bmwelby's blog

In Richard Pope’s essential book Platformland he pitches a new ‘anatomy of public services’ and uses helpful images to dismantle, reconstruct and develop for 2024 ideas that he was first part of putting forward in 2015. Given how well the book does this we should all hope that he’s moving the conversation forward in terms of the UK’s inertia in this area.

Over the weekend he popped up on BlueSky and LinkedIn to ask what people who had seen those original drawings in 2015 thought about them at the time.

Question for digital government people c 2015. Did you see these at the time? What worked / didn't work about them (conceptually, not graphically)

— Richard Pope (@richardpope.org) 2024-11-23T10:14:39.259Z

This will be the nudge I needed to finally finish a series of posts (I don’t think you want it as a single post 🙃) I started writing about my reflections on Government as a Platform so if you want them straight to your inbox then do subscribe.

I remember being treated to a sneak peek of what Richard and others were cooking up and being persuaded pretty quickly that they were absolutely spot on. When it was subsequently presented to GDS more widely at an All Staff (I don’t think any public version exists but Tom Loosemore’s October 2015 Code for America talk captures a lot of it) it was one of those moments at GDS, of which there were more than a few, that made me feel so lucky to be working alongside such inspiring minds.

So perhaps I was too close to the thinking and experiences that produced the visualisation to be an impartial observer – I was already fertile soil for these seeds to land in. I expect people who were further away from the conversation (and the shape it subsequently took) might give more insightful responses to Richard’s immediate question for his purposes in 2024. Nonetheless, here are some thoughts of my own.

---

Mark Foden’s “Gubbins of Government” was another reference point at the time and I thought these images and the ideas they put forward were a great complement to that and spoke of a similar ambition in ways that could land with a not-inside-GDS audience.

---

I liked how the visuals were helpful beyond the ‘whole of government’ perspective. The obvious takeaway is that the data, consent and components layers are about the role of the centre to enable vertical services at the top. I think it’s also a helpful cross section for specific services to think about as well. Any end to end service trying to meet a whole need is going to do that through a composite of elements (micro-services if you will), that sit on top of a service-wide approach to data and identity.

---

At the OECD I wrote the Government as a Platform pillar of the Digital Government Policy Framework. That exercise was really helpful for me in marshalling my thinking as a partial retrospective on my work as the Lead Product Manager for Government as a Platform in the UK (more on this in those upcoming blog posts).

One of the biggest things I felt when it came to writing it up was the need to take a wider-angle lens on how you enable and equip teams to move quickly, at scale, and with quality. This 2015 visualisation sits alongside a whole host of contextual assumptions about things we at GDS didn’t exactly take for granted but which we saw as self-evidently important: fixing procurement, controlling spend, assuring quality, building capability, etc.

So as powerful as I think the visuals are and were, I think they only tell a partial, more technical story, about what it means to create a Government as a Platform ecosystem

Below I’ve partially recreated a table from the Digital Government Policy Framework that represents the needs which Government as a Platform ecosystems can meet. I’d argue each line is integral to the foundational Government as a Platform model, that of ‘an ecosystem supporting service teams to meet needs’. Many of them are also part of the second order ambition of ‘a marketplace for public services’ and some of them help create a route to the most optimistic vision of ‘rethinking the relationship between citizen and state’.

Transforming procurement to improve supplier relations

Training and equipping of in-house capability

Internal tools for civil servant users such as authentication

Standards and controls for spending

Guidance on “what good looks like”

Reusable common components that respond to common user needs

Reusable designs and patterns that respond to common needs

Standards for ensuring the design of services

Standards for technology

Canonical, discoverable data

Standards for publishing and handling data

Cross-governmental networks for delivering services that avoid silos of delivery

Interoperability of data

Transparency of access to personal data and effective models of citizen consent for their reuse

These ideas went on to be expressed through the Enablers pillar of the OECD Framework for Service Design and Delivery (first developed for work in Chile) and then applied as part of the OECD Digital Government Review methodology. Under that lens those 14 things became 7:

• Best practices and guidelines (including style guides and service manuals)
• Governance, spending and assurance (including business cases, budgeting thresholds, procurement, and service standards and assurance processes)
• Digital inclusion focused activities (including digital literacy, accessibility and connectivity)
• The channel strategy (emphasising an omni-channel model)
• Common components and tools (including design systems, hosting and infrastructure, digital identity, notifications, payments, and low code)
• Data-driven public sector approaches (including strategic, tactical and operational activities in line with the OECD Data-Driven Public Sector Framework)
• Talent (including recruitment and professions, communities of practice, consultancy and coaching, skills training and skills transfer in line with the OECD Framework for Digital Talent and Skills in the Public Sector)

---

You’ll notice that while the original 2025 visuals treat ‘trust and consent’ as a distinct layer that digital identity is not given its own focus as an enabler (with the tool folded into ‘common components’ and questions of consent reflected in a wider conversation about data). From an OECD perspective that’s because for many countries digital identity is already a fairly settled and functional tool. So it is comparable to other technical components rather than needing the conceptual discussions that dominate the context in which this conversation happens in the UK.

Back in 2015, and then in the period 2016-2018 when I was involved with the Government as a Platform team, that ‘trust and consent’ layer belonged to the GOV.UK Verify team and because of its scale and scope and importance that made sense. But ‘trust and consent’ isn’t simply ‘digital identity’ and so with hindsight I think what Richard and co were proposing about the interplay between that trust layer and underlying data in 2015 required more of a challenge to ask whether the emerging GOV.UK Verify orthodoxy was going to get us where we needed to be as a country.

I’m not going to extend an already pretty long post to discuss digital identity in detail. However, in light of leading the work to develop and agree the OECD Recommendation on the Governance of Digital Identity I’m afraid I don’t think we’re much closer to realising the capability which the 2015 vision required. Maybe that’s unfair but with different approaches for public sector (but only available if you’re in central government) and private sector services, no sign of any nod to legal persons, people being disenfranchised because in-person identity has been treated separately, as well as a distinct but highly adopted NHS solution, we are still spinning our wheels and off the international pace (if the interest expressed at G7 and G20 levels for genuinely interoperable, cross-border identity were to materialise I don’t think we’re at the races).

So in 2024 if I think about the value of this 2015 visual and what it says about ‘trust and consent’ then it remains critical as something to state explicitly. Digital identity in the UK (with all that comes with it in terms of credentials, proofs, attributes and the rest) is a challenge to solve. And a challenge where the solution really needs to be something that functions on an integrated, whole of society basis as a genuinely foundational layer for all sectors and all people in the country (at home and abroad).

---

But ‘trust and consent’ is of course only as useful as the layer that sits beneath it. And so whatever the state of digital identity it is inextricably bound up with that of data. It is absolutely correct that data was the foundation of those 2015 visualisations because it really does need to be the basis for everything.

And yet, in September 2018 after Richard and I had sat down for a bit of a retrospective about my time as Lead Product Manager for Government as a Platform he had to ask me why I hadn’t mentioned data. Because somehow despite our wide ranging discussions we hadn’t done so.

It’s not a good excuse, or a legitimate reason, but I think in that 2016-2018 period when I was involved with Government as a Platform the data layer, as expressed through the GOV.UK Registers work, was housed outside the programme. The Government as a Platform Programme was very much focused on that slice of technical common components and not the broader suite of enablers. That mismatch between the concept and the delivery vehicle and organisation structures is something I wish I’d been capable of doing something about.

One of the first things I did at the OECD was take on the baton of the Data-Driven Public Sector (DDPS) Framework from Charlotte van Ooijen (finalising her working paper and then coordinating the follow-up policy paper). That’s the basis for the measure of DDPS maturity used in the Digital Government Index and despite a generally strong performance across the board (though showing increasingly strong performances from other countries compared to 2019) it is the measure against which the UK performs least well. The questionable abandoning of GOV.UK Registers no doubt contributes to that (because however established or otherwise they ever were, they would have ticked several of the boxes in the data collection sat behind the 2019 index and could not for the data for the 2023 index).

UK performance in the OECD Digital Government Index

Though the situation is even worse when it comes to tracking the performance of the country when it comes to Open Government Data. For a country that was in the vanguard of those conversations the Open, Useful and Re-useable Data Index paints a very sorry picture indeed.

UK performance in the OECD OURdata Index

All of which is to say that judged against international benchmarks and compared to the countries we imagine as our peers the UK has a lot of work to do in establishing a truly effective approach to data in the public sector that satisfies what the OECD judges as the essential mix of:

• Governance which covers leadership, capability, legislation, operationalising the Government Data Value Cycle, architecture and infrastructure
• Delivering public value which means the nuts and bolts of how government uses data for thinking about the future, delivering in the present and evaluating the past.
• Building trust in terms of consent, privacy, transparency and ethics

The UK does do some bits well but it remains stymied by ongoing structural and political obstacles that make it really hard going. It’s nine years on from these visuals and while the last data strategy had much to commend it, the country is still only having vaguely hopeful conversations about a National Data Library that may, or may not, be the answer to some of the data architectural and infrastructural questions that must be resolved to create the kind of foundations that Richard and these visuals demanded.

---

My final comment loops back to my earlier observation about the whole of government versus individual services. While sorting out data as a foundational enabling layer is critical to ambitions for the public sector, it’s just as important for every service team to be as keenly aware of what the Government Data Value Cycle looks like in their context and how data flows underpin what they’re trying to achieve.

This model needs to be baked into the planning, designing and delivering of individual services. That isn’t about adding data science capability (which is important) or making everything about AI (which it can help), it’s about that core appreciation for recognising data as the foundational building block, and a valuable output, for responding to users and their needs.

---

My overall conclusion is that until the UK gets serious about data and identity then the country isn’t actually close to being where we imagine we are, let alone where we want to be.

Thinking back to 2015 it was absolutely correct for Government as a Platform to be visualised on the basis of data at the bottom with trust and consent layered on top before you got into the technical components. I can only speak to being in the Government as a Platform team 2016-18 and evidently we didn’t get those layers done. We skipped to the common components and trusted (hoped?) Verify as a separate thing would do trust and consent, and that GOV.UK Registers would handle the data.

But, here we are in 2024 with neither.

It really is too early to judge the decision to bring DSIT and CDDO/GDS/i.AI closer together but I hope, and have to be optimistic, that what is going on in this newly combined entity and in the minds of those advising on the future of the ‘Digital Centre’ is keenly aware of that. I trust that they’re coming up with excellent plans and compelling ambitions to make those of us on the outside, looking in, regain the sense of enthusiasm and inspiration that accompanied all the chat in 2015. And maybe they’ll have some similarly insightful graphics to go with it.

---

Well done for reading all this way – if you’re still here then maybe I can also persuade you to sponsor me this Movember. I’ve supported the fight for men’s health every November since 2007 and while it’s always good to hear about the progress that has been made, there is still so much more to do. Your sponsorship is so gratefully received. Still not convinced? Maybe the Movember inspired post I wrote about American Presidents and their facial hair will do the trick?