Beyond the Test Drive: Turning prototypes into sustainable public service reform

The spending review has given government departments some medium-term budget certainty and with it, the opportunity to launch multi-year public service reform programmes.

In this article, Mark Smith and Andrew Laird draw on their collective experience of delivering central government programmes locally to explore the roles of Central Government, Strategic Authorities, and Councils in prototyping and proliferating public service reform.

In recent years, a growing chorus of voices in the public service reform space has lamented that too many promising prototypes have failed to deliver systemic change. Some see these projects as exciting and even transformative in isolation, yet ultimately ineffective because they don’t shift the machinery of government or deliver lasting improvements at scale. But this critique misplaces the blame. Expecting a prototype to change the system is like blaming a car’s starter motor for not powering a cross-country journey. That’s not its job.

Prototypes serve a distinct and valuable role: they help us understand how to do things better—or (more importantly!) how to do better things. They are designed to generate insights, test ideas, and expose the system’s friction points. But we must recognise that prototyping and proliferating those ideas are two fundamentally different tasks, and it is time we designed the system to reflect that distinction.

The role of prototyping in public service reform

Prototypes are, at their core, efficacy tools. They help us test whether an idea works in practice, what impact it has, and where it runs into trouble within the existing system. When designed well, a prototype is not only a small-scale implementation of a new approach, it is also a diagnostic tool. It reveals the points where the system supports innovation—and where it resists it.

This diagnostic function is crucial. Prototypes can show us what kind of evidence is missing, or what kinds of learning are most valuable. They help policymakers, practitioners, and communities reflect more honestly on what is working and why. They provide clarity about what we should change, what we should preserve, and where new energy might be best spent.

Yet despite their value, prototypes are often dismissed as failures because they don’t “scale up” or trigger wholesale transformation. The reality is that a prototype can never be anything more than a starter motor, essential to beginning the journey, but never designed to carry the vehicle forward on its own.

Why proliferation is a different task

Proliferation—by which we mean the spread, adaptation, and embedding of successful ideas across systems—is an altogether different endeavour. It is not about copying prototypes wholesale, nor about enforcing a top-down model of “scaling.” Rather, it is about growing ideas through distributed leadership, learning, and iteration.

Where prototyping is about asking, “Can this work?”, proliferation is about enabling others to ask, “How might this work for us?” It requires a different set of capabilities, a broader mandate, and a more systemic vantage point. Most importantly, it requires a power base that is positioned to influence multiple local actors without being too far removed from their operational realities.

Proliferation is also a political and relational task. It demands the legitimacy to convene, to influence, and to support learning across diverse settings. These are not things prototypes are designed to do. They require scale, positional authority, and long-term commitment—attributes not typically found in innovation teams or short-term pilots. Mutual Ventures work on Radical Place Leadership explores much of the culture and shared endeavour needed to create the enabling environment to support prototyping and then proliferating innovation. 

Differentiating the roles of government

This distinction between prototyping and proliferation has major implications for how we organise public governance. In particular, it suggests that different levels of government should take on different roles in this process.

Central government is well placed to support prototyping. It can fund innovation, provide safe policy and regulatory spaces for experimentation, and support the generation of new evidence. It can create national programmes that test novel ideas in different contexts and gather insights about what works.

But central government is poorly positioned to proliferate those ideas. Its distance from delivery, its need to maintain neutrality across diverse localities, and its often-rigid structures mean it struggles to adapt ideas to different settings or to support local learning processes.

In her landmark book, Radical Help, Hilary Cottam points out the danger of a well-intentioned centre trying to scale good ideas.

“Of course, we want population level change and we need to find ways to make these successful small-scale experiments work nationally …[However] Scale is a linear process – but growth is modular.” - Hilary Cottam, Radical Help

In the book, Hilary provides some examples of where central efforts to “scale” good ideas have not always worked. The experience of the Troubled Families programme is informative. It was born out of a Radical Help experiment called “Life”. A visit from the then Prime Minister David Cameron, led to an instruction to Civil Servants that this was to be rolled out nationally. Instinctively, this should have been the best possible outcome for the fledgling programme.

However, the central government instinct was set it up as a market place with incentivisation, despite protestations from the experiment’s leaders. What should have been grown in a modular way which allowed that local nuance, was packaged up into “a linear programme with outputs that could be measured and controlled” and reported on at a national level.  It became the Troubled Families programme. Many see the Troubled Families programme as a great success, and it definitely did many good things. The point is that it lost much of the essence which made the original experiment work. It could have been better.

So, we need to think about the right level to drive proliferation of a promising prototype. This is where local and regional authorities need to come in.

Local councils, and particularly Mayoral Combined Authorities (MCAs), are better placed to lead proliferation. They sit at a “Goldilocks” level—close enough to delivery to understand context, but broad enough to influence change across systems. They have the positional power to convene stakeholders, align incentives, and support long-term learning. They are also more likely to recognise that spreading an idea is not the same as replicating it. Instead, it’s about adapting it to local needs, capabilities, and opportunities.

The Opportunity of Devolution

Devolution offers a powerful opportunity to put this model into practice. England’s regional devolution arrangements are uniquely placed to bridge the gap between innovation and system change. Strategic Authorities (MCAs) may not have full control over delivery or policy, they occupy a strategic space that allows them to shape how ideas travel and take root.

This “in-between” position, often seen as a weakness, can actually be a strength. With the right support, Strategic Authorities can act as amplifiers of promising ideas—gathering insights from local prototypes, convening actors to explore their implications, and guiding broader system change. They can also steward the long-term relationships needed to turn new ideas into new norms.

No one beats the drum for proper devolution of powers harder than Dr Simon Kaye and colleagues from the Re:State think tank. Their continuing work on what powers should sit where provides a solid basis for getting what power and responsibility should sit where right.

We must be clear about what each part of the system is responsible for. Central government should continue to invest in experimentation without expecting instant systemic results. Local and regional authorities must be empowered to lead the spread and adaptation of ideas. And both must work together to ensure that learning is not lost at the handover point.

Next Steps

There is much to be optimistic about.

Test and Learn (and grow) are the new watchwords of public service reform. We are hopeful that this heralds a new more flexible relationship between the centre, where national policy and funding pots sit, and places and neighbourhoods where the rubber hits the road on any new reform idea.

There are people working in central government who really get this - but it is important that Test, Learn and Grow doesn’t become captured by the centre’s inclination to grip, standardise and measure above all else.

Nick Kimber has set out the principles of the Test, Learn and Grow approach – these are very welcome and appreciate the power the centre has traditionally had over the proliferation of ideas and innovation. The acknowledgement of the need to focus on relationships not transactions in itself is a recognition that the variable and non-standard work on the ground is what makes the difference – not some grand one-size fits all blueprint imposed from the centre. As is the principle of continuous learning as ideas are proliferated. While lessons can be shared nationally, the centre cannot drive it.

Final thoughts…

It is time to rethink how we pursue public service reform. Prototypes are not failures because they don’t change the system; they are only failures when we treat them as the system. They are the starter motors—critical for beginning the journey, but never sufficient to carry us to our destination.

To achieve real, lasting change, we must build a better engine for reform. That means recognising the distinct roles of prototyping and proliferation, and designing our governance structures accordingly. It means positioning central government to support early innovation, and empowering local and regional authorities to lead adaptation and adoption.

Above all, it means treating reform not as a single breakthrough moment, but as a collaborative, multi-stage process—one in which learning and leadership are shared, and in which the power to act is distributed more wisely across the system. If we can get that right, we just might find ourselves not with more failed prototypes, but with a system finally ready to run.

Mark Smith, Independent consultant and visiting Professor of public service innovation at the Policy Evaluation and Research Unit (PERU), Manchester Metropolitan University,

Andrew Laird, Chief Executive, Mutual Ventures

If you are interested in discussing the themes of this article further, we would love to hear from you  - andrew@mutualventures.co.uk