A barium meal: what the Government Digital Service’s story reveals about a Civil Service that struggles to reform itself

Source: A GDS story 2013. The first birthday of GOV.UK

The creation of the Government Digital Service in 2011 was the start of one of the boldest and most impactful reforms of the last 60 years. The GDS story provides a compelling playbook for the design of high impact reform; it is a story of the rise and fall and moderate recovery of an atypical start-up unit in the centre of government. The story shows how context, leadership, events and transitions both help and hamper the course of a reform over 15 years; and it acts as a barium meal exposing the embedded dysfunction of the Civil Service and how it resists fundamental change.

This week feels like Groundhog Day as the patient rebuilding of the capability, coherence and clout across digital, AI and data over the last two years is thrown up into the air by a machinery of government change. It is a damaging version of what happened in 2021 when data was abruptly split out from GDS and shoved into another department. Those changes were reversed in 2024 after serious analysis and review  - yet that promising reset was upended this week in a move that has the appearance and coherence of something cobbled together in 10 minutes on a cabinet office whiteboard.

I am now knee-deep in the writing up phase of a substantial case study on the GDS reform. With 500,000 words of interviews and over 400 artefacts and articles it has become quite the undertaking for me and my research associate Daniel Thomas. A book I didn’t realise we were writing, the doctorate I never did. The full case study is due to be published in early November 2026 – along with a shorter non-geek edition.

The case study will tell the full story of GDS from its origins in the early 2000’s to where we are in 2026. Our primary research was 35 interviews with a remarkable set of public servants who came from inside and outside government, working both in GDS and in nascent digital teams in government departments. Having worked on previous UK reforms including the PMDU and Capability Reviews I felt an affinity with the driven, risk-taking, highly committed and maverick crew that joined the GDS reform mission because they thought there was an opportunity to make a difference. There are always plenty of civil service sleepers who are up for a bit of revolution.

The compelling mission and inspiring founders who articulated it attracted a diverse group of outsiders  - coders, designers, marketeers, user researchers, and product managers  - into a civil service world that didn’t have or comprehend those roles.

Many of the GDS founders had worked with or helped create the pioneering organisations of the UK civic tech movement in the noughties: a loose network of largely voluntary reformers who saw the potential of the internet to reconstruct the relationships between the state and its citizens. The tactics of the movement were to demonstrate rather than advocate, it was politically ambiguous and focused on users and their needs rather than institutions.

In the same timeframe civil servants were already pioneering digital in some departments, in the Cabinet Office E-gov Unit and Directgov (the forerunner of GOV.UK). The mission of GDS was presaged by the Transformational Government: enabled by technology report of 2005 brokered amongst departmental Chief Information Officers (CIO’s) by the first Government CIO and head of the E-gov Unit Ian Watmore.

The key actions required are (a) to increase understanding of customer needs and behaviours; (b) to define customer groups and appoint directors to lead the overall development of services to those groups; (c) to create a Service Transformation Board to define and enforce common service design principles; and (d) to develop modern channels and manage the migration to them. (Cm.6683, 2005)

That was followed soon afterwards by David Varney’s review: Service Transformation: A Better Service for Citizens and Businesses, a Better Deal for the Taxpayer (HM Treasury, December 2006). Varney argued that improving the user experience and saving public money were complementary rather than competing goals. His most compelling critique was that because government organised itself around its own siloed transactional convenience, they left the citizen to try to stitch together fragmented services themselves:

“underpinned by a mass of helplines, call centres, front-line offices and websites… The end result is that the citizen who needs multiple services is left to join up the various islands of service to meet his or her needs.” (Varney, 2006)

His findings and recommendations surfaced the evolving discussion of elements that became part of the GDS model: identity verification, strong emphasis on user needs, joined up service leadership and product management rather than policy making dominance.

In the run up to the 2010 election in parallel with the civic tech movement Conservative thinkers, activists and advisers were developing a strong agenda for changing how technology was used in government that aimed to end what was seen as a catalogue of high-profile, hugely expensive IT failures in government – in which big system integrator companies were ‘ripping off’ government.

So the ideas that fuelled the GDS insurgency had bubbled and brewed for a decade as these unlikely three tribes – civic tech, pioneering civil servants and politicos  - engaged outside and inside government. The source code for the revolution came from the civic tech movement along with many of the leading figures in GDS. The standing battalions of Civil service policy makers were largely peripheral although the (since abolished) Prime Ministers Strategy Unit was a significant interlocuter.

After the election the essential political leadership came from the new Cabinet Office Minister Francis Maude who had unusually clear views on how to get things done at the centre of government informed by decades of experience around reform. His ambition and urgency were underpinned by a vanishingly rare grown-up partnership with the Chief Secretary to the Treasury (the liberal democrat Danny Alexander) and by cross party support for the agenda.

A playbook for high impact reform

The GDS story matters because it offers a playbook for successful reform lead by a central team that changed the mindset and capability of swathes of civil servants. GDS built a broad-based movement that was passionate about the potential of digital to transform services, passionate about working in a different way: with a bias for getting things done rather than producing endless briefings, policy papers, reports, strategies and project updates. As movements do, it outlived the catalytic host unit and sustained its followers through the organisational ups and downs of the digital transformation agenda in government.

At its best GDS showed what enlightened high capability, high trust management looks like in a team. It upended the secretive and detached nature of too much policymaking as well as the suffocating micro control of civil servant communications. They worked in the open, shared work in progress, reflected on what was going well and what wasn’t in real time. The reform showed the importance of stable and bold political leadership. GDS imported new methods and ways of working that were easy to learn and spread. They eschewed the crushing tedium of so much of the process and bureaucracy of planning, project, programme and performance management that has accumulated in the civil service.

They chose to trust high-quality people who were committed to the missions, gave them the tools and support they needed to succeed, and backed them when things sometimes went a bit wrong. Not a sentence you will hear from many senior policy makers in government these days.

It shone most brightly from 2011 to 2016, when a combination of internal and external changes meant GDS lost its mojo and its thread. Those changes combined to create a cold climate for the evolution of GDS: the loss of political backing and therefore its mandate; a decline in the clarity of their mission; the departure of many of its founders; a retreat into the default operating model of the cabinet office functions offering support but reluctant or unable to lead or direct; the omnishambles of Brexit, Prime Ministers Johnson and Truss; global shocks including COVID and the Ukrainian war; and, a succession of ineffectual Cabinet Secretaries.

The case study is also a story of the lifecycle of an atypical, insurgent start up at the heart of government. What happens when it grows 300% in a matter of months? What happens when the politics changes and when the inspiring founders leave and the mandate crumbles? How do you balance providing support and building capability with running ‘decision gates’ saying no to proposals for digital spending and services? How do you change your operating model to share power and controls as the capability of peers across Whitehall develops? How do you balance pragmatism with purism? What happens when new leaders reject the strong culture of an existing team? What happens when passion and mandate takes on the appearance of arrogance and a lack of empathy with the challenge of delivery and digital transformation in departments and agencies?

Despite the difficulties of the last 10 years, there is a very significant GDS legacy that enables departments and agencies to provide better services to citizens: GOV.UK; the ‘standard for digital services’; design patterns for common transactions and services; common platform components that departments use enthusiastically like Notify, Pay and Forms; the Digital Data and Technology (DDaT) professions; and, communities of practice – are all embedded even if the full vision of the GDS founders has not been realised. The most enthusiastic user facing agencies like the passport office, DVLA – as well as crucial government services like universal credit and tax collection have transformed the user experience and how government delivers services by drawing on the methods and the diaspora of GDS.

Yet the biggest impact is less visible: the change in mindset amongst those who chose to engage with the movement. A change from:

·      policy-first to services first;

·      specifying to discovering;

·      secrecy to working in the open;

·      vendor dependency to in-house capability; and,

·      project to product.

And alongside the mindset changes GDS embedded new routines and ways of working that have spread widely – from formal methods like agile to the personal practice of blogging openly on work in progress, and rituals like show and tell and retrospectives. Regardless of the ups and downs of the central unit these routines live on in the professional identity and personal practice of the GDS diaspora and in the formal and informal communities of practice that have developed around the new roles that the GDS revolution established.

A barium meal exposing civil service dysfunction

The GDS story also matters because it acts as a barium meal that vividly reveals the dysfunction of the Civil Service and the embedded blockers to meaningful reform. Once the ambitions of the GDS insurgency became clear  - that it was about more than a website and sorting out crap IT choices  - digital transformation threatened the established supremacy of the policy classes that hold power and dominate the senior civil service.

GDS exposed the barriers to reform that hamper all serious reform efforts: federalism; the Treasury’s siloed approach to resource allocation and fantastical approach to business cases (making the numbers add up theoretically but not asking ‘is it any good’); the dominance of lofty and remote ‘policy making’ over user focused product and service design; the absence of serious collective leadership of the Civil Service (with the honourable exceptions of two Cabinet Secretaries Richard Wilson and Gus O’Donnell and their serious efforts to build collective leadership); and policy making that is ‘open’ in name only. 

The Civil Service tends to tolerate innovative approaches so long as they are in a sand pit peripheral to what senior leaders see as their principal role: the supreme policy adviser to their ministers. But once innovative units threaten to disrupt the mainstream policy, programmes and funding senior leaders usually ignore them, say yes and do no, shut them down or shove them into an inconsequential department without clout. See for example: Social Exclusion Unit, Total Place, cross cutting PSA’s, Policy Lab, Test Learn and Grow.

The implications for civil service reform

For all that 60 years of near continuous civil service reform have transformed many aspects of the capability of the Civil Service, few would argue that it is fit enough for the future. Institutionally it continues to resist the accumulated logic of what needs to change. So what it would take to rethink how government works? To make it fit for a government with ambitions to focus more on citizens and place, to devolve policy, funding and tax raising powers, to enable the redesigning of whole services around the user and citizen?

The hard wiring and accountabilities of a federal civil service is of course a substantial barrier – but the learned, endlessly reinforced behaviour, mindset and identity of senior civil servants is a much greater one. Machinery of government changes have never and will never change that.

The Civil Service substitutes process for leadership, reorganisations for transformation, assurance rituals for taking personal accountability and being bold, business cases for doing something, micromanagement for trust and empowerment. That needs to change but how will it?

To keep on trying to reform through the essentially managerial interventions that have repeatedly failed to make a lasting change is futile. As one of the greatest public administration academics reflected shortly before his premature death:

“The record shows that managerialism alone is not enough — either in words or in deeds. It turns out that better public managers cannot restore fiscal balance; they cannot save the welfare state and they are certainly not the prime guardians of democracy… But the huge, 40 year-old international exercise of glorifying management as an academic subject and, in the practitioner world, turning top managers into a generously reimbursed elite, can now be seen as a mistake.” (Pollitt, 2016)

In other words, politics matters. Transformation is not coming from managerialism or by importing business methods (or senior executives) from the private sector – all that has been repeatedly tried and substantively failed.

A smaller government that acts as a system steward?

The failure to make bold change to the civil service is not for the lack of a tangible vision of what government could and should be like in the future. We now have over 20 years of strong arguments and evidence for a better approach by Government to its role as a system overseer. These arguments for a change in the role of government and its performance regimes have been brewing since the early noughties (Stoker, 2006)(Osborne, 2006) (Christensen & Lægreid, 2007) (Hallsworth, 2011) (Osborne, 2018) (Ansell & Torfing, 2021) (Sørensen et al., 2021)  - and most notably the prescient work by the Prime Minister’s Strategy Unit on public service reform (Kelly, Gavin et al., 2002). But these ideas have had little traction on most senior officials whose identity and illusion of control as supreme policy adviser would be imperilled by such a change.

The implications for political and managerial leaders of acting as a system steward are profound. Earlier experiments in government like Total Place in 2009, and the continuing Test Learn and Grow programme (2024 onwards) both show it is a path that can be taken, and that civil servants can rise to the challenge. But it requires a very different mindset from the defaults of ‘doing to’, high control needs and micromanagement that is pervasive in government:

“success depends on the building of successful relationships through networks and partnerships… ‘efficiency is not achieved by handing over the job to bureaucrats or managers… the key is learning exchange and mutual search for solutions.’… ‘no one is in charge but leaders at various levels play a role. It is not a linear relationship between a principal and agent.” (Stoker, 2006).

Governments and especially the Civil Service that advises them are stubbornly resistant to letting go of centralised control of funding streams, revenue raising, performance management and policy direction (Ferry et al., 2023). The deeply entrenched ‘doing to’ mindset within Civil Service colludes with the politically seductive notion that Ministers are in control of their domains and make effective policies that direct change in the system and ensure compliance. History suggests otherwise.

To overcome these barriers would involve bold leadership from determined ministers and some reformist permanent secretaries backed by the Prime Minister and Cabinet Secretary. So far, such leadership has not been forthcoming.

Reflecting on the GDS story and the familiar patterns it repeats from 60 years of UK reform efforts, I increasingly think you cannot reform the Civil Service by trying to reform the Civil Service. Instead, something big bold and political is needed to change the game. The Civil Service has great capabilities and brilliant people who respond to all sorts of challenges and crisis but the leaders of the Civil Service have shown themselves unable to change their collective destiny.

The radical devolution of swathes of policy making (and policy makers), fiscal devolution including tax raising powers to the new tier of strategic authorities and mayoralties would force a change. If you have 150,000 less central civil servants especially in the policy making empires, then you need think differently about what they are for and how they add value. Having been in the Civil Service when Gus O’Donnell succeeded in reducing it to around 385,000 staff (compared with the subsequent bounce back under Boris Johnson’s regime to 523,000) – it is not clear that today’s civil service is better for that rapid 36% increase in size.

Only politics will make this happen. Civil service leadership will at best react enthusiastically or at worst offer dogged federal resistant to change. But they will not propose it.

References

Ansell, C., & Torfing, J. (2021). Public Governance as Co-creation: A Strategy for Revitalizing the Public Sector and Rejuvenating Democracy (1st edn). Cambridge University Press. https://doi.org/10.1017/9781108765381

Christensen, T., & Lægreid, P. (2007). The Whole-of-Government Approach to Public Sector Reform. Public Administration Review, 67(6), 1059–1066. https://doi.org/10.1111/j.1540-6210.2007.00797.x

Cm.6683. (2005). Transformational government: Enabled by technology. GOV.UK. https://www.gov.uk/government/publications/transformational-government-enabled-by-technology

Hallsworth, M. (2011, April). System Stewardship. The future of policy making? Institute for Government.

Kelly, Gavin, Mulgan, Geoff, & Muers, Stephen. (2002). Creating public value: An analytical framework for public service reform. Cabinet Office Strategy Unit.

Osborne, S. P. (2006). The New Public Governance? ¹. Public Management Review, 8(3), 377–387. https://doi.org/10.1080/14719030600853022

Osborne, S. P. (2018). From public service-dominant logic to public service logic: Are public service organizations capable of co-production and value co-creation? Public Management Review, 20(2), 225–231. https://doi.org/10.1080/14719037.2017.1350461

Pollitt, C. (2016). Managerialism Redux? Financial Accountability & Management, 32(4), 429–447. https://doi.org/10.1111/faam.12094

Sørensen, E., Bryson, J., & Crosby, B. (2021). How public leaders can promote public value through co-creation. Policy & Politics, 49(2), 267–286. https://doi.org/10.1332/030557321X16119271739728

Stoker, G. (2006). Public Value Management: A New Narrative for Networked Governance? The American Review of Public Administration, 36(1), 41–57. https://doi.org/10.1177/0275074005282583

Varney, D. (with Great Britain). (2006). Service transformation: A better service for ciizens and businesses, a better deal for the taxpayer. Stationery Office.