License or Company: Why Out-Licensing Goes Nowhere When the Market Isn't There Yet

Portfolio review in the IP department of an industrial group. On the table: a patent family, filed in eleven countries, annual fees running. The technology behind it works in the lab, a prototype exists. It doesn't fit the core business. Then comes the line that comes up in every one of these meetings: “Why don't we just license it out?”

Six months later there is a list of forty potential licensees, forty cover letters, three replies and no deal. The fees keep running.

A license is a good tool. Just not for every technology. Use it for an unfinished technology in an unfinished market, and you're looking for a buyer who doesn't exist yet.

The market for patents is smaller than it looks

Licensing out sounds like a functioning market: here sits the right, over there sits someone who wants to use it, you agree on a fee. The data paints a more sober picture.

Gambardella, Giuri and Luzzi used the PatVal survey of European inventors to examine how many patents actually get licensed. The result: around 11 percent. For another 7 percent, the owner wanted to license but found no partner. So the market could be almost 70 percent bigger than it is. At large companies the gap is especially stark: they would like to license 16 percent of their patents, and 9 percent actually are licensed.

Here's the interesting part: the patents that found no licensee were no worse than the others. The authors read this as a sign of transaction costs. It doesn't fail on quality. It fails because buyer and technology never find each other.

Why unfinished technology finds no buyer

The problem is clearest where technology is systematically licensed at an early stage: at universities. Jensen and Thursby surveyed the technology transfer offices of 62 US research universities. More than three quarters of the licensed inventions were, at the time of licensing, no more than a proof of concept (48 percent, no prototype) or a lab prototype (29 percent). Only 12 percent were ready for commercial use.

And now comes the part that matters for every licensing strategy: for 71 percent of licensed inventions, respondents say commercialization only succeeds if inventor and licensee keep developing it together. The transfer offices don't believe a licensee can take such an embryonic invention to success on its own.

This isn't a university quirk. It's the nature of early technology. A licensee buys a right of use. What it would actually need for an unfinished technology is the knowledge of the people who developed it. And that isn't written in any patent. Agrawal analyzed 124 MIT license agreements: licensing strategies that engage the inventor increase both the likelihood and the extent of commercial success. Even so, a third of the agreements don't engage the inventor at all.

Translated to the corporate world, that means: the license separates the technology from exactly the thing that would make it work. The patent goes to the licensee, the knowledge stays in the lab. Or it retires.

Radical technology tends to end up in a new company

Scott Shane examined 1,397 MIT patents from 1980 to 1996. 26 percent of them were licensed to newly founded companies, the rest to existing firms. The question was: which technologies end up where?

The pattern is clear. More important, more radical and broader patents are more likely to be commercialized through a new company. Shane's explanation: established companies often don't pursue radical opportunities. They don't fit the existing business, and sometimes they even threaten it. So they're left for independent founders.

That is exactly the situation of many technologies in corporate labs. They didn't end up on the shelf because they're weak. They ended up there because they don't fit the core business. And for exactly the same reason, an established licensee won't want them either.

In fairness: most of the time the license is right

Before anyone wants to spin off every patent: the license is the normal case, and for good reason. Of the 8,769 new licenses and options US universities granted in 2021, around 17 percent went to start-ups. The majority went to existing companies.

In a second study, Shane shows that inventions with effective patent protection are licensed mainly to established firms and earn higher fees there. Gans and Stern reach a similar conclusion: when the IP is strong and established companies control the complementary assets, meaning distribution, production and regulatory approval, cooperation beats entering the market yourself. And Lowe points out that the start-up is often only a way station until the invention goes to a company with the right assets.

All of that is true. It just shows when the license is the right tool: when the market exists, when there's a buyer with the right assets and when the technology is close to deployment.

And one more truth belongs here. If you bet on licensing revenue as a business model, you're playing the lottery. In Jensen and Thursby's data, each university's five best inventions brought in 78 percent of licensing revenue. A Brookings analysis of AUTM data shows: over twenty years, on average 87 percent of US universities didn't even cover the costs of their technology transfer.

The rule of thumb: does the market already exist?

All of this boils down to a simple test. Licensing is right when the market already exists, when there's a buyer who brings production and distribution, when the technology is close to commercial use and when the necessary knowledge can be documented.

A company is right when the market still has to be built, when the technology is at prototype level, when the decisive knowledge sits in the developers' heads and when it's radical enough that no incumbent wants to touch it.

In my view, the thinking error is hidden in the word “exploit”. If you license, you sell a right. If you found a company, you build a market. A license monetizes a technology once. A company monetizes a market.

That doesn't mean the license disappears. It just changes roles. The new company becomes the licensee itself, exclusive in its field of use. The source keeps a stake and a royalty, and the developers take their knowledge to where it's needed.

The license has to survive the first funding round

This is where the second classic mistake happens. The license to your own venture gets negotiated as if the venture were a corporation: a high royalty from the first euro, milestone payments, plus a hefty stake on top. Each clause looks reasonable on its own. Together they make the company unfundable.

The independent review of university spin-outs commissioned by the UK government says it very clearly: if you take the top end of the ranges on equity, royalty and milestones alike, you make a bad deal for the spin-out and lower its chances of attracting investors. The UK USIT Guide, which the universities themselves developed as a standard, recommends royalties between 0.5 and under 2 percent of net sales for very early technologies, and licenses designed to protect early cash flow. Reductions for additionally needed third-party licenses are market standard there, with a floor. In one of the Guide's case examples, the royalty only kicks in at cumulative sales of 25 million pounds.

The principle behind it: the royalty is due when there's revenue, not before. Exclusivity is cleanly cut to a field of use. And there's no clause a seed investor has to renegotiate. If you have to reopen the license in the first round, you lose months a young company doesn't have.

The wrong question

Back to the IP department. “Who buys our technology?” is exactly the right question for mature technology in existing markets. For everything else, it's the wrong one.

The right question is: who builds the market in which this technology is worth something? If the answer is “nobody”, even the best license won't help you.

Sources

• Gambardella, Giuri, Luzzi, “The market for patents in Europe”, Research Policy 36(8) (2007): approx. 11% of PatVal patents licensed, willingness to license without a license for another 7%, market “almost 70% larger”, unlicensed patents not of lower quality — Working Paper LEM 2006/04
• Gambardella, “The Market for Patents in Europe” (RIETI presentation, 2008): willingness to license vs. actual licensing, large companies 16% vs. 9%, small 37% vs. 26% — rieti.go.jp
• Jensen, Thursby, “Proofs and Prototypes for Sale: The Licensing of University Inventions”, American Economic Review 91(1) (2001): survey of 62 US universities, 48% proof of concept, 29% lab prototype, 12% ready for use, 71% need inventor involvement, top 5 inventions 78% of revenue — scheller.gatech.edu
• Agrawal, “Engaging the Inventor: Exploring Licensing Strategies for University Inventions and the Role of Latent Knowledge”, Strategic Management Journal 27(1) (2006): 124 MIT license agreements, inventor engagement increases likelihood and extent of success, one third without engagement — doi.org
• Shane, “Technological Opportunities and New Firm Creation”, Management Science 47(2) (2001): 1,397 MIT patents 1980–1996, 26% licensed to new companies, more important, more radical and broader patents more likely commercialized through new companies — stanford.edu
• Shane, “Selling University Technology: Patterns from MIT”, Management Science 48(1) (2002): with effective patent protection, licensing to established firms, higher fees — doi.org
• Gans, Stern, “The product market and the market for 'ideas': commercialization strategies for technology entrepreneurs”, Research Policy 32(2) (2003): strong IP and complementary assets held by the incumbent favor cooperation over market entry — doi.org
• Lowe, “Who Develops a University Invention? The Impact of Tacit Knowledge and Licensing Policies”, Journal of Technology Transfer 31(4) (2006): start-up often a way station until sale to a company with complementary assets — doi.org
• National Science Board, Science & Engineering Indicators 2024, Table INV-3 (AUTM data): 8,769 new licenses and options in 2021, of which 1,480 to start-ups — ncses.nsf.gov
• Valdivia, “University Start-ups: Critical for Improving Technology Transfer”, Brookings (2013): own analysis of AUTM data, over 20 years on average 87% of universities did not cover their transfer costs — brookings.edu
• Tracey, Williamson, “Independent Review of University Spin-out Companies”, Department for Science, Innovation and Technology (2023): high equity, royalty and milestones combined are a bad deal and lower the chances of investment — gov.uk
• TenU et al., “USIT Guide” (2023): royalty ranges by maturity (very early 0.5 to under 2%), cash-flow-friendly licenses, anti-stacking with a floor, case example with a sales threshold of £25 million — bioindustry.org