Enterprise asset management: deciding what to do with a machine over its whole life

Criticality ranking is the decision that directs everything else

Not every machine deserves condition monitoring, held spares or a standby unit, and treating them equally spreads limited attention until nothing is well protected. Ranking assets by what a failure would actually cost — lost output, safety exposure, environmental consequence, a customer commitment that cannot be recovered — produces a short list where investment is justified and a long tail where run-to-failure is the correct policy. The exercise has to be done with production rather than by engineering alone, because the equipment engineers worry about and the equipment that stops shipments are frequently not the same items.

Condition data matters only where a decision hangs on it

Vibration readings, oil analysis, thermal surveys and process signatures accumulate quickly and are read rarely. The discipline that makes them worth collecting is deciding in advance what a deteriorating trend would trigger and who has authority to act on it. Where the answer is that the machine would be run until it failed anyway, the monitoring is expenditure without a decision attached. Where the answer is a planned intervention in a specific window, the data is doing real work. This is also the honest test for adding monitoring to a machine: name the decision it will change before buying the instrumentation.

Repair, refurbish or replace becomes an argument with evidence behind it

The case for capital replacement usually rests on rising maintenance cost, falling availability, obsolescence of parts or controls, and an inability to hold tolerance or rate. Each of those is recorded somewhere, and none of it is persuasive as anecdote. A system that can produce cumulative maintenance spend against an asset, downtime attributed to it, and the age profile of its control equipment turns a request into a proposal finance can examine. It also exposes the opposite case, where a machine everyone complains about turns out to have a modest failure history and a real problem with how it is operated.

The same asset exists twice, described differently

Finance holds a fixed asset record for depreciation and tax purposes; engineering holds a functional item that gets maintained. They rarely align, because a capitalised project may cover a whole line while maintenance manages individual machines within it, and a component replacement may be capitalised while a repair is expensed. Trying to force a single record usually fails. What is achievable is a mapping between the two, so that spend can be attributed and disposals do not leave equipment maintained on the books after it has been scrapped, or vice versa, which auditors find with unhelpful regularity.

Multi-site standards versus what each plant actually runs

A group system promises comparable reliability figures across plants and shared learning about common equipment. The obstacle is that each site codes its assets differently, defines downtime differently and has a maintenance culture shaped by its own history. Imposing a template without reconciling those definitions produces league tables that plants dispute and quietly game. The workable sequence is to agree the small set of definitions that must be identical, allow local detail below that level, and accept that comparability arrives gradually rather than at the moment the system goes live.

Frequently asked questions

Is this just a larger maintenance system?

There is real overlap, and many products cover both. The distinction is in the questions asked. Work management is concerned with executing and recording jobs efficiently. Asset management is concerned with the whole life: acquisition, criticality, condition, cost of ownership, obsolescence and disposal, usually across multiple sites and connected to capital planning. A plant with one site and stable equipment may never need the wider scope; a group replacing ageing lines across several countries certainly does.

What data do we need before lifecycle decisions become possible?

A stable asset identity that has not been renumbered, maintenance history attributed to it consistently over a meaningful period, downtime expressed in production terms rather than labour hours, and spend that includes contracted work and parts rather than internal labour alone. Most manufacturers find the first two years of data are unusable because coding changed as people learned the system. That argues for settling the structure early and accepting a period during which the history is being built rather than analysed.

How should we handle equipment obsolescence?

Track it deliberately rather than discovering it during a breakdown. The exposures are control systems no longer supported, drives and sensors out of production, and software that will not run on any current operating system. Recording a support status against critical assets, and reviewing it on a schedule, converts an eventual emergency into a planned upgrade. It also strengthens replacement cases considerably, because an unsupported controller on a critical line is a risk finance can understand without a reliability argument.