Five connected areas.

Research commercialization

The research is strong, the patents are filed, and nothing has left the building. That is the problem he is usually called in on, and it is rarely a selling problem. It is a question-selection problem.

When industry is absent from the room where the research question is chosen, the result is a portfolio of answers to questions nobody asked. That is technology push. Market pull starts from a problem somebody already has and will pay to solve. Same laboratory, same people, different starting point, different ending. A patent is evidence that something is new. It is not evidence that anyone wanted it.

Then one decision carries most of the value: what this discovery should become. A license suits a result that slots into a line somebody already runs. A spin-out suits a result that needs a company built around it. The test is not how novel the science is. It is who has to build what is missing. Get it wrong and a license ties up the asset for years, or a spin-out costs a founder a decade.

Dr. Nadir Yildirim has built this function twice from nothing. First at Bursa Technical University's technology transfer company, as founding general manager. Then at Hamad Bin Khalifa University, where the offices, the innovation management companies and the operating model were built as one system: technology transfer and intellectual property, licensing and spin-outs, innovation funds, startup programs and commercialization studios.

If you lead a research office or a funding body, the temptation is to start with the numbers. Patent counts rise without anything underneath them changing. Ask instead whether something reached someone who needed it. Get that right and the figures follow. Chase the figures and figures are all you get. And when a first contract arrives, make it the thing the institution celebrates. People repeat what gets applauded.

Related: A note on industry's share of research.

Innovative entrepreneurship

Innovative entrepreneurship means companies offering technologies and services that are both new and needed. Both halves matter. New without needed is a patent nobody licenses. Needed without new is a business somebody already runs.

Dr. Nadir Yildirim has been on the founder's side of this. Revolution Research Inc., co-founded in Maine in 2014, commercialized environmentally responsible advanced materials, backed by competitive U.S. federal and state innovation grants: the National Science Foundation, the Environmental Protection Agency, the Department of Agriculture and the Maine Technology Institute.

What that side teaches, and the office side forgets, is where a research-based company's first two years actually go. A buyer's procurement cycle. A certification nobody mentioned. A supplier minimum order ten times the pilot volume.

Two things decide whether a company survives its own paperwork. The first is a license it can live with after a Series A, because founders sign the first agreement offered, usually with equity and royalty terms an investor will object to two years later. The second is an early and spoken decision about who runs the company. Not whether the inventor can learn the job. Whether the inventor wants it.

And the part most programs will not say out loud: most research should not become a company. Pushing a result into a spin-out when a licensee would carry it better costs a founder years and a fund its credibility, usually because somebody needs the spin-out count to rise.

He teaches and mentors on this, and holds a graduate certificate in Innovation Engineering from the University of Maine alongside his PhD in sustainable bioproducts.

If you are a researcher weighing a company, the question to answer first is not whether the technology is good enough. It is who the first paying customer is, by name, and what would have to be true for them to sign.

Intellectual property management

Managing university IP is less about filing than about deciding what not to file.

A patent costs money every year in every country where it is kept alive, and a portfolio can grow for a decade while nothing in it is ever used. Three questions do the filtering. Is there a buyer whose product this touches? Is the claim worth working around? Could anyone here detect infringement? A claim nobody can police is a certificate, not a right.

The patent is never the whole asset either. Manufacturing know-how, process tolerances, the data a model was trained on, the reason this formulation works and the five that failed: none of it is in the specification, and all of it is what a licensee needs. Pricing a deal on the patent alone is the most common and most expensive error in university licensing.

Ownership is the other half. Who owns a result produced by a doctoral student funded by a company, in a lab paid for by a national program, using a material supplied under a transfer agreement? The answer sits in four documents that rarely reference each other. Making them agree before the result is worth arguing over is early engagement in its least glamorous form, and it prevents more losses than any later negotiation.

There is also a collision of calendars. A researcher's deadline is the conference. A patent's deadline is the day before the first public disclosure. Most of the daily work of an IP office is being close enough to the research to see a disclosure coming.

Dr. Nadir Yildirim has worked on this from both sides of the table. As a researcher and founder whose results were licensed, and as the office deciding what to file and what to let go.

If you are negotiating with a university, the question that saves the most time is not about royalty rates. It is what, besides the patent, comes with this, and who on your side will still be there in three years to receive it.

Related: A note on what a patent cannot protect.

University–industry–government collaboration

Most of these agreements are signed with goodwill and no mechanism. A memorandum of understanding commits two institutions to being well disposed toward each other. Neither can be held to that.

Early engagement is the mechanism. A partner brought in at the end is a buyer being sold to, and can only say yes or no. A partner brought in before the question is fixed helps shape the answer, and is already most of the way to being a licensee. It is the cheapest intervention in the system and the one most often skipped, because it asks the university to accept a question it did not choose.

Government is the third seat at this table, and usually the one that sets the terms. A matching-fund program decides what a company has to put in. The IP rules attached to a public grant decide what can be licensed afterwards. In many sectors the first serious buyer is a public body, which makes procurement a market-access question rather than an administrative one. A collaboration designed without that third seat gets redesigned later, by someone else.

Three instruments can actually be held to. A sponsored research agreement with the IP terms settled before the work starts. An option agreement that gives a company a defined window and a price to decide in. A consultancy arrangement that is honest about being consultancy. A framework agreement with no project underneath it cannot, and after two years it is quietly embarrassing to both signatories.

Two things do more than any instrument. A pilot in a real setting, which produces the one thing a patent cannot: evidence that somebody wanted this. And proximity, which sounds soft until you count how many partnerships began in a corridor rather than at a signing ceremony.

Dr. Nadir Yildirim runs this interface at Hamad Bin Khalifa University, across AI, biomedical sciences, cybersecurity, advanced materials, health analytics, smart platforms and energy, from campus research to national programs and international partners. He earlier contributed to establishing a university science and technology park in Turkey.

Counting signed agreements is the trap on this side. Agreements are easy to sign and easy to report. What matters is how many of them changed what either side actually does.

If you lead research or strategy at a company, the most useful thing to bring a university is not a budget line. It is one problem you would pay to have solved, stated plainly enough that a researcher can tell you whether they can solve it.

Sustainable innovation ecosystems

An innovation ecosystem is not a set of buildings or a set of programs. It is the set of decisions about who gets to decide.

Dr. Nadir Yildirim's mandate sits above the individual offices and companies. He sets innovation strategy and governance for several universities and organizations at once. The position is deliberate, because the failures that matter are structural. An office with the wrong mandate works hard and produces nothing. A fund rewarded for deployment distorts every company it touches. A startup program rewarded for applications produces applications.

That last pattern repeats everywhere. Any part of a system rewarded for a count will produce the count and quietly stop doing the harder thing the count stood for. The way out is not a better indicator. It is to hold people to the result and let the indicators report it afterwards.

Innovation strategies fail on authority more often than on strategy. An office that needs three signatures to issue an option agreement will not issue option agreements, whatever its plan says. So the first question is not what the strategy is. It is who can say no, and how long they take to say it.

Sustainable means something specific here. A system is sustainable when it survives the departure of the person who built it, the end of a funding cycle, and the failure of its best-known company. Most do not. They run on one person's relationships, and when that person leaves the relationships leave too. Designing for that at the start costs almost nothing. Repairing it afterwards costs years.

Sequencing decides whether this takes years or decades. The instinct is to start with the fund, because capital looks like the constraint. It rarely is. A fund built before there is a disciplined path from a result to a decision will buy whatever arrives, and spend its first years learning that. Build the path first and the fund has something worth choosing between.

If you are designing an innovation system, build it around the result you want rather than the figure you will be asked for. A system built to produce reportable figures will produce them and little else. The track record on the front page of this site was the consequence of working in that order, not the target.

Related: A note on the statistic that built technology transfer.

Working together

Universities, national programs and public bodies building or repairing an innovation system. Companies looking for a research partner on a specific problem. Researchers deciding whether a result should become a license or a company.

Get in touch · Read the field notes