A UK domestic energy supplier: four governed master data domains behind an S/4HANA transformation

The client supplies gas and electricity to millions of domestic accounts and runs one of the largest domestic field service organisations in the UK — boiler installation, servicing, breakdown cover and smart metering. Those two businesses meet at the same customer, the same premise and often the same appliance, but they had grown up in different systems.

The S/4HANA programme made that unavoidable. A transformation of that size does not improve master data on the way through; it carries whatever it is given into a system that far more of the business depends on. Customer, product, asset and vendor records that disagreed across IS-U, ECC, CRM and the field service estate would have disagreed in S/4HANA too, with a larger blast radius.

So the governance layer was scoped as a precondition rather than a follow-on workstream. MDG had to be able to issue a golden record, and the business had to be able to steward it, before the domains it governed were migrated.

Customer, product, asset and vendor were governed as one programme rather than four. That is a harder delivery shape, and it was the right one here because the domains are not independent in this business.

An asset — a meter, an installed appliance under service contract — is meaningless without the premise and the customer relationship it sits under. A field service job draws on the product catalogue, the engineer's parts availability from the vendor base, and the customer's contract terms at once. Governing customer first and asset later would have meant stewarding customer records against asset data that had not yet been given a definition.

MDG was deployed as a hub on S/4HANA, with BRFplus carrying the approval workflows, the Data Replication Framework distributing approved records to the consuming estates, and SAP Data Services and Information Steward handling profiling and remediation on the way in. The hub pattern matters: it gives one place where a record becomes official, rather than several systems each believing they hold the master.

Millions of metering points — MPAN and MPRN — including a large and growing SMETS1 and SMETS2 smart meter population, sit in the asset domain. They do not behave like the assets in a manufacturing master data programme.

A metering point has a regulated identity that exists outside the supplier's systems, a lifecycle driven by industry processes rather than internal ones, and a relationship to premise and customer that changes without the supplier initiating it. Governance rules written for an internal asset register do not survive contact with that. The domain model had to accommodate a record whose authoritative identity is partly external, and stewardship had to recognise when a change originated in the industry estate rather than in the business.

The vendor domain carried a comparable complication in a different direction: meter asset providers, installation partners, parts and appliance suppliers and contact centre partners all participate in delivering the same service, so vendor records feed operational field service as directly as they feed procurement.

MDG now sits between the source estates and the platforms that consume master data: S/4HANA, demand planning, CRM, field service delivery and the BI and analytics layer. Each takes approved records through the replication framework rather than reconciling its own version.

The commercial model was built on the same basis the engineering was measured on — fixed-price milestone delivery with an outcome-based data quality element and a risk-reward pool tied to cutover and adoption. That alignment is worth stating because it changes the delivery conversation: data quality stops being a reporting line and becomes a term.

What we would flag: a governance hub is only as good as the stewardship behind it. The technical layer can route a record for approval, apply the rules and replicate the result, but it cannot decide what the business means by a customer. That definitional work was the slower half of the programme and the half that determines whether the golden record stays golden.