Rail won the innovation argument and never moved the money

I used to think the gap between a good rail trial and a real rollout was a problem of difficulty. Adoption is hard. Operations resist change. The culture needs time.

I no longer believe that. The culture already agreed. Our tenth anniversary review found a sector that is open, engaged and actively pursuing innovation. Innovation sits in strategies, job titles and conference agendas.

What did not move was the money.

Where the money actually goes

Most of our funding still flows into discovery and trials. Those phases are visible. You can announce a competition. You can open a testbed. You get a photograph and a press line.

Adoption is the phase that changes daily operations, and it has no owner, no budget line and no name attached.

No single body decided this. It is the sum of many sensible decisions taken separately. Committing recurring revenue means defending a budget line, carrying performance risk and working across funding cycles. Announcing a trial means none of that.

So the system keeps funding the search and rarely banks the result.

What the numbers show

A system's chosen metrics tell you what it wants more of. Measure activity, and you get more activity.

Our 2023 report set out independent analysis of forty-five SMEs and start-ups that received innovation grant funding in 2018/19, tracked across the following four years using publicly available Companies House and funding data. Only a handful showed real success four years on in company value, job creation or securing further investment. Across the portfolio the return was negative on grant funding of around £40 million. The companies' own matched funding was not available to the analysis, and the report estimated that total investment could be as high as £60 million to £80 million. No allowance was made for opportunity cost.

We published that in April 2023, in Small Voices Ask for Big Changes. The National Audit Office reached a compatible conclusion in May 2026, working independently and by a different method.

Things did move in that period. The Procurement Act came into force and gave authorities more flexible tools. The NAO found the department had raised the profile of innovation and had identified barriers it is now addressing. What did not move was the allocation.

The cheap fix stayed unimplemented. The expensive problem carried on.

Move the fight to the argument that decides it

There are two arguments here. We think we are having the cultural one, and that is settled. The allocation argument has not been had yet.

The cultural argument being won is not straightforwardly good news. Consensus is a weak position for money. Budgets go to contested proposals with champions who have to fight for them. When a proposal is universally approved, it acquires goodwill, and goodwill does not appear in a business plan. We won the argument and then discovered that winning arguments is not how allocation works.

It did do one useful thing. While the resistance was cultural, every failure could be blamed on attitude. Now the attitude has changed and things stall at the same point, which leaves the structure fully exposed.

There is also a difference in what each position costs. Being enthusiastic about innovation is free. Putting it in a strategy is nearly free. Committing recurring revenue to embed something in operations is expensive, it runs through a process, and that process has no opinions about what the strategy says.

Allocation is a finance and business planning matter. It gets settled in reporting frameworks and business plan templates, the dull machinery almost nobody chooses to lobby.

That is exactly where the useful work sits.

I should apply the same criticism to ourselves. We measured the cultural argument by counting what was easy to count. More events, more attendees, more innovation roles. That is the same measurement failure I have been describing, and we won the argument we knew how to measure.

Hand one named person one measurable duty

The NAO found the department's R&D Board spends its time monitoring spend and moving underspends between programmes rather than deciding where money should go. That is what happens to a body with no budget and no delivery obligation. So this cannot sit with a new committee.

Accountability should land on people who already control budgets. Accounting officers at the Department for Transport and their equivalents in arm's length bodies already carry a duty to secure value for money. That duty should extend to banking innovation outcomes.

Make it measurable with two published figures:

• The ratio of money spent on discovery and trials against money spent on adoption.

• The number of innovations entering service each year.

This costs almost nothing and it gives everyone a basis for scrutiny.

One caveat worth naming: making someone answerable for the ratio only works if you also protect them for the decisions that follow. Asking people to back adoption without stating how much risk the department will carry, and who is responsible when something underperforms, makes the safe choice even safer by comparison. Accountability without cover just adds consequence, and funding another trial stays the safe move.

Repeat the number until it cannot be filed away

I learned this the slow way. In 2023 we published the finding, watched it get politely absorbed, and moved on. That was a mistake. A truth published once is easy to file. A number that keeps reappearing is much harder to set aside, because a series is a track record rather than one organisation's view.

If I were doing it again I would repeat the analysis every two years and publish the same figure each time. Same method, same public data, same single number. The test I would apply is straightforward: if we run it again in 2028 and the picture is unchanged, then the theory was wrong and a different approach is needed. 

That is why this is a coalition task. Trade bodies and research organisations that see the same pattern can carry the same ask and the same measurement until the weight of it lands.

The window may now be open

GBR is still being designed, so nobody in my position can say whether the reporting framework is settled, drafted or not yet started. That is part of the problem. If the financial architecture is being decided now, this is the moment to get one line into it. If it was decided six months ago, we have already missed it and nobody outside would know either way.

Either way, the ask belongs with whoever is designing the financial and business planning architecture, not just the innovation function. The question to put is whether the reporting framework covers what happens after a trial is down-selected, and where the money for that phase actually sits. If the answer is that adoption will be funded when business need is confirmed, it has not been designed. It has been deferred.

One line in the GBR business plan requiring reporting on benefits realised from adopted innovations would do more than another strategy document.

The point is not to punish slow progress. It is to keep one named person answerable for the balance of the spend, so the question cannot quietly disappear again.