imitating startups that develop substantial improvements to established products or services.
7. Venture capitalists will be most interested in investing in startups that larger companies will be most interested in acquiring.
8. As a result of 1 through 7, talented engineers and scientists will be channeled into working on improvements to already established products or services.

If you agree with all of that thinking, and you agree that our most talented engineers and scientists should be working both on established products and services and on new products and services, then you might also agree with the conclusion that all of this thinking seems to lead:

Both Either a strong IPO market and or strong intellectual property rights for individuals and startups are necessary to encourage talented engineers and scientists to build new products and services.

Is there a flaw in this logic? Consider also the fact that many talented engineers and scientists work at universities, which will never IPO.

UPDATE 12/27/9: Switched the conjunction of IPO market and strong IP rights to disjunction. In principle at least, IP rights could be licensed into large companies from individuals or startups -- if they were mutually respected.

Um. How is this different from any other kind of secured credit?

Financial engineering "news" from Tech Transfer Tactics:

An emerging paradigm called royalty-based financing — applied to early-stage start-ups — may be another, according to Gregory T. Huang of Xconomy Seattle. The concept is simple. Instead of buying equity in a young company, an investor agrees to receive a percentage of the company’s monthly revenues - but the upside is capped at a limit of, say, three to five times his or her investment.

It's a bond, folks. And people have been making venture loans in Silicon Valley for a long time.

What makes a market concentrated?

Does this qualify?

These days, bulge-bracket banks -- notably Goldman, Sachs & Co. and Morgan Stanley -- pretty much have a lock on underwriting new issues. (Goldman Sachs, for example, was co-lead manager on five of the seven IPOs that late-September week, and only Vitacost.com turned to bookrunners who weren't bulge-bracket banks.)

Not only does their "brand" dominate, but they've done a good job of convincing pre-IPO companies of their worth. "The common joke is that no CEO gets fired for hiring Goldman Sachs," says Cowen's Biele.

Domination means that these larger banks can pretty much dictate IPO terms. With very few exceptions, they demand IPOs of $100 million or more "to make it worth their while," one venture capitalist says.

That translates into a minimum valuation of $400 million to $500 million, which excludes a whole universe of high-growth, middle-market concerns.

But the FTC and DOJ by themselves cannot solve this problem. They need to work with SEC to clear the hurdles to smaller companies making offerings. Without an IPO option, these companies have to negotiate with potential acquirers with zero leverage.

Commentary Worth Reading in view of Stanford v. Roche

From an interview of Dr. Gerald Barnett at the excellent ip-advocate.org, which I just discovered:

Bayh-Dole focused on harmonizing agency patent policies and distributing opportunities to promote inventions to universities across the country. It wasn't directly concerned with innovation, but with using the patent system to promote use of inventions, development of collaborative relations with industry, and support for U.S. small business and U.S. labor.

The law itself hasn't constrained innovation - but the way in which universities have implemented practice under it has had mixed outcomes. In response to the law, many universities established Technology Transfer Offices (TTOs) to act as a central reporting agency and to facilitate the selection of inventions that will benefit from use of the patent system. This mixed a regulatory function (reporting) with an invention management function (patents). The result was pressure to monetize patent rights rather than to create an environment that promoted innovation.

The TTO works great for paperwork reporting, but not so well when you have a diversity of inventions, from biotechnology and nanotechnology and software and composters, moving through a single office. Even if the TTO is well-funded and staffed, you still only have a one in a hundred chance of a monetary success worth talking about.

What universities haven't done well is the use of the patent system to build working relationships with industry, support the development of research and early adopting practice communities, or support inventor-based insights into commercialization directions. For this last bit, TTOs find themselves cornered between processes designed to identify inventions with great "commercial potential" and a demand that everyone get equal treatment. As a result, it's difficult, especially in public research universities, to mobilize resources in support of any particular invention - beyond filing a patent application.

Second, and just as important, many universities also established employment policies that compel inventors to agree to assign their inventions to the TTO. This is a subtle but huge transformation in how invention is managed within the context of university research. A hundred years ago, when Cottrell invented the electrostatic precipitator, he was working with a company, at the University of California. He used the patents to found Research Corporation, which then supported research throughout the country and served as an agent for other faculty who had inventions they felt could benefit from Research Corporation's expertise. The system was voluntary, not compulsory. And it gave rise to the idea of the research foundation, as later established as university affiliated at Wisconsin, Kansas, Washington State and elsewhere.

Bayh-Dole gave universities an incentive to wrap voluntary inventor interests with a process-based policy, anchored in contracting regulations. These were then extended for consistency, if not convenience, to cover all instances of invention, not just those made with federal support. As you know, this move to compulsory claims based on employment, rather than voluntary offer based on recognition of resources, has created tremendous issues.

When you think about it, Bayh-Dole set up a classic struggle between bureaucracy and innovation. After all, innovation threatens institutional norms that bureaucracy seeks to codify. Given the opportunity to do so, institutions - agencies and companies included - inadvertently bottleneck and suffocate bright ideas with process. It turns out that consistency, compliance, efficiency and avoidance of risk are the dominant virtues of a process approach to research innovation. Mobilization for a purpose, risk-taking to achieve an outlier advantage, and preparedness for rare events don't count for much.

A very Drucker-esque view of innovation. Here's an important question for tech transfer offices to consider: given the bandwidth issues they face in addressing the many demands on their time, wouldn't now be a good time to consider working with a private intermediary like the Research Corporation?

In economic terms, the Research Corporation acted as a market-maker in those early days of university patenting. The Bayh-Dole Act seemed to have disrupted that model when a handful of universities were successfully able to patent and license innovations without the help of such an intermediary. But what the last decade has shown, I believe, is that entities like the Research Corporation performed an even more important economic function than the provision of professional (even legal) services. Intermediaries like Research Corporation help close feedback loops between universities and industry that are difficult because of the Prisoners's Dilemma of incentives and the asynchronous time horizons of universities and for-profit industry. Without such intermediaries, the tech transfer market is going to become more and more like the patent secondary market, with its litigation-driven norms and valuations. We need more experienced and culturally sensitive intermediaries to foster transactional norms.