In Autumn 2025 we started our accelerator projects: small, multidisciplinary teams working alongside places to design, test and scale solutions to complex public service challenges.
Rather than designing a solution centrally and rolling it out, an accelerator brings together people from central and local government to work side by side on a shared challenge. It will include people with a variety of skills such as service design, data, product, policy, analysis and delivery. This way of working means that solutions are designed for and specific to the needs of the communities we’re serving.
The accelerators are also trying to prove that test and learn is a legitimate and effective way of developing public policy close to the people we're serving. We are continuously evaluating, learning and adapting as we go. What we learn during the course of the accelerators directly shapes how we think about public service reform more broadly.
There's no fixed length to an accelerator: the timeline depends on the nature of the challenge and what the team and place partner learn along the way. What's consistent across every accelerator is the approach.
Starting with people, not assumptions
Every accelerator begins the same way: by listening, understanding and building relationships with places and our partners. Before anyone designs a solution, teams spend time understanding the problem from the perspective of the people living with it, whether that's residents, frontline practitioners, schools, the voluntary sector or local delivery partners. In most cases this means spending time in places to understand how communities experience their daily lives such as travelling to work or support services. It could also mean interviews with service users where they outline any issues they are facing, or sitting with local teams to understand how services currently work and where they fall short.
This stage matters because it's easy to wrongly assume you know what the problem is. Test and learn only works if you're honest about what you don't yet know.
Co-designing together
Once a team has a clearer picture of the problem, the next step is co-design: bringing local partners, delivery staff, and often service users themselves into the room to collaborate and generate ideas together. The people who understand the local context and deliver services day to day help shape what gets tested. This is central to how we work.
Rather than committing to a large-scale rollout, accelerators test the riskiest assumptions first on a small scale. This might mean testing whether a particular group can be identified and reached before testing whether an intervention changes behaviour, or trying an idea with a handful of people before expanding it.
Starting small lowers risk. It means that if an assumption turns out to be wrong, the team finds out quickly and cheaply, rather than after investing months of effort and funding into something that doesn't work. It also means we can adapt fast when we learn something unexpected, rather than being locked into a plan that no longer fits the evidence.
Learning, always
We treat learning as continuous, not as a single phase with an endpoint. Every test, whether it succeeds or fails to produce the expected result, generates evidence that shapes what happens next. Teams regularly review findings, adjust their approach, and sometimes go back to first principles when something doesn't work as expected.
This means being honest about what doesn't go to plan. If a test doesn't produce the results we hoped for, that's not a failure of the programme, it's exactly how the model is supposed to work. We'd rather learn that something doesn't work on a small scale, quickly, than find out after a much larger rollout.
While each accelerator is rooted in a specific place, the aim is never just local improvement. Every accelerator is designed to generate insight that can inform national policy and practice and highlight wider, systemic barriers. This could include data-sharing arrangements, funding structures, or ways of working, that get in the way of good public service delivery everywhere.
That's why in Test, Learn, Grow we work hard to build feedback loops with central government departments, feeding what they learn locally back into national policy development and guidance.
Working as partners, not hierarchies
Underpinning all of this is a particular way of working. Accelerator teams don't sit above local partners directing activity, they work alongside them, as equals with different roles and expertise. Ideas, feedback and decisions flow in all directions.
This is important because accelerators are tackling challenges that are usually too complex for any single organisation or level of government to solve alone. Progress depends on genuine partnership between national and local government, alongside the wider ecosystem of place, which includes community organisations and citizens.
Nicola Bailey, Early Years and Family Hubs Service Lead in Manchester said:
“What Test, Learn and Grow does is create space for people to think differently and by turning up today the government is showing the level of commitment they’ve got to changing the outcomes for children and families and for us as a city..
“I always feel that my contribution is valued.”
What's next
Each of our accelerators is at a different stage and each is dealing with a genuinely difficult public service challenge. Over the coming months we'll share more from each place: what we're testing, what's working, and what isn't. We'll also introduce you to some of the people delivering this work on the ground in the places themselves.
If you want to follow along, you can subscribe for email notifications using the sign-up link, or join the Test, Learn and Grow community on LinkedIn.
https://testlearnandgrow.blog.gov.uk/2026/09/09/what-is-an-accelerator/
seen at 10:30, 9 September in Test, Learn and Grow.