Delivering better government after three decades of disappointment
For three decades, governments have promised that technology would transform and improve the state - yet the same failures keep recurring. The missing ingredient is to find a better way to prove that a policy works
By
James Findlay and Jerry Fishenden
Published: 24 Jun 2026
UK governments have spent over three decades telling us that technology will modernise the public sector, cut costs, improve services for citizens and businesses, and boost productivity. Yet despite small pockets of success, progress continues to fall well short of politicians' bold ambitions.
In place of the wholesale transformation and modernisation of the public sector, technology is all too often misused to create a thin digital veneer over the status quo. This is a cultural and systems-level failure where:
technology is engaged far too late, automating rather than improving policymaking;
digital transformation degrades into digitised bureaucracy;
service design only gets applied downstream, well after the most important decisions have already been taken, undermining its value and role in the policymaking lifecycle;
structural and political constraints - not technical capabilities - drive failure.
Beneath each of these failures also lies a more fundamental problem. Government has no equivalent of a research and development environment, no controlled environment in which to discover whether a proposed reform works before the entire country is made to depend on it.
We would not license a medicine, certify an aircraft, or sign off a new bridge without testing it first. Yet governments routinely launch policies that touch tens of millions of people with no trial of any kind - only to belatedly discover the real, often disastrous, consequences once they're live and difficult and expensive to reverse.
They spend billions and hope for the best. No serious organisation would implement change this way, particularly change that impacts the lives of so many people and businesses.
It's time to adopt a better approach and apply the lessons of the past 30 years
The policy-technology fracture
The routine sidelining of digital expertise from the political and policymaking heart of Whitehall, underscores the mismatch between the longstanding ambition for the “digital transformation” of government and the reality.
Digital teams typically get engaged only well after all the most important and critical policy decisions have already been locked in. As a result, "service design" is misplaced and misapplied in the policymaking lifecycle, relegated to working out how best to deliver onto a screen a policy long since decided.
Years of Computer Weekly investigations into troubled government programmes illustrate what happens when policymaking fails to properly exploit digital technologies and practices. The landscape is littered with the debris of failed initiatives implemented without adequate exploration and evaluation of alternative policy options, reliable evidence and research - especially with vulnerable groups - and iterative testing and improvement. Examples include:
DWP's Universal Credit: widely cited as a significant failure in delivery and impact, including breaches of the rule of law. National Audit Office (NAO) reports have documented payment delays, hardship for claimants, including vulnerable people, and a reluctance to fully acknowledge the problems caused. There have been court cases arising from its flawed digital implementation leading to arbitrary hardship. It's been criticised for pushing people into debt, causing stress, and failing to deliver on promised efficiency or work incentives in the way initially claimed. Its "digital by design" approach was rushed and insufficiently tested for real-world edge cases.
The Home Office's EU Settlement Scheme: although it processed millions of applications quickly, it also caused significant hardship and frustration. Problems include backlogs on late applications, high later refusal rates, issues with the digital-only "View & Prove" system - described as a fragile "mirror" database - ongoing difficulties with travel, benefits, proving status, and exclusion risks.
HMRC's Making Tax Digital: NAO reports and the Public Accounts Committee have been highly critical. It's suffered repeated delays, massively underestimated costs and complexity, adopted unrealistic timetables, and had insufficient early engagement with taxpayers and agents. It's increased the burden on many businesses without delivering the promised benefits or additional revenue in the way originally intended.
What these initiatives all share in common is that not only was digital expertise engaged far too late to inform and improve the outcome, but there was also nowhere to learn and fail safely.
Those left to suffer the consequences - the vulnerable claimant, the EU applicant left with a fragile “mirror” status check, the small business buried by new regulations - only surfaced once the cost of correcting a broken policy and reversing course had become prohibitive.
The former Gov.uk Verify digital identification service is another perfect illustration of the problem. The NAO found the programme cost £154m between 2011 and 2018, yet could only successfully verify the identity of 38% of Universal Credit claimants. This failed system, which most of its intended users could not use, was designed, built, run, and abandoned without ever being proven on a smaller scale first. Had these ambitious objectives been explored and tested in a real, but contained, setting with real users and real data, the most damaging failures might have been caught while they were still easier and cheaper to fix.
Integrated policy and technology
Government needs to urgently overhaul its outdated "big bang" approach. It needs to integrate digital technologies and practices into policy development as part of an improved "test and learn" approach that prioritises:
A policy-first focus: integrate digital and policy expertise from the earliest possible stages of development, using prototypes, data and evidence, feedback, and simulations to explore options before decisions crystallise.
Consistent benefits: link all work to cross-government / manifesto priorities, a reduced administrative burden, improved productivity, and better outcomes for those affected.
Adaptability: build flexible approaches using open standards and a modular. composable infrastructure so that policy and operational changes can be implemented and maintained rapidly while preserving legal and historic integrity.
Security, control, and sustainability: ensure strong protections, appropriate UK jurisdictional safeguards, data minimisation, operational resilience, and assessments of environmental impact throughout the lifecycle.
Value and reuse: prioritise common components, such as shared platforms, "once only" data models and citizen/business touchpoints, eliminate duplication, and ensure value for money.
Ethical innovation: ensure transparency, human oversight, safety, assurance, and accountability, particularly for the likes of AI and algorithmic decision-making.
Clear mapping of capabilities: establish a single, authoritative view linking strategy, organisations, processes, information, resources, and technology to policy needs.
Robust governance and measurement: integrate risk management, benefits tracking, and performance measures focused on adaptability and sustained public value.
A place to start
If this better approach is to succeed, government also needs somewhere real it can test, learn and improve policy initiatives before releasing them at national scale. Somewhere it can trial integrated policy and technology, observe cause and effect more effectively, and model the likely effects of a change before committing to it.
Regular political pronouncements that the
James Findlay & Jerry Fishenden
Take, as a topical example, the government's current policy proposals for digital identity. A new national digital ID system was announced in 2025, only for its mandatory element to be dropped in January 2026 after a widespread public backlash.
In May, the Home Affairs Committee found government's handling of the policy was “nothing short of a fiasco”, with its rationale only assembled after the public announcement had already been made. Yet the scheme is still due to reach every UK citizen and legal resident by the end of this Parliament, relying on the same One Login and eVisa foundations the committee found fundamentally flawed.
It’s deja vu all over again, as they say - the Gov.uk Verify lesson, unlearned, heading for national scale once more.
But let’s stop for just a moment to imagine a different approach to digital identity. It could first be explored, tested, and improved on the Isle of Wight - real GPs, real benefit claimants, real landlords, real employers, real digital verification service providers, and the very real and important edge cases the design must survive - such as the 90-year old with no smartphone, the person with no fixed address, the carer acting on behalf of someone else. If the new system fails them, it does so in one contained place, visibly, cheaply, and reversibly, not across the entire country at once. Failures surface where and when they can still be fixed, and what later gets deployed at national scale already meets both policy objectives and reality.
Using a policy testbed moves the government's digital ID proposals away from being perceived as a potential surveillance state intrusion into the very opposite - a democratic exemplar of citizen-state engagement where scope, needs, and safeguards are agreed and proven before the public is asked to trust them. An exemplar that applies and complies with principles such as those we outlined above from the very start, including consent, data minimisation, clear UK jurisdiction protections, transparency, and human oversight. As we have previously commented:
“For islanders, this is not a research project imposed on them, and it is emphatically not about treating them as test subjects. It is the most credible, practical and inclusive route to addressing problems that Westminster has been promising to fix for decades and has never quite delivered on.”
A controlled environment can demonstrate that a system collects only what it needs, works for people most likely to be excluded or marginalised, and can be put right quickly and effectively when things go wrong.
Approached this way, the Isle of Wight - all too often treated as a problem to be funded -becomes instead a national asset: an environment where government can actively engage with citizens and businesses to better understand their needs, and test, measure, adapt, and improve policy before it stakes its reputation and citizens’ wellbeing on it.
Applying the lessons
Regular political pronouncements that the "magic" of technology is about to fix an increasingly broken system ring increasingly hollow after 30 years. It's time for governments to apply the lessons of the last three decades and fundamentally rethink and improve how they design and deliver better policy outcomes.
Adopting a better approach isn't simply some "nice to have" - it's existential. It will not only help reduce the endless succession of headline-grabbing failures, but also help governments learn, react, and adapt more rapidly to an ever changing and more challenging world; provide better, more timely outcomes for citizens and businesses; and fulfill manifesto commitments more quickly and effectively, helping, in turn, to restore public trust.
This is why government should now seize the opportunity to:
move digital technologies and practices upstream, integrating them into policy exploration and formulation to inform and improve policymaking and its outcomes.
treat data as a core asset for moving towards better, evidence-based and more responsive, adaptive, and successful policymaking.
design and build systems for a world of constant policy change, enabling quicker and more effective responses to ever changing needs and expectations.
It's time for government to close the 30-year fracture between its bold "digital transformation" ambitions and the reality.
James Findlay is a former CTO/CIO in the UK government. He is the author of the Isle of Wight Living Lab proposal, which advocates for the island's designation as a Special Policy Zone for cross-departmental government trials.
