The Logistics Group's Backbone Question
A Europe-based logistics group runs IT as a shared service for 14 business units. Its CIO funded a two-year integration programme after an audit found 14 overlapping tool stacks, 38% of production services with no named owner, and an average of nine days to answer the question of who approved a service and at what price. Procurement creates supplier contracts in an ERP module, the service desk keeps its own list of services, and monitoring knows only hostnames. The architecture team has drawn four value streams on a page but cannot explain how a new service moves from an idea to a billable offering, or which record should be trusted at each step. A steering committee wants to start by building a CMDB and a service catalogue, then do IT4IT later. You are asked to advise on sequencing and on the authoritative record for a service contract.
1. Which value stream should be authoritative for the Service Contract data object, and which other value streams depend on it? Justify from decision rights and lifecycle state, not from tool ownership.
Read the worked answer
In IT4IT the Service Contract is a backbone object that records the agreed relationship between a consumer and a provider: what service, at what price, under which terms, for which period. The decision rights for creating it sit where a qualified demand is matched to an offering and the consumer commits: Request to Fulfill. R2F already owns the offer, request, subscription, entitlement and charging flow, so it is the only value stream that can transition a contract from proposed to active to terminated with the consumer's acceptance.
S2P is not the owner but constrains it: Service Portfolio Management and Policy define whether the offering may be sold, at what price band, and to which consumer segment. R2D supplies the release and service state that says whether the offering is consumable; a contract should not activate against a service that has not passed deployment. D2C consumes the contract to know service levels, entitlements and support obligations when handling events, incidents and problems.
The test is not which tool holds the record, but which value stream has authority to change its state. Procurement's ERP contract with a supplier is an Agreement with a supplier, a different object; treating it as the Service Contract is the root cause of the nine-day question.
2. The steering committee wants CMDB, then catalogue, then IT4IT. Give a reasoned critique and an alternative sequence for the first 90 days.
Read the worked answer
The sequence inverts cause and effect. A CMDB is an implementation of configuration information; a catalogue is an implementation of offering and request information. Neither resolves the missing decision rights and object ownership that produced 38% unowned services and a nine-day answer. Building them first typically produces a technically accurate CMDB that no value stream trusts, and a catalogue that duplicates the ERP and service desk lists.
Alternative: first agree the value stream scope and the authoritative component for the backbone objects (Service, Service Offering, Service Contract, Configuration Item, Incident, Change). Second, define the minimum identifiers and state transitions that let one object reference another. Third, choose one service family and three business units as a thin slice, and implement the flow from S2P portfolio entry to R2F contract to D2C support. Fourth, measure traceability on that slice. Only then extend the CMDB and catalogue, because by then you know which attributes are needed and which system is authoritative.
The critique is not that CMDB and catalogue are wrong; it is that they are downstream consequences of decisions the committee has not yet made.
3. What single operational test would reveal whether the backbone is working, and what would it measure?
Read the worked answer
Pick one live service and ask for its full history in one query: portfolio decision, offering, contract, deployment, current configuration items, open incidents and changes, and charges for the last quarter. The test is whether the records resolve to the same service and contract identifiers without manual reconciliation.
Measure elapsed time and number of hand-offs to assemble that view, and the percentage of records that cannot be linked. A target of under one hour and under 5% unlinked is more meaningful than tool counts or CMDB class coverage. If the nine-day question persists, the backbone is decorative. This test also exposes false ownership: a record may exist in two tools, but if neither can be shown to be authoritative, the value stream has no single state.