Building an asset register that's actually useful.

The asset register is the least interesting document a facilities team owns and the one everything else depends on. A maintenance schedule is a set of instructions about the register. A compliance report is a claim about the register. An audit is, functionally, someone checking whether the register matches the building. Get it wrong and every system built on top of it inherits the error while looking entirely healthy.

The characteristic failure is not inaccuracy. It is incompleteness that nobody can see — three items of plant that were never entered, and therefore have never been scheduled, never inspected, and never appeared as a gap in any report.

Decide what an asset is before you start

This sounds like a philosophical question and is in fact the most practical decision in the project. Is a fire door an asset? Every radiator? Each individual emergency light, or the emergency lighting system? There is no universally right answer, but there is a usable test: something is an asset if you need to hold a maintenance obligation or a service history against it individually.

Apply that consistently and the register sizes itself. A boiler is an asset because it is serviced individually and its history matters. Two hundred radiators are probably not, because nothing is scheduled against radiator number 147 and nobody will ever ask for its history. Emergency lights usually sit somewhere in between, and the deciding factor is whether your testing regime records results per fitting or per circuit.

Record less than you think

The register that never gets finished is the one that demanded forty fields per asset on day one. Every additional mandatory field multiplies the survey time, and fields that are hard to obtain are the ones that end up guessed — which is worse than empty, because an empty field is visibly unknown and a guessed one is not.

• –What it is — asset type, from a controlled list rather than free text
• –Where it is — building, floor, room or plant area, precise enough to find it
• –Which one it is — make, model and serial where they are readable
• –How much it matters — criticality, recorded on a scale you will actually use
• –When it arrived — install or commissioning date, where known
• –Who looks after it — the contractor or trade responsible, if that is fixed

That is enough to schedule maintenance, run compliance and answer an audit question. Everything else — warranty terms, replacement cost, expected life, energy rating — is genuinely useful and should be added as it becomes available rather than blocking the survey. A register at eighty per cent that is in daily use beats a perfect specification nobody has filled in.

Controlled lists are not bureaucracy

Free text asset types are the single most reliable way to make a register unsearchable. "AHU", "A.H.U.", "Air Handling Unit" and "Air handler" are four different things to a filter and one thing to a person, and once a few thousand rows exist in four spellings the only fix is a data cleanup project. Fix the vocabulary before the survey starts, not after.

Surveying an estate without stalling

The common mistake is attempting the whole estate at once, in detail, as a project with a completion date. It stalls somewhere around the third building, usually when a plant room turns out to contain something nobody expected. A better shape is building by building, with each one finished and in use before the next one starts.

• –Do one building end to end, including putting its schedule live, before starting the second
• –Survey plant rooms first — they hold the assets with the heaviest obligations
• –Photograph rating plates rather than transcribing them on site; transcription is faster and more accurate afterwards
• –Record what you could not identify as an explicit unknown, not as a blank
• –Reconcile against existing evidence — service certificates, contractor visit sheets, O&M manuals — to catch plant nobody walked past

That last one finds the assets a walk-around misses. A contractor invoice for servicing something you have no record of owning is a very direct way to discover a register gap, and every estate has a few.

Keeping it accurate afterwards

A register decays from the day it is finished. Plant is replaced, rooms are repurposed, buildings join and leave the portfolio, and each change is small enough to not get recorded. Within three years an unmaintained register is describing a building that no longer exists, and the worst part is that it still looks authoritative.

The durable fix is to attach register updates to events that already happen rather than to a periodic review nobody prioritises. An engineer attending an asset is the cheapest possible moment to confirm it exists, is where it says it is, and still carries that serial number. Holding service history against the asset rather than in a chronological visit log is what makes that confirmation a by-product of the work instead of a separate task.

• –Update the register when plant is replaced, as part of closing the job that replaced it
• –Confirm existence and location as part of routine attendance, not as a separate audit
• –Run a reconciliation when a building joins or leaves the portfolio
• –Review criticality ratings annually — they change with occupancy and use more often than people expect

What good looks like

A useful register answers three questions without anyone having to think. What do we have, and where. What is each item obliging us to do. What has actually happened to it since we took it on. If your register answers all three for every building, the planned maintenance regime built on it is standing on something solid, and the compliance reporting above that is describing the building rather than describing the spreadsheet.