In the first two articles in this series, I explored how we built today’s fragmented public sector technology estate, and why our operating models have failed to keep pace with technological change. The next question is what we build in its place.
The technology estate is not the only part of the public sector shaped around organisational boundaries, because behind every departmental system, council system and health system sits an institution funded, governed and held accountable for its own responsibilities. Our technology fragmented because the wider system fragmented, and our applications became increasingly self-contained because the organisations buying them were expected to demonstrate control over their own people, processes, information, expenditure and outcomes.
That institutional model brought clarity and accountability, but it also shaped what we were able to see. Problems entered the public sector through a particular service, were translated into the language and eligibility criteria of that service and were then managed inside the organisation responsible for it. Technology reinforced the same pattern by turning a person into a case, transaction or record held within the application belonging to that institution, usually with only a partial view of the wider circumstances around them.
As long as the challenge was to help an individual organisation perform its existing role more efficiently, this model worked reasonably well. It becomes much less effective when the outcome depends on several organisations acting together, when preventing demand in one part of the system requires action somewhere else, or when the people, knowledge, relationships and capabilities needed to create the outcome are distributed across a place. The wicked problems facing public services expose the limits not only of our technology but of the institutional logic around which that technology was designed.
Moving beyond centralisation and decentralisation
We have usually responded by moving boundaries, redistributing responsibilities or creating another layer of coordination, which is why the public sector repeatedly finds itself debating centralisation and decentralisation as though every reform must ultimately decide where control should sit. Should something be national or local, owned by a department or devolved to a place, standardised across the country or adapted around local needs?
These are understandable questions, but they assume power, capability, delivery and accountability must all sit at the same level, and that improving the system means finding the correct box in which to place each responsibility.
I increasingly think this is the wrong debate, because the future requires national coherence and local flexibility at the same time. It requires some capabilities to operate consistently across the country, some decisions to be shaped around populations and places, and some services to remain deeply local, relational and rooted in communities. Trying to force all of that into a single hierarchy will either centralise away the knowledge that makes local delivery work or devolve away the coherence needed for the wider system to function.
The language I keep returning to is federation.
Federation is not another word for decentralisation, and it is not a technical integration pattern hiding inside an operating model. It is a way of organising government in which decision-making and service delivery are distributed to the level at which they can be done best, while the shared capabilities allowing the whole ecosystem to function are protected and stewarded nationally.
The more we devolve delivery, the more important national standards become.
The more autonomy places have, the more important interoperability becomes.
The more organisations compose services around local population needs, the more valuable common data models, trust frameworks and reusable components become.
Local flexibility and national coherence are not competing ambitions in this model, they are mutually dependent, because places can only exercise meaningful freedom without creating further fragmentation when they are building from foundations that allow their decisions to remain connected to the wider system.
This changes the way we should think about devolution. Central government should not devolve applications for places to implement, exporting the complexity of individual departments into local government and allowing each department to do so differently, because that simply creates more layers of variation and leaves places responsible for joining together fragments designed elsewhere. It should devolve the ability to shape delivery through regional operating models and local priorities while maintaining the common digital spine that allows services, data and organisations to remain connected.
A national spine, with local freedom to deliver
The principle is simple:
National spine, providing the rules, standards, trust and common capabilities that must work everywhere.
Regional orchestration, understanding populations, shaping operating models and bringing organisations together around place-based outcomes.
Local delivery, combining professional practice, community knowledge and trusted relationships around the circumstances of individuals and families.
Citizen-centred experience, allowing people to experience joined-up support without needing to understand which institution owns each intervention.
This is not an integration architecture, it is an operating model for a state that needs to combine national scale, place-based coordination, local delivery and personal experience without forcing them into a single hierarchy.
The centre’s role in this model is not to build an ever-larger suite of national applications, and it is not to describe every local service pathway. Its role is to steward the common digital spine that allows a distributed system to remain coherent, interoperable and trusted.
Some things make sense nationally because fragmenting them would degrade progress for everybody, including identity, common standards, national data foundations, cyber assurance, trust, consent patterns and the interoperability needed to move safely between organisations. These are infrastructure for government rather than products to be recreated by every department, council or public body, and we need to start treating them in that way instead of continuing to procure them as features buried inside separate applications.
Building capabilities, not more applications
The composable capabilities built around the spine should include things recurring across many service journeys, including:
- Identity verification, eligibility checking and consent
- Referrals, appointments, status updates and secure messaging
- Evidence sharing, outcome reporting, audit and safeguarding controls
- Payments and entitlement integration where services need them
The point is not that every service should become identical, because population, geography, economy, community capacity and local operating models matter. The point is that difference should sit where difference creates value, in service design, prioritisation, relationships, professional judgement, workflow and delivery, rather than in twenty separate versions of checking eligibility, making a referral, uploading evidence or reporting an outcome.
This requires a different relationship with the market as well as a different relationship between public bodies. Suppliers need to engineer this future with us, building components, platforms and services able to participate in a wider ecosystem rather than requiring every customer to enter another closed one, and many are already moving in that direction because the wider technology industry is moving there too. Public sector buyers need to meet them in that future, creating room for reuse, interoperability and collective investment rather than rewarding the product promising to contain the most functionality behind a single contract.
Architecture becomes less about controlling individual systems and more about establishing the rules through which a changing ecosystem can remain coherent. Standards stop being documentation produced around delivery and become the means through which delivery can happen across boundaries. The role of the technology leader changes with it, becoming less about selecting the one product that promises to solve everything and more about creating the conditions in which multiple organisations, suppliers, platforms and professions can work together without recreating complexity for the citizen.
Our enduring assets become the data, standards, trust and shared capabilities underneath the service, while applications become increasingly skinny and temporary ways of interacting with them.
