Data standards are cheaper than integration projects
Every pair of government systems that needs to talk creates an integration project. A shared data standard eliminates all point-to-point work for the cost of one agreement.
Government agencies in Africa commonly build their digital systems independently — health, education, agriculture, revenue, civil registration — each with its own database schema, identifier formats, and API design. When a cross-cutting programme needs data from two or more of these systems, the default response is a point-to-point integration: a new interface between system A and system B. The next programme needs A and C, so another integration. Over five years, the number of interfaces grows quadratically with the number of systems, and the annual integration maintenance cost exceeds the original build cost of any single system.
The alternative is a shared data standard: a common agreement on core entity definitions, identifier schemes, and exchange formats that every new system adopts as a condition of funding. A person registry, a facility registry, a shared list of administrative boundaries. These are not complex systems — they are curated lists with open APIs. But their existence means any new system enters the ecosystem capable of exchanging data with every existing system from day one, without a single integration project.
The barrier is not technical. It is institutional. Each agency controls its own procurement and has no incentive to adopt a standard that makes another agency's job easier. The fix is a policy mandate from the central digital transformation authority: no system funding without compliance with the shared data standards. The mandate is cheap to enforce (checklist item in procurement approval) and expensive to bypass (separate interfaces for every pair of systems).