A finance director opens a £14m substation transformer replacement paper on the Tuesday of a capital review. The engineering case is written up, the sector benchmark is included, and the historic spend line supports the ask. Three slides in, the director puts a finger on the summary page and asks a single question: “Where does this transformer appear in the asset register as a candidate for replacement, and what did its condition score look like twelve months ago?” The room goes quiet. The register carries the transformer as a functional location with a serial number and an installed date. It does not carry a defensible condition trajectory, and the criticality score has not been reviewed since 2015. The paper is deferred.
That scene is recurring across UK utilities, transport authorities, and asset-intensive manufacturers. Capital allowances are shifting from historic-spend defence toward record-of-condition defence. Ofwat’s April 2026 proposals for PR29 will lean less on past expenditure and more on forward-looking asset health data. Ofgem’s RIIO-3 final determinations, published in December 2025, treat asset data submissions as a first-order regulatory artefact. ISO 55001:2024 has tightened the expectation that capital replacement decisions are traceable to the Strategic Asset Management Plan and the register beneath it. The register is moving from ledger to underwriter, and most operators have not re-engineered it for the load.
The register in most estates cannot yet carry a capital case unaided, and the fix is not another tool. It is five questions the head of asset management should be able to answer before the paper reaches the CFO. Each has a mechanism, and each has a named home in IBM Maximo Manage or the equivalent EAM of record.
Do the assets in the register match the plant on the ground?
The first failure mode is boring and expensive. The register lists functional locations and equipment records that no longer match the plant. Assets have been retired without being marked out of service. Assets installed under capital projects have not been onboarded. Serial numbers on the record do not match nameplates. On the brownfield estates we see in utilities and transport, the reconciliation gap between register and walk-down survey commonly runs between eight and twenty per cent on rotating equipment, and higher on instrumentation.
The mechanism is a scheduled walk-down cycle, banded by asset class and criticality, with a documented owner. In Maximo Manage this is a set of PMs against LOCATIONS and ASSET records with a task list that reconciles nameplate against record, and an escalation on discrepancies over thirty days old. It is the precondition for every question below, and it is why our note on governing changes to asset master data after go-live treats master data governance as an operating discipline rather than a project deliverable.
Does the criticality score reflect how the operation runs today?
The second failure mode is quiet drift. Criticality is often the output of a workshop held once, at implementation, with a scoring method inherited from the previous operator. Consequence bands were written against the risk appetite of a decade ago. New assets have been onboarded with a default score or with the score of the nearest neighbour. Under the record, the score has not been reviewed by the safety, environment or reliability lead in years, and the register contains critical assets that were not critical when the score was set.
The mechanism is a criticality method that carries a version, a board-endorsed consequence matrix, and a review cadence tied to changes in the operation, not to the anniversary of go-live. Where the estate carries a Health score alongside criticality, the two are not the same artefact: criticality speaks to consequence of failure, Health speaks to probability. Both go into the capital paper, and the accountability for both belongs on the same nameplate. Reading an APM criticality score without confusing the board covers the presentation problem in more detail; the register problem is upstream of that.
Does the failure history support the replacement case?
The third failure mode is the record everyone assumed was there. When the CFO asks why a transformer needs replacing rather than refurbishing, the reliability lead reaches for the failure history. If corrective work orders on that asset carry blank problem codes, missing cause codes, or a remedy code that reads “repair complete”, the history is decoration. The paper falls back on engineering opinion, which the CFO discounts, because engineering opinion is not a record.
The mechanism is closeout discipline enforced at supervisor level, backed by a completeness KPI on corrective work orders raised against critical assets in the register. The FAILUREREPORT structure in Maximo Manage carries the problem, cause and remedy codes. A dashboard reporting the share of critical-asset corrective work orders closing with all three populated is a short build; adoption is the harder half, and it does not move without supervisor accountability. The failure code library itself has to be usable, which is the argument in our note on failure codes that produce reports worth reading.
Does the cost history separate maintenance from capital replacement?
The fourth failure mode is a cost record that flatters the asset. Repeat repair work has been booked against opex maintenance budgets, sometimes deliberately, sometimes because the work type on the work order was never set correctly. The register then shows an asset whose lifetime cost of ownership looks manageable, while a walk-down finds a unit that has been kept alive by three years of unbudgeted crew hours.
The mechanism sits on WORKTYPE and the general ledger account structure. Corrective, preventive, project and capital-replacement work types have to be separable at the closeout screen and reconciled monthly against finance’s ledger, not annually against a benefits case. Where the estate runs with a handful of active work types and no discipline on which is used when, cost history against the asset is not usable evidence and the capital paper reverts to opinion. The fix is a joint working session with the maintenance planning lead and the finance controller; the Maximo configuration around it is small.
Who signs when the record and the engineer disagree?
The fifth question is the operating-model one. Every estate contains assets where the record says one thing and the reliability engineer says another. The score is low, the failure history is thin, and the engineer is convinced the unit is on borrowed time. That disagreement is real and it should be traceable. The register cannot pretend to be the sole author of the capital plan; it can and should be the primary author, with a named override path.
The mechanism is an override register with three fields: the asset, the reason for departing from the record, and the name of the accountable engineer who signed the override. The head of asset management reviews it at the same cadence as the capital plan itself. Overrides are not failures; they are the seam between judgement and record, and they belong in the audit trail alongside the register they qualified.
What stays hard
The register does not become an underwriter in one quarter. Reconciling nameplate to record on a large estate is measured in shifts, not sprints. Criticality reviews depend on safety and environment attendance that competes with statutory work. Failure code adoption moves at the pace of supervisor discipline. No register carries the whole capital case unaided; replacement decisions still involve sector economics, financing costs and regulatory windows outside the EAM. The register is the primary evidence, not the whole paper.
The position
The next capital review at most asset-intensive operators will not be won by a better slide deck. It will be won by an asset register that survives the CFO’s finger on the summary page. The head of asset management who invests in the five conditions above owns the underwriting of the capital plan; the one who does not will keep watching papers deferred by a single quiet question. That is a governance decision, made in the operating model long before the paper is written.
For the sector-specific version, our note on Ofwat’s PR29 asset health direction covers the water regulator’s move; the shape of the shift is the same across the regulated network sectors.