FST JOURNAL

New ideas can provide a welcome stimulus for creativity and innovation. Equally, some accepted ideas can constrict and frustrate progress.

This is really two comment articles in one, a warm welcome and a much hoped for farewell. First, then, a welcome to ARIA.

The UK has a new research and invention agency. It has been a while in arriving, making its way from idea, to policy, to announcement, to funding, to legislation and now to the appointment of a Chief Executive, Dr Peter Highnam. Those of us who have watched closely, sometimes from the cheap seats, occasionally from the orchestra pit, have certainly experienced the nerves and uncertainty, and have written ARIA off more than once. Yet – here we are.

We have been gifted an incredible and exciting opportunity. ARIA could be many different things: better funded, with a more developed model, with a stronger axis to UKRI and it could have greater assurance on forward funding to make long-term programmatic bets. Despite all of this we have a new agency that has the operational freedom and, in good measure, the resources to make a difference and to position the UK in some key frontier areas of technology.

As a community, in science and technology, as well as in policy, we need now to step forward in a collective effort to ensure ARIA works. This means, it almost goes without saying, a high tolerance of failure when it comes to the choices it makes. It also means, more challengingly, a willingness to accommodate new funding modalities and models that may not be either familiar, or necessarily optimal for the recipients. We will need to show some institutional flexibility.

We can reasonably expect ARIA to be collaborative; independence should not mean disregarding the strengths of the UK research and innovation landscape. We can also expect it to be complementary – bridging gaps – rather than purely going it alone.

We must demand that ARIA sets aside the narrow compartments into which the research and innovation system has been divided (more on this below). These are not, however, major constraints. All this being said, again, welcome ARIA. You represent an important advance for the UK and you should find many friends. I am one.

And farewell…

There are some models that are so simple, so seductive and so often repeated, that even when you know them to be wrong – in fact, worse than wrong, actively misleading – it remains almost impossible wholly to set them aside. My example, and my principal bête noire, is the ‘pipeline’ metaphor, that relates science and research to innovation and commercial exploitation. I wish it gone.

It would not really matter, were this relationship not of such overwhelming importance and of most immediate relevance to how we frame policy.

The idea of the pipeline is that scientists shovel ideas into one end of the pipeline, then innovation actors ready them for industry to commercialise. The process is regular and staged, albeit leaky. It is easy in this model objectively to characterise the state of a technology – its level of readiness.

Everybody has their fixed station on the pipeline. If you are a researcher, you are not suited to the role of innovator – too ‘other worldly’. If you are an innovator, you are unlikely to have the capability to scale and commercially exploit the innovation you have nurtured – too ‘undisciplined’. If you work in industry, your job is to turn the handle and deploy the capital – too staid to innovate, too impatient to undertake research.

Of course, we all know this is wrong. It was never right and has become ‘wronger’. There is no pipeline – it is a jacuzzi. Research scientists based in universities not infrequently prove to be excellent innovators capable of attracting talent and spotting commercial opportunity. Innovative small companies often undertake ground-breaking research, even if they do not choose conventional routes to publication. Larger industrial organisations preserve a substantial reserve of deep technical expertise, the residue of a research capacity, and ‘intrapreneurs’ who can act with agility.

After an extended period of disappointing revenues, we came to a collective conclusion: none of us could sell for toffee.

Systemwire started with a joint programme of research on software development environments. The programme was sprawling, with some core funding and research muscle from the Research Councils but it also drew in students, funded on scholarships, a couple of students funded on discretionary accounts the group had accumulated, a key researcher funded by an overseas national foundation, and so on. Floating in and out of the lab were some former researchers, now working for another spinout. They were role models and later, occasional mentors.

The core concept we came up with was certainly a collective product and gained some interest from the research community. I had rather inflated and grandiose ideas about the potential of the technology as an alternative to ‘the semantic web’. So much as expected. One, rather random, early evening, a former student, then working on an internship in a large investment bank, visited the lab and leaned over the shoulder of a researcher developing an early prototype. ‘That looks just like a big problem we have at the bank’ and, in that moment, we became fintech innovators.

The next day, an extraordinarily able student started work on the banking data standards. Meanwhile, with exceptional support from UCL Business (the TTO), the idea and associated implementation were patented and we started a spinout, i.e. Systemwire. We were joined by an experienced innovator who took on the role of COO and we secured some innovation funding (a Smart Award from, then, BIS). We started to look for opportunities to apply our technology and found, in some investment banks, a small number of early adopters.

Unfortunately, and this is key to the story, our initial technology did not really work! When confronted with the real scale and operating constraints, as well as some important technical features of the data, our approach fell over. So it was back to the proverbial drawing board and a rethink of some of the fundamentals which, by the way, led to some of the best purely scientific work I believe that I have published. We solved the problem and developed a set of optimisations that underpin the technology to this day. The papers are still being cited.

There followed an extended period of successful technical demonstrations and disappointing revenues, after which we came to a collective conclusion: none of us could sell for toffee. We did not really understand the market, did not have the contacts and could not talk the talk. We then merged with a small company, Message Automation, that had a related, though considerably simpler, product but which was run by a team who had the fintech knowledge and address book we so clearly lacked.

The rest was hard work, by that team and by talented former students who joined the company. My colleagues and I drifted in and out, offering advice on no very sound basis. Sales grew organically, and the recurring revenue from licences built up. There were a few early anxious moments during market crises. After a while, we exited to a large US fintech platform.

Complex pathways

Why do I tell this story? Well, first because research and innovation have complex contingent pathways. Good ideas come from all sorts of places and are enriched through a network of different engagements and opportunities. Research feeds innovation but equally innovation feeds research. Technology can be in an uncertain zone between research, application and innovation for lengthy periods. Researchers, innovators and industry expertise can build sustainable partnerships and share financial incentives. Background knowledge is often more important than foreground intellectual property.

All of this is a massive distance from any simple linear rendering of the research and innovation process. Building our funding structures and schemes in accordance with an imaginary pipeline, separating research from innovation and expecting technologies to conform to a simplistic readiness schema, runs counter to building a connected research and innovation system.

An alternative is possible. We should be funding multiple pathways open to the different ways through which technologies arise. We could be building sustained relationships with innovative networks and following technology as it loops back through research, is tested, fails, pivots and is rebuilt. We can apply funding both strategically and tactically in active partnerships. We can be proactive rather than waiting for the non-existent pipeline to deliver commercialisable research. To do this we need greater engagement and a more integrated approach.

In the meantime, farewell and good riddance to the pipeline metaphor.