It's just rained. The sky's a kind of pinky grey blue. There's a rainbow, I'm out with the dog and I'm really hoping that Crystal Palace beat Manchester City this evening. (Note from the future: they did not.)
I'm four weeks into a new role as deputy director in the Test, Learn & Grow (TLG) programme in Cabinet Office. We're working with ten places and local authorities on six missions - broadly around neighbourhood health, early years support, SEND, economic inactivity, violence against women and girls, and supporting children's social services. We call each of those teams 'accelerators'.
The work is twofold:
- try some new interventions out, alongside communities and place partners, and see whether those interventions move the needle
- keep learning about the test and learn methodology, what works, what doesn’t, and to share that knowledge more widely and do what we can to change the central government system to make it easier for other teams to ‘do’ test & learn
We're pretty convinced based on the first three demonstrator projects the team did in 2025 that this way of working works. It's taking a load of the standard digital practices - agility, collaboration, user-centred design - and adding in some others like working hand-in-glove with places and communities (because every place is different and potential solutions will vary), mixed methods research and evaluation approaches embedded into teams).
I'm excited about the first part of the work, and I'm totally buzzed about the second. Nick calls the work 'syncretic' - "combining or merging of various distinct beliefs or schools of thought". My layman's interpretation of that is that we are trying to be of three things: Of local or place Of Whitehall and Of the internet
All of those things are signifiers for something else - where ‘place’ is relational practice, human learning systems, participation, or 'the internet' is working in the open, Kanban boards, lean UX, GDS myths and legends. Part of the work is observing and recognising and learning and talking about the signified, to reduce the risk of misinterpretation between us or more widely.
I am pretty sure that my role in the programme is to lead being 'of the internet' - culturally, delivery, practices - while simultaneously learning from others who are more 'of place' or 'of Whitehall', including Nick and my fellow DD Christine. There is only so long you can be successful while going against the grain of your wider operating environment, so I'm very happy to have colleagues who know government inside out rather than to be in a GDS-Mark-Two, sitting outside of the centre as a tool to widen the Overton window of possibility.
(For what it's worth, I believe that often is an entirely necessary thing to do. It's what we did with FF at Sport England. But there always follows a painful process to reintegrate (including the risk of tissue rejection), and things get left behind in the process. In TLG we are lucky enough to have the mandate not to need to go entirely against the grain of Whitehall. We are able to use edge language at the centre and be greeted with hope, not hostility.)
The company that I wanted to start before I ended up starting FF with Eliot was based on the premise that quantitative and qualitative research could be incredibly powerful working together when approaching services and projects. That's only one of the aspects or layers we're integrating to attempt new methods for systems change. Imagine my delight in my second week when I stumbled across the Public Design Evidence Review and the introduction where Georgia Gould puts TLG as next in the timeline for public design, building on the "great legacy of design in UK government that started over a decade ago with the Government Digital Service formed in 2011 and Policy Lab in 2014". At FF we talked about "pushing the practice forwards" almost incessantly. So I'm buzzed that this now gets to be my job.
The major thing that we need to get right is making the work itself sing and speak for itself. That is the only way in which the approach begets legitimacy and earns the right to spread more widely. So most of my time for the foreseeable is going to be working with our SLT and our accelerator teams, helping to get the teams to fly.
On a personal front, I feel at home; that's no mean feat within four weeks. I feel like I'm exactly where I want to be. And this is the first time that work hasn't felt like a bit of an uphill struggle since around autumn 2025.
It's not all roses: my laptop is more locked down than any machine I've ever had. Life without sudo in the Claude Code era is...less fun. I was conscious before starting the job that TLG is adjacent to rather than deep in the digital delivery environments I've spent most of my career in. I wondered if I'd be stepping off the AI wave, unable to get a job in pure digital stuff again at the end of this. On balance I feel mostly fine about that: the intellectual challenge of this work is a better fit for me than that of LLMs and agentic experiences, and I will figure out how to spend time with Kuba, Tom and others doing more of the Fun Computers In Government and bring it all back to TLG. And I'll do it all without installing 1Password, because I'm not allowed.