Reconciling three asset counts at Meridian Energy
Meridian Energy's asset register holds 18,400 records, its CMDB holds 27,000 configuration items, and the discovery tool reports 31,500 devices. A reconciliation exercise finds 4,100 probable duplicates, 12% of records with no owner, and 700 servers with no physical location. ERP contains 16,900 capitalized fixed assets. The change team relies on the CMDB for impact assessment, security relies on discovery for vulnerability scope, and finance closes the fixed asset register quarterly. A new CIO asks why three numbers exist and demands a single figure by the next board meeting. The previous ITAM lead left six months ago; data stewards were never appointed. The discovery tool creates records from MAC addresses, while the asset register uses purchase order lines and asset tags. Cloud subscriptions are recorded manually in a spreadsheet. The CMDB has not been reconciled with discovery since January.
1. Diagnose why the counts legitimately differ and why a single number may be misleading. Propose a reconciliation approach for the board.
Read the worked answer
The counts measure different things: ERP holds capitalized fixed assets, the asset register holds managed assets, the CMDB holds configuration items, and discovery reports observed devices. Duplicates arise because discovery creates identity from MAC addresses, virtual machines change identifiers, and purchase-order lines may not match physical devices. A single blended number is misleading unless the question is defined. The board should see three figures with a bridge: capitalized assets, managed assets, and discovered devices, plus explained exceptions. Reconcile by entity type and strong identifiers such as asset tag, serial number, and cloud resource ID. Treat discovery as evidence, not automatically as truth. Publish an exception report showing duplicates, unmatched records, and missing owners. This gives the CIO a defensible position rather than a false precision that collapses at the next audit.
2. Define the authoritative source and data ownership for serial number, owner, location, cost, patch status, and service relationships.
Read the worked answer
Use different authoritative sources by attribute, not one source for everything. Procurement or ERP is authoritative for cost, purchase order, and supplier. The asset register is authoritative for asset tag, ownership, lifecycle status, and contract. Discovery is authoritative for observed presence and network attributes. The CMDB is authoritative for service relationships and impact. Security tooling is authoritative for patch status, but it must be mapped to a stable CI or asset identifier. Assign data owners: finance for cost, the asset manager for lifecycle and ownership, service owners for relationships, and security for vulnerability status. Location should come from the asset register unless discovery provides better physical evidence, in which case reconcile exceptions. Appoint data stewards and an escalation path. Without named ownership, reconciliation becomes a periodic cleanup rather than a controlled process.
3. Recommend data-quality controls and measures that would prevent the counts from diverging again.
Read the worked answer
Preventive controls should include an onboarding gate before payment, automated discovery import with de-duplication rules, and a change process that updates CI relationships and asset status. Run quarterly reconciliation with exception ownership and closure dates. Measure the data-quality dimensions that matter: accuracy, completeness, consistency, timeliness, uniqueness, and validity. Practical KPIs include duplicate rate, owner completeness, location completeness, discovery coverage, reconciliation exceptions, and time to update after change. Avoid manual cleaning of all 31,500 records; fix the process that creates duplicates. For cloud subscriptions, replace the spreadsheet with a controlled register tied to contract, owner, and cost. Report trends to the CIO and data stewards so improvement is visible and sustained.