Clarity of outcomes: knowing what success looks like

In 2021, the UK Government’s identity verification platform, Verify, was shut down after struggling with cost overruns, delays, data issues, and low adoption.¹ The failure of GOV.UK Verify wasn’t a rare exception. It’s all too often the rule for digital projects in the public sector. And it’s not a secret.

Digital and AI transformation programmes frequently fail to achieve their intended aims for a host of reasons, including poorly defined outcomes. Government urgently needs a fresh formula for delivering lasting digital change – especially as we move into a new era of opportunity created by AI and other technology innovations.

At Baringa, we believe the first step in achieving successful outcomes for digital and AI transformation lies in creating clarity of outcomes. Before a project starts, teams need to understand what they want to achieve and where they want to end up. Jean-Paul Sartre’s idea of “being-for-itself” (pour soi) suggests that people have more control and confidence when they clearly define what they want and make conscious choices to achieve this. And arguably more importantly, they must clearly define what this is not. Only then will they take true agency and can progress confidently towards their intended objectives.

This approach led to success for NHS England’s Federated Data Platform (FDP), which was designed to integrate operational data across NHS organisations. From the outset, its main objective was clear – connecting data to reduce waiting lists. Just over a year after its launch, the FDP has exceeded expectations for user adoption, and multiple trusts are reporting tangible improvements in productivity and patient outcomes.²

Deciding on core objectives is easier said than done, especially in these early days of seeing tangible benefits from AI. How can public sector organisations achieve clarity of outcomes to drive value from digital and AI transformation efforts?

1. Addressing the problem of 'disco(very) fever'

Over the years, government departments have conducted numerous discovery phases for digital programmes. Conducting a discovery phase is essential to understanding the problem that needs to be solved, the user needs, any technology or integration constraints and the underlying policy intent that a digital service is intended to address.

However, there's a tendency to start discovery efforts from scratch with each new digital initiative, leading to:

• redundant efforts and wasted resource
• delays in project initiation and delivery
• missed opportunities to build upon previous learnings
• inconsistent approaches across departments.

According to the National Audit Office (NAO), this pattern of repeatedly starting anew contributes to the "consistent pattern of underperformance"³ in the public sector's digital transformation efforts.

To solve the issue of ‘disco(very) fever’, we recommend leveraging existing insights and fostering collaboration across government departments. This begins with establishing a central knowledge repository, which categorises past discoveries by sector, technology, and citizen needs. Regularly updated and maintained, this repository would provide a foundational resource to inform new initiatives.

In parallel, introducing a mandatory discovery review process will ensure that departments consult the repository, identify knowledge gaps, and avoid redundant efforts. By summarising relevant insights and focusing new discovery phases on genuine knowledge gaps, departments can save time and resources and get projects off the ground faster. As part of the discovery review process, projects should consider the very real and new dynamics and opportunities that AI presents. Each new automation or digital change is now an opportunity to lay down structured data that can informal future endeavours in this area and many others.

Discovery should become as much about “what we can learn” as what the current problem and solution may be. This focus on learning reinforces the need for a central knowledge repository as more projects navigate their way through discovery in the age of AI.

2. Dynamic outcome and learning mapping

Current funding and governance processes can present a major barrier to establishing clear project outcomes. The Green Book encourages upfront assessments, but following this kind of one-and-done process can lock teams into inflexible plans. It conflicts with the iterative, test-and-learn approach often required for digital and AI transformation projects.

The recent State of Digital Government review observes that the current public-sector spending model needs to evolve to the requirement for “ongoing funding of persistent teams”.⁴ Recognising this limitation, HM Treasury and the Government Digital Service (GDS) have published supplementary Green Book guidance specifically for digital and AI programmes. It allows for phasing, testing, and learning, which is more aligned with digital transformation principles. However, this guidance still may not go far enough in addressing the dynamic requirements of digital and AI transformation programmes. In particular, guidance fails to address the upside and growth opportunities from data and AI.

To bridge the gap, we recommend UK government implements a dynamic outcome mapping approach to business cases. Dynamic outcome mapping involves:

• Continuous outcome and learning refinement. Replacing fixed, long-term projections with a rolling horizon of outcomes and learning objectives from data that is regularly updated based on real-time data and emerging insights. This satisfies the intent of the Green Book in managing uncertainty and risk.
• Cross-departmental outcome and learning alignment. Identifying and visualising how outcomes from one digital initiative impact or support those of initiatives led from other departments. This will foster a more holistic approach to transformation across government.
• Outcome and learning-linked funding model. Tying funding releases to the achievement of interim outcomes, learning from data, and the continuous refinement of long-term goals. This will encourage agility and responsiveness to changing needs and insight gleaned from data.

Dynamic outcome and learning mapping transforms the traditional static business case into a living, evolving document that can adapt more easily to the rapidly changing societal, economic, political, and technological landscape. By embracing a more flexible and collaborative approach, government departments can better navigate the complexities of digital and AI transformation, increase stakeholder buy-in, and ultimately deliver better outcomes for citizens.

3. Legacy-aware and future-aware outcome mapping

Across government, outdated digital systems remain a key source of inefficiency and a major barrier to modernising services. It’s estimated that 28% of systems in central government departments are legacy systems.⁴

All too often, attempts at digital and AI transformation overlook the legacy problem, prioritising the delivery of new systems and solutions instead. They also overlook future opportunities for change beyond the immediate system refresh, such as failing to understand what value could come from new, clean data and applied learning. To achieve true transformation and deliver the desired outcomes for digital public services, government must first modernise its legacy infrastructure.

Part of the solution is a legacy-aware and future-aware outcome mapping approach. This involves integrating technical debt reduction and legacy system migration into the core objectives of every digital and AI initiative, combined with exploration of what could be learnt from new and better data in that field.

By mandating that all new digital projects explicitly outline how they will contribute to modernising legacy infrastructure, alongside their primary goals, departments can ensure a dual focus on immediate service improvements and long-term sustainability. It would require project teams to collaborate closely with both business units and Digital and Data operations to define clear, measurable outcomes that address current needs while systematically tackling technical debt.

Legacy modernisation should be an integral part of outcome definition, rather than a separate concern. In this way, government departments can avoid the trap of perpetuating outdated systems and ensure that every digital investment contributes to a more agile, efficient, and future-proof technological landscape.

Getting clear on how digital change will be delivered

By implementing these recommendations, government departments can define clearer outcomes for their digital transformation projects. In our next article, we look at how public services can tackle the second dimension of digital and AI transformation – achieving certainty of outcomes.

If you’d like to learn more about how Baringa can help you define clearer outcomes for digital and AI transformation, please get in touch.

Find out how to build confidence in digital and AI transformation for the public sector

^{1. The challenges in implementing digital change - NAO insight}

^{2. “Trusts using FDP treat 114 more patients a month on average” - Digital Health}

^{3. The challenges in implementing digital change - NAO insight}

^{4. State of digital government review - GOV.UK}