The shared-platform thesis
The suppliers, products and customers in this chapter are in the landscape database, filtered to shared platforms, with known spend and sources. The graph shows how they connect.
If the eight horizontal layers were built once as public infrastructure and consumed by all sectors, what changes? This page sets out the thesis with order-of-magnitude figures and states its limits.
Summary of the argument
UK public-sector buyers currently pay separately, sector-by-sector, for the same eight functional capabilities, and pay again to integrate them. A small set of well-built, openly-developed platforms, one per horizontal layer, consumed by every sector that needs them, would reduce duplication and vendor concentration and lower long-run cost. GOV.UK Notify shows this working: one communications platform is used by several hundred services, each of which would otherwise have bought its own.
What the shape would be
The model is eight platforms, each with a small number of well-defined responsibilities, openly developed, with well-documented APIs, a self-hostable reference deployment, and a managed instance for institutions that prefer not to host themselves. Services in several sectors would use each platform, with sector-specific configuration on top.
| Layer | Existing nucleus to build on | Largest single uplift |
|---|---|---|
| Identity | GOV.UK One Login + ScotAccount | Extension to NHS staff identity, council customer identity, police staff identity |
| Case management | None at scale; HMCTS digital case has reusable patterns | Social care, probation, regulatory casework, planning, RMS |
| Scheduling | Fragments inside individual products | NHS appointment booking, council services, court listing, dispatch-adjacent |
| Documents/records | Alfresco / open archives where adopted | Council records, court files, FOI processing, prison case files |
| Payments | GOV.UK Pay (inbound) + Pay.UK rails | Local-government and NHS adoption |
| Communications | GOV.UK Notify | NHS and council adoption beyond central government |
| Geospatial / asset | QGIS / PostGIS / GeoServer ecosystem | Council and infrastructure-asset integration |
| Reporting / analytics | Statutory-return tooling, NHS Federated | Public-sector-wide analytics that respects data residency |
Order-of-magnitude savings estimate
Public-sector digital spend on the eight horizontal layers, summed across sectors, is in the region of £3–6 billion per year Low. Much of this is duplicated: each council, trust and agency pays separately for capability that is not domain-specific.
If shared platforms replaced a third of that duplication over a decade, the order-of-magnitude saving is £1–2 billion per year Est at steady state, against £500m–£1.5bn in build, run, and adoption-support cost across the same period Est. On these estimates the investment returns more than it costs over a decade, but the build cost is substantial and the savings take years to arrive.
The survey treats the sovereignty and adaptation benefits as more important than the savings. The sector itself can direct changes to a shared platform, which it cannot do with a vendor product.
What has to be true for this to work
Versions of this thesis have been argued for decades. Several international programmes (covered in the international page) have succeeded with it, and several UK attempts have failed. The factors that determine success or failure include:
- Long-term governance and funding. Institutions do not adopt public infrastructure that may lose its funding at the next political cycle. Stable governance with an independent budget is the largest single factor.
- Engineering quality of the reference implementation. Services adopt platforms that work well. Notify is widely used because it is good software; its use is not mandated.
- No loss of capability. Buyers must be able to use the shared platform without losing any capability they currently have. Where adoption means losing capability, it does not happen at scale.
- Onboarding investment. Open-source software still costs money to adopt. The migration cost has to be supported, particularly for smaller institutions.
- Limited scope. Building eight platforms at once is not realistic. Starting with three and adding the others later is more achievable.
How the CIC fits
The CIC’s plausible role
Building eight national platforms is beyond the CIC. Its contribution would be at component level: writing parts of the platforms and showing in individual sector engagements that the components work. For example, a scheduling component used in three CIC council projects could serve as a prototype for a national scheduling platform, and a case-management framework used in social care, probation and regulatory casework could be adopted by a national body.
The shared-platform thesis would take decades to realise. In its early years the CIC would contribute components and build a delivery record, which would give its arguments for shared platforms more weight.