The Sheen Gate
You could count on one hand the amount of times I’ve written weeknotes from somewhere that isn’t home. No, two hands, maybe. Trains have hosted writing spaces the most, after the homes of relatives. A café or hotel, perhaps. But this is the first time I’ve written weeknotes from a park.
It’s sunny and the birds are loud, and spring is in the air. My partner is doing her Sunday long run around Richmond park. I’ve had a long walk, talking to my mum on the phone and reading articles, and now I’m sat in the car. I got sick over Thursday and Friday, so didn’t much feel like doing a 27km run. Nor a 3-hour bike ride to rest my ankle. So I’m taking it easy and enjoying the sun instead.
Mowed the lawn yesterday. Had lunch outside. We’re thinking about paving the bottom corner of the garden so we’ve somewhere to sit and sip a soupçon of a summer evening.
Isn’t it nice that the light and warmth are back? It doesn’t get dark until after 6pm. There’s a little more life and happiness in each day.
Focusing on these positive attributes because I’ve got a stinking cold. It started developing on Thursday, took me out of action for half of Friday, and has had me tired, coughing and slathered in Vicks Vaporub all weekend.
Look at this lovey chicken schnitzel I made.

Anyway, shall we move on to how I earned my crust this week?
Service handbook
My main job for the week was finishing off the mapping of our value stream / high-level process / data lifecycle on planning.data.gov.uk. Still not really sure what to call it. High-level process means the most to most people, but that’s also a value stream. And it happens to follow our data lifecycle, so all apply really.
As documented, the mapping is a high-level overview of how we do things. It doesn’t call out specific processes because that would be overkill, but it does model the ‘operating model’, including
- each step of the ‘operating model’
- how long a step lasts (on average)
- aims to achieve at each step
- a more detailed explanation of activities and processes
- who does what
- clearly denoted leading roles and supporting roles
- space for considering interactions with data consumers, Open Digital Planning and other communities
- questions we have, and
- notes on things we need to do or build.
It needs reviewing by the product managers – and psd as service owner, obviously – but I’m quite proud of the work. Before I joined the team, a few of these steps weren’t defined, so it’s nice to leave having shaped them, iterated them, and documented their form now. It’s a manual of how things work that I can leave for the next person.
I’ve only got two weeks left on the team, and I’ve done much more besides this. It’s going to be tough to remember it all but I need to make sure to document some of it as case studies.
Discovery and prototyping
A week of two halves with Sport England. We spent the majority of the week blocked from making progress as we didn’t have access to essential source data and systems, so there was nothing we could do. But that was resolved by Thursday, meaning we could dive into the early phases of prototyping on Friday.
I’m starting with two artefacts: a proof-of-concept for solving an operational bottleneck with Copilot; and a provok-otype1 of what might be possible if a digital & data strategy were put into action. The proof-of-concept is clunky and will introduce friction to a workflow, despite making some aspects better. And the provok-otype is riddled with assumptions, but paints a picture of a possible future.
Both exist to generate discussion and feedback, at this stage. The goal is to open up the space for co-design and learning more about the problem space. I’m fairly sure we can come up with something that’ll solve operational bottlenecks in the short-term, but I’ve already written a piece to share on how it’ll only be a sticking plaster: real change needs a paradigm shift.
Magic patterns
I used Magic Patterns to pull the provok-otype together, as I had a year’s free subscription as a gift. It helped me emulate the Sport England brand just enough to be recognisable, but not well enough to be production-ready. It’s deliberately shonky.
Iterating an interface with it has been…interesting. I’ve had to nudge it a few times by describing the structure I want, in HTML terms, and the styling I want, in CSS terms. There’s one aspect I need to re-design that’s likely going to happen best if I can describe structure and style. You can see why AI-powered prototyping tools get in a mess.
LARPing as a designer
For what feels like the first time, I’m acting more as a designer than a product manager. There’s a couple of product roles I’m filling, but the main heft of this job is design.
I’m not trained in design. I haven’t studied it. Some of it feels intuitive, but a lot of this I’ve picked up from being around good designers. So I wouldn’t say ‘I’m a designer’, at best I’m LARPing as a designer.
What’s giving me joy is the creative aspect of the job. It’s fulfilling. It’s speedy work too: being an information sponge, assimilating context, plotting a line from now into the future, and charting steps along the way, using pictures and words to bring it to life.
I’m happy!
Starting a side project
I’ve made a start on trying to bring to life something which I think should exist and would be useful. I’m trusting my gut for a while, shaping up the vision, principles and constraints before diving in to research and market validation. My goal isn’t to make money or solve a big problem, it’ll serve a niche audience if anything, so there isn’t a great need to validate this upfront. That can come later.
It’s a vehicle for learning too. The majority of things I’ve worked on had the technical direction set, with many of the founding decisions made. Here I’m having to consider all of those, as an amateur, and it’s giving me an appreciation for technical architecture and design.
I won’t say much more about it now, but I’ve already bought the domain name. Watch this space. It may be ready in April.
Running
Despite being fine after the half-marathon last week, my ankle started playing up on Tuesday’s speed work session. My visit a local osteopath on Thursday was well-timed, and she resolved some of the issues. To give it some rest, I switched to cycling in to the co-working space, which is about 30km each day. Decent chunk of zone 2 exercise, especially the traverse up and down South Norwood hill each day.
This cold put shot to any activity the last few days, so I’m hoping to get back on it on Tuesday. The marathon is a month away, so the next two weeks have the most volume and distance. Provided I can get through those, I’ll have the two taper weeks to get enough rest before the big day.
Bookmarks
- Why ATMs didn’t kill bank teller jobs, but the iPhone did, 19 mins. Great long read on the dynamics of the automation of human labour. Although it’s an economics-flavoured article, which is why I probably liked it.
- Where is service design headed?, 6 mins. Lovely piece on how service design is a complex, quiet set of creative acts that shapes trust and human experience inside organisations. Made lots of good highlights on how it deals with slow, hard-to-measure work that AI cannot easily replace.
- Maude on missions: Treasury was always going to be a block, 2 mins. Was hoping this piece would have more on the specifics, but honestly I’ve seen enough of the problems over the last 8 years. Treasury brain combined with departmentalism is a problem.
- What that viral Anthropic jobs chart really means, 7 mins
- Companies House flaw exposed five million directors and enabled company hijacking, 3 mins
- AI is coming for the lanyard class, 5 mins
- Maude on missions: Treasury was always going to be a block, 2 mins
- Revealed: UK’s multibillion AI drive is built on ‘phantom investments’, 10 mins
- Event-sourced Claude Code workflows, 5 mins
- AI Doesn’t Reduce Work—It Intensifies It, 10 mins
- Perplexity drops advertising as it warns it will hurt trust in AI, 3 mins
- How Generative and Agentic AI Shift Concern from Technical Debt to Cognitive Debt, 2 mins
Footnotes
A tiny inconsistency here. Previously I’ve spelled this ‘provoc-otype’ – from the adjective, ‘provocative’ – but here I’ve leaned into the verb, ‘provoke’. The former is likely more correct but possibly less recognisable. A portmanteau that’s not recognisably so just sounds like a mistake. Hmm. ↩
