public digitalThe public digital logo

Test and learn; going beyond programmes in the public sector

This article is included in our book, Adopting the practice of test and learn. It was written by Audree Fletcher. It builds on a previous blog post published in 2024. 

You can download a PDF version of this book, or visit this page to request a printed copy.

In the public sector internationally and in the UK, programmes have long been governments’ go-to vehicle for making major changes to public services.

The UK’s Government Functional Standard calls programmes “unique, temporary, flexible organisations created to coordinate and oversee the implementation of projects and related work”. In brief, they are there to produce a set of pre-specified changes or outputs, and then vanish.

This mode of delivery holds a broad appeal for central government: As well as being codified and institutionalised, the size and flexibility of programmes enable useful opportunities for scope creep, while political figures benefit from their big numbers and strategic importance.

Programmes have their place in public service delivery. But where they are implemented in order to create a service or transform a service, this approach of discontinuous, one-off change carries a heavy cost.

There is an alternative approach for building, evolving and sustaining public services over their whole lifecycle. That approach, at its heart, is test and learn.

Programmes and the fantasy of one-off change

The change infrastructure of government operates under the assumption that transformation is a one-off: a significant and abrupt change to the status quo.

The expectation is that once the change is complete the programme can be wrapped up, making way for other projects in the government portfolio. At the point of programme closure, the programme director moves onto the next big thing, the delivery team is wound down, and responsibility for maintaining and running the service passes over to a service management team expected to “keep the lights on”.

This logic is reinforced by government frameworks - across funding, governance and delivery - which are often split into two categories: “change” and “business-as-usual” (BAU). After change is finished, the logic follows, only the running of the service and occasional maintenance is required.

Even at a conceptual level this is deeply flawed, because - unlike programmes - services endure and must continue to meet evolving needs. They exist as long as the public needs them and must continuously improve with proactive lifecycle management, supported by appropriate stewardship, governance and funding.

The cost of discontinuous change

The strategy of building services through ‘big bang’ programmes creates a landscape of services which consistently fail to deliver outcomes.

Some programmes become ‘cinderella’ services which, funded at or below subsistence levels, steadily decline over time. A notable example is the Ministry of Justice prisons visits service, designed to help citizens book visits with people in prisons. The service was well researched, well-designed, given adequate BAU headcount and funding, and saw reassuring take-up after its launch in 2014, but a 2020 report revealed that once built, “the funding disappeared, the service atrophied, the tech stack atrophied, and take-up fell right away”.

Other programmes, despite their pre-set closure dates, fail to actually end. Instead, they get caught in an endless loop of programme resets and renewals, unable to escape their ‘unfeasible’ label. In these cases, scope creep becomes ever harder to resist.

A model for sustaining public services: test and learn

Our current approach is defined by high-scrutiny programme governance, project-centric business cases, and stage-gated project phases. It doesn’t fit the risk profile, shape and pace of service transformation. It doesn’t deliver continuous improvement or support service stewardship, because as an approach it’s over-engineered upfront and under-supportive over the longer term.

There certainly is a place for programmes in the public sector. But we also need an approach to delivery that works for building, evolving and sustaining public services over their whole lifecycle, and which acknowledges the reality that improving services is not a short-term deliverable, but a continuous process.

A test and learn approach meets those needs, but it requires major culture change in central government:

  • We need to fix funding in traditional ‘run and maintain’ phases so that it covers significantly more than just software patches and licensing. If service funding disappears when a programme closes, this effectively sets an expiry date on the service - the start of its steady decline.

  • We need long-term stewardship of services by an enduring service team. That means moving away from teams disbanded at the end of a programme, and moving towards funding and organisation design that allows leaders to build permanent, cross-functional, multidisciplinary teams.

  • We need governance appropriate for sustaining services, supporting decision-making at the pace of delivery (rather than at pre-defined stage gates) with a focus on continuous improvement against outcomes over the long-term.

  • We need to recognise spending on sustaining services as the strategic investment that it is: something that will build and retain asset value, and avoid service decline and obsolescence.

Much of this change is held back by current government accounting and spending control models: the Treasury focuses largely on projects with high net present value (NPV) and a calculated whole lifetime cost, concepts which aren’t easily applied to services without an expiry date. What’s more, funding in the form of capital expenditure is more readily available than the operational expenditure budgets required for building enduring service teams.

Despite these obstacles, the last few years have seen a slow but steady emergence of service-centric operating models across central government and agencies. HM Treasury has indicated intent to shift away from traditional, waterfall governance towards service-focused funding models, and the Test, Learn, Grow programme established by the Cabinet Office in 2025 aims to rewire central government to enable these approaches.

In fact, since the start of the Labour government in 2024, there have been stronger signals of intent - and occasional real examples - toward adopting continuous models of change. In 2024, Pat McFadden, Chancellor of the Duchy of Lancaster, called for the adoption of a test and learn approach within the civil service. At the same time, outliers such as the Infected Blood Compensation Authority (IBCA) have taken a new approach to delivery based on sustaining services.

Signs of culture change: the work of the Infected Blood Compensation Authority

In 2024, Public Digital began work with IBCA to design a compensation service for the thousands of people affected by the scandal in which men, women and children in the UK were given infected blood and blood products by the NHS.

The work of IBCA, as an arms length body set up to take a digitally enabled approach to delivery, is a striking example of a test and learn approach in government. As its CEO recently explained “by starting small, testing and learning, we can develop and improve the service quickly, meaning payments will be made sooner for everyone”.

The impact of this continuous approach to change - in contrast to typical ‘big bang’ approaches to delivering public services - has increased the pace of delivery. It can typically take approximately 2-3 years to set up a compensation scheme after regulations are laid, but IBCA was able to make the first payments within 4 months.

Changing how we change

The work of IBCA offers us a glimpse of what can be achieved when we shift our collective mental model for public services from discontinuous ‘big bang’ change towards a test and learn approach which promotes significant continuous change.

If we change the way we think about change in public services more broadly, we can shift the way we fund, govern and buy services in complex, uncertain and evolving contexts. We can grow the legitimacy of alternative, complementary delivery frameworks that can sit alongside programmes in public service transformation.

Service leaders in government can finally feel encouraged and empowered to adopt the last target operating model they should ever need, one that maximises adaptability and prevents atrophy. Programme leaders can plot a route to escaping their ‘unfeasible’ programme delivery labels.

And our public services will be all the better for it.

Previous:

About Audree Fletcher

Audree is a Director at Public Digital, where she leads digital transformation and design programmes for major public sector organisations including the Home Office, UK Parliament, Arup and the NHS. Prior to joining Public Digital, Audree held roles as Chief Digital Officer at the Department for International Trade, and Service Director at NHS Test and Trace.