Somewhere in a national laboratory, a result that could become a company is being written up for a journal and then forgotten. This is not a failure of science; it is the default outcome of a system that rewards publication and provides no path to market. Governments across India, the Gulf and Southeast Asia have spent the past decade building that path — research foundations, proof-of-concept funds, technology-transfer offices, scouting agencies — with results that range from transformative to invisible. This article distils what the evidence and four decades of practice say about the difference, and lays out a program design that research councils and universities can commission.
The idea in brief. Research commercialisation is a funnel with a well-documented shape: for every hundred invention disclosures, a handful become companies, and most value arrives through licences and partnerships rather than spin-outs. The system-level lever is not more patents; it is the institutional machinery — technology-transfer capability, inventor incentives, proof-of-concept funding across the “valley of death”, and surrogate entrepreneurs who can carry a technology the scientist cannot. India’s new national research foundation, the Gulf’s university-anchored innovation hubs and Thailand’s science-agency model are all bets on that machinery. A commercialisation program that works scouts broadly, triages ruthlessly, validates with the market before it patents, and treats team formation as its central product.
The gap, measured
The scale of the opportunity is easiest to see in the inputs. India’s gross expenditure on research and development has hovered around 0.6 to 0.7 per cent of GDP for years, according to the Department of Science and Technology — low by the standards of innovation-intensive economies, but on a base large enough to produce one of the world’s biggest scientific workforces and a rapidly rising patent output. The Indian patent office’s annual grants crossed the one-hundred-thousand mark in the 2023–24 fiscal year, a multiple of the figure a few years earlier, following procedural reforms and a sharp rise in domestic filing. The Anusandhan National Research Foundation, established by an Act of Parliament in 2023 with a stated budget of ₹50,000 crore over five years and an explicit mandate to bring industry into research funding, is the institutional response to the gap between that output and the economy.
The chart makes a point that is often lost in national debates: research intensity and commercialisation capability are different things. Israel and South Korea sit at the top of the intensity table because their systems convert research into firms and exports with unusual efficiency, and that conversion capacity is what justifies the spending politically. A country that raises intensity without building conversion capacity buys more journal articles. A country that builds conversion capacity first creates the demand that justifies raising intensity. The Gulf states, whose research bases are young but whose universities — KAUST, Khalifa University, Qatar’s Education City institutions — were built with commercialisation in their charters, are testing the second path at speed.
What Bayh–Dole did, and what it did not do
Every commercialisation policy conversation eventually reaches the United States’ Bayh–Dole Act of 1980, which allowed universities to own and license inventions arising from federally funded research. Its reputation as the origin of American academic entrepreneurship is partly deserved and partly myth, and the distinction matters for anyone copying it.
David Mowery, Richard Nelson, Bhaven Sampat and Arvids Ziedonis examined the evidence in a landmark 2001 paper in Research Policy. Their conclusion was that university patenting and licensing were already rising before the Act, driven by the growth of biomedical research and the emergence of biotechnology, and that Bayh–Dole accelerated and formalised a trend rather than creating it. The Act’s real contribution was institutional: it gave universities a reason to build technology-transfer offices, standardise licensing, and share royalties with inventors — machinery that the science alone had not produced.
Why some universities then generated far more start-ups than others was the question Dante Di Gregorio and Scott Shane asked in a 2003 Research Policy study. Two of their findings translate directly into policy. Institutional eminence and the willingness of a university to take equity in start-ups were associated with more company formation. Counter-intuitively, giving inventors a larger share of licensing royalties was associated with fewer start-ups — because a generous royalty makes licensing to an existing firm more attractive to the scientist than founding a company. The design of incentives shapes the pathway, not merely the volume.
Donald Siegel, David Waldman and Albert Link’s 2003 analysis in Research Policy examined the transfer offices themselves and found large differences in productivity explained by organisational practices — staffing, incentives, and the business skills of the people in the office. Markus Perkmann and colleagues’ 2013 review in Research Policy added a broader frame: academic engagement with industry (collaborative research, consulting, informal exchange) is far more widespread than formal commercialisation, is driven by different motives, and produces much of the economic value that patents are credited with. A program that counts only patents and spin-outs misses most of the traffic between labs and firms.
The four pathways
Commercialisation programs go wrong when they treat the spin-out as the only respectable destination. There are four, and the right one depends on the technology, the market and the people.
| Pathway | Best when | Value captured by | Typical timescale | Common error |
|---|---|---|---|---|
| Licensing to an existing firm | Technology improves an existing product line; incumbent has distribution | Royalties, milestones; inventor share | 1–3 years to first revenue | Over-valuing early-stage IP; exclusive licences that go unused |
| Spin-out company | Platform technology; no natural acquirer; team available | Equity; jobs; follow-on research | 5–10 years to exit | Scientist as CEO by default |
| Contract research and consulting | Industry needs capability, not a product | Revenue; relationships; future licences | Immediate | Undervalued and unreported |
| Open dissemination / standards | Public-good technology; adoption matters more than rent | Impact; reputation; ecosystem growth | Variable | Treated as failure to commercialise |
The licensing pathway carries the most volume in mature systems. The spin-out pathway carries the most attention and, per venture, the most risk and the most upside. A research council’s portfolio should contain all four, and its metrics should count all four; the temptation to report spin-outs alone because they photograph well is the same input-counting error described in article 32.
The funnel per hundred disclosures
The single most useful expectation-setting device for a minister or vice-chancellor is the shape of the commercialisation funnel. The figures below are approximate ratios drawn from the orders of magnitude reported over many years by the Association of University Technology Managers’ licensing survey of North American institutions; they are illustrative, and individual institutions vary widely.
Several things follow from the shape. A program that hopes for ten companies must expect to process on the order of a few hundred disclosures, which means it must scout far beyond the researchers who volunteer. Most patents will never be licensed, so patent counts are a cost metric, not a success metric — filing should follow market validation, not precede it. Licences outnumber start-ups by a wide margin, so a transfer office that cannot license is failing at its main job, whatever its spin-out count. And the funnel’s bottom is small enough that a single well-supported spin-out changes an institution’s numbers, which is why the quality of support for the few matters more than the breadth of support for the many.
Technology readiness and the valley of death
Between a laboratory result and a product lies a stretch that neither research funders nor investors want to pay for: the work of building a prototype that works outside the lab, testing it with real users, and establishing that someone will pay. Public research grants stop at the publication; venture capital starts at the prototype with traction. The gap is the valley of death, and crossing it is the specific job of proof-of-concept funding.
The instruments are well established. India’s Department of Science and Technology runs the NIDHI family of programs — seed support, prototyping grants, an entrepreneur-in-residence scheme — and BIRAC‘s Biotechnology Ignition Grant has become the standard early instrument for life-science ideas, providing grants to researchers and start-ups to establish proof of concept. The UK’s Innovate UK and Research England fund pre-commercial validation through schemes that pay researchers to leave the lab for months and talk to customers. The Gulf’s university innovation centres typically bundle proof-of-concept funding with incubation. The design principles that recur across the successful schemes:
- Small, fast and staged. Amounts sufficient to build and test, released against milestones, with a decision point at each stage.
- Market validation before patent spend. A structured customer-discovery sprint — dozens of conversations with potential buyers — before any significant IP investment.
- Time bought for the scientist. Teaching relief or a funded sabbatical, so that the researcher can actually do the work rather than add it to a full load.
- A named counterpart in industry. A proof-of-concept project with an industrial partner who has agreed to evaluate the result has a customer built in.
Three regional models, side by side
The corridor and its neighbours offer three instructive models, each with different strengths.
India: scouting and scale. The Office of the Principal Scientific Adviser‘s AGNIi mission, housed at Invest India, was built around a simple insight: the binding constraint was not the absence of technology in India’s national laboratories but the absence of anyone whose job was to find it, package it and connect it to industry. Scouting, translation and match-making — done by a dedicated team rather than left to individual scientists — is a model any research council can adopt without waiting for legislation. The national research foundation now provides the funding layer beneath it.
Thailand: the science agency as anchor. The National Science and Technology Development Agency combines national research centres, a science park and incubation on one campus, and has for years run programs that pair its own researchers with industry and with university partners such as Chulalongkorn University. The lesson is institutional proximity: when the lab, the transfer office, the incubator and the industrial partners share a location and a governance structure, the handoffs that kill projects elsewhere happen over coffee.
The Gulf: commercialisation in the charter. Universities founded in the past two decades — KAUST in Saudi Arabia is the most prominent — were designed with entrepreneurship, industry partnership and proof-of-concept funding as founding functions rather than later additions. Their research bases are still growing, but the absence of legacy incentives means the pathway from lab to market is short and expected. The risk is the opposite of India’s: capability ahead of pipeline.
HexGn’s own experience in this space comes from delivering research-commercialisation programs for research councils and university systems in which, cumulatively, more than two hundred scientist teams have been taken through structured translation — from technology audit to market validation to a licence-or-spin-out decision. That experience informs the playbook that follows.
The playbook: an eight-step commercialisation program
- Portfolio scouting. Systematically audit the research base — every lab, every grant, every thesis — for results with plausible application. Scouting is a full-time function with a target: disclosures per hundred researchers per year. Waiting for volunteers yields the same enthusiastic few.
- Triage. Score each disclosure on technical readiness, market pull, freedom to operate and team willingness. Most will stop here, kindly and quickly. The triage panel should include industry and investment voices, not only scientists.
- Market validation sprints. For the shortlist, a six-to-eight-week structured sprint of customer and industry conversations, run with the scientist by a commercial lead. The output is evidence of demand, or evidence of its absence — both valuable.
- IP strategy after validation. Decide what to protect, where and how, in light of what the market said. Provisional filings buy time cheaply; full international filings should follow a licensing or spin-out decision, not precede it.
- Team formation. The central product. Identify whether the scientist will lead, co-found or advise, and recruit the surrogate entrepreneur — the operator who will actually run the venture — from alumni, industry and the program’s own network. Most spin-outs that fail do so here.
- Proof-of-concept funding. Staged grants tied to the sprint’s findings: build the thing the customer said they needed, not the thing the paper described.
- Industry linkage. Every project gets a named industrial counterpart — evaluator, pilot customer, potential licensee or acquirer — engaged before the prototype is finished.
- The decision. At a fixed point, choose the pathway: licence, spin out, partner, or publish and release. Record the decision and its reasons; the record is the program’s learning system.
The program’s KPIs follow the funnel and the four pathways: disclosures scouted, projects triaged in, validation sprints completed with evidence, proof-of-concept projects reaching a decision, licences executed, spin-outs incorporated with an operating team, industry partnerships signed, and — at 36 months — revenue, capital and employment in the resulting ventures.
The program’s economics should be modelled before launch and reported alongside the KPIs. The useful unit costs are cost per validated project (scouting, triage and sprint costs divided by projects reaching a decision), cost per licence executed, and cost per spin-out reaching its first external financing. In mature systems the first of these is small, the second moderate and the third large, and the ratio between them tells a research council where its money is actually going. A program whose cost per validated project is rising while its licence count is flat has a triage problem; one whose spin-out cost is rising has a team-formation problem. Either way, the economics diagnose the design.
A composite case: the research council that commercialised by accident
A composite from several engagements; details altered.
A national research council had, for a decade, funded excellent science and reported patents filed as its commercialisation metric. The count was respectable and rising. A review commissioned by its ministry asked how many of those patents had generated revenue or a company. The answer was a small handful — and, awkwardly, the two most successful spin-outs from the council’s labs had been formed by scientists who had left the system to do it, because the council’s rules had made it easier to leave than to stay.
The redesign began with an audit rather than a fund: a scouting team spent a year visiting every laboratory and produced a portfolio of several hundred candidate results, most of which had never been disclosed because no one had asked. Triage cut the portfolio to a few dozen. Validation sprints — the first time most of the scientists had spoken to a potential customer — eliminated half of those and transformed the rest. The council changed its rules to allow researchers to hold equity and take leave, introduced a surrogate-entrepreneur program that recruited operators from industry, and funded proof-of-concept projects against sprint findings. Four years later its patent count was lower, because it filed less and later, and its licence revenue, spin-out count and industry partnerships were several multiples of the earlier figures. The ministry’s review question was answered in the council’s annual report for the first time.
What could go wrong
- Patents as the target. The office optimises for filings; the funnel fills with unlicensed IP. Antidote: count licences and decisions, not filings.
- The scientist as CEO by default. A brilliant researcher becomes an unhappy, unsuccessful chief executive. Antidote: surrogate-entrepreneur recruitment as a program function, with the scientist as founder, chief scientist or adviser.
- Incentives that block the pathway. Rules on leave, equity and conflicts make staying in the system incompatible with founding. Antidote: policy audit before program launch.
- Proof-of-concept funding without a customer. Prototypes are built for the paper, not the market. Antidote: validation sprint before funding, industrial counterpart during it.
- Transfer offices staffed for compliance. Lawyers and administrators without commercial experience. Antidote: hire deal-makers; measure them on deals.
- Exclusive licences that sit idle. A licensee blocks a technology it never develops. Antidote: diligence milestones and reversion clauses in every licence.
Questions research leaders actually ask
“Should we patent first or validate first?” File a cheap provisional if there is a disclosure risk, then validate. Full filings should follow evidence of demand. Patents are a cost centre until licensed.
“How do we get scientists to participate?” Remove the penalties (leave, equity, time) before adding incentives; go to them with a scouting team rather than a call for proposals; and make the first step small — a sprint, not a start-up.
“How many spin-outs should we expect?” Few, per hundred disclosures. Set targets across all four pathways, and judge the program on the funnel’s conversion at each stage rather than on the bottom alone.
“Where do the entrepreneurs come from?” From deliberate recruitment: alumni in industry, executives seeking a second career, MBA cohorts, and the program’s own accelerator alumni. A research-commercialisation program without a surrogate-entrepreneur pipeline is a patent office.
“How long before we see results?” Licences within two to three years of a functioning program; spin-out revenue and employment within three to five; exits, if any, beyond that. The horizons should be written into the mandate.
Methodology & data notes
National R&D intensity figures are approximate, drawn from World Bank and UNESCO compilations for the latest available years, and vary with definitions and reporting lags; the ranking is the point, not the decimals. The funnel per hundred disclosures is an illustrative set of ratios reflecting orders of magnitude reported in AUTM’s licensing surveys over many years, not a single year’s data. Indian patent-grant and research-foundation figures are as reported by the relevant ministries and the Act’s published text. The composite case combines several engagements with details altered. Companion articles cover campus entrepreneurship (article 34), ecosystem measurement (article 32) and national innovation-agency design (article 39).
References & further reading
- Mowery, D., Nelson, R., Sampat, B. & Ziedonis, A. (2001) — The growth of patenting and licensing by U.S. universities: an assessment of the effects of the Bayh–Dole Act of 1980, Research Policy
- Di Gregorio, D. & Shane, S. (2003) — Why do some universities generate more start-ups than others?, Research Policy
- Siegel, D., Waldman, D. & Link, A. (2003) — Assessing the impact of organizational practices on the relative productivity of university technology transfer offices, Research Policy
- Perkmann, M. et al. (2013) — Academic engagement and commercialisation: A review of the literature on university–industry relations, Research Policy
- AUTM — annual licensing activity survey
- Department of Science and Technology, India — R&D statistics and the NIDHI programs
- Office of the Principal Scientific Adviser, India — AGNIi and technology translation
- BIRAC — Biotechnology Ignition Grant and biotech commercialisation schemes
- UK Research and Innovation — Innovate UK and Research England commercialisation programs
- NSTDA, Thailand — national science and technology agency
- World Bank Open Data — research and development expenditure (% of GDP)
- WIPO — patent statistics and technology-transfer resources
HexGn designs and runs research-commercialisation programs for research councils and universities — scouting, validation sprints, surrogate-entrepreneur recruitment and proof-of-concept management — with more than two hundred scientist teams taken through the process across the corridor.