Outsourcing support is easy to buy. Outsourcing administration means someone outside your organisation can change production — and that is a governance question before it is a service question.
A procurement platform needs an administrator. Not a help desk — an administrator: someone who configures the chart of accounts, maintains approval chains, loads budgets, manages supplier records and keeps up with three platform releases a year.
For a national cultural institution that is a role, not a full-time job. The usual answer is to give it to someone who already has another job. They are trained, they are capable, and they are busy — so configuration drifts, release changes go unreviewed, and the institutional knowledge sits with one person who eventually moves on.
Buying that capability from outside solves the availability problem and creates a sharper one. An external administrator needs production access.
Administration as a service. Valtatech took the work that would otherwise fall to a certified platform owner: onboarding and offboarding users, chart of accounts configuration, master data for users, suppliers and account codes, creating and loading budgets, supplier record changes, approval chains, content groups and blockers, order forms and buyer-managed catalogues — across Procurement, Invoicing, Expense and Supplier Information Management.
Level 1 support alongside it. Platform defects under normal operation and after version releases, plus the supplier-facing problems that generate the most volume in any procurement platform: suppliers unable to register on the portal, not receiving purchase orders, or unable to submit invoices. Issues needing vendor escalation were raised with Coupa support but managed by us through to resolution, rather than handed over and tracked by the institution.
Response targets by severity, stated honestly. Five priority levels from a system-wide outage to a feature request, each with a defined response time. The service documentation says plainly that a response target is not a resolution target — a distinction that is often blurred at the point of sale and then discovered later.
The controls that make external administration defensible. Routine production changes are described in writing and confirmed by an authorised counterpart at the institution before execution. Substantial changes are made in test first and require approval from both Valtatech management and the institution before promotion to production. Batch data uploads are reviewed and approved before loading. Where resolving a support ticket needs a configuration change, that change goes through test and acceptance testing like any other.
A deliberate scope boundary. Integration support was explicitly not part of the plan. Integration incidents were monitored and routed to the institution's own system integrator. Saying so up front is better than discovering the gap during an incident.
Governance with a fixed cadence. Weekly activity reporting, timesheets in thirty-minute increments, and a standing service review call. Unused hours could be borrowed forward within the period, so a quiet month did not simply evaporate.
Knowledge transfer, not dependency. Coupa training was delivered to the institution's own team as a separate engagement, and a knowledge base was built from live ticket history so recurring questions could be answered without raising a ticket at all.
Valtatech is software vendor agnostic. The technology named here is what this client selected; we start from requirements, operating model and existing systems. Talk to us about managed services →
The part of this engagement worth copying is the change control, because it answers the objection that stops most administration outsourcing before it starts: you are giving an outside party the ability to change a system your finance close depends on.
Dual approval is the answer, and it has to be real. A written description before routine changes. Test-first and two sign-offs for anything substantial. Review before batch loads. None of it is sophisticated, and all of it is the difference between a service an audit committee will accept and one it will not.
The second thing worth copying is the training. An administration service that deepens dependency every year is good business and bad value. One that transfers what it learns — through training and a knowledge base built from the tickets people actually raise — shrinks its own demand.
Talk to us about what this would look like in your organisation.
Start a conversation