The question before the conclusion.
- Lens
- maintenance
- Journey
- 8 decision points
- Reading
- 7 min
The question comes up in the management committee: should this forklift be repaired or replaced?
The maintenance manager has an opinion. They defend it with conviction and are probably right—they know the machine. But when asked how much it has cost over the past three years, the answer comes with caveats: the parts could be found; internal labour cannot, because it is allocated nowhere; the service provider’s interventions are in invoices mixed with those for other equipment; and no one has ever calculated the cost of downtime.
The decision will therefore be based on experience. In other words, on one person’s intuition—usually excellent, always unverifiable, and lost on the day that person leaves.
Total cost of ownership is not an accounting concept
It is an operational concept, and that is what makes it difficult.
Accounting knows depreciation, supplier invoices and parts purchases. It classifies them by nature and cost centre. What it never does, because that is not its role, is attach every expense to a specific item of equipment.
Yet that is exactly what is needed to make a decision. The total for the year’s ‘spare parts’ line says nothing. A technical director needs to be able to say: this particular machine cost this much, comprising this much labour, this much in parts and this much in external services, and it was unavailable for this many hours.
Four components, three of which are almost never allocated:
Parts—the only component that is tracked reasonably well, when the parts store is properly managed.
Internal time—the hours worked by in-house technicians. They are paid anyway, so they ‘cost nothing’. That is false: an hour spent on one machine is an hour not spent elsewhere, which is precisely what needs to be arbitrated.
External services—present in invoices, rarely broken down by equipment. A service provider’s monthly invoice often covers around ten interventions on as many machines.
The cost of downtime—the heaviest and least measured component. A line at a standstill means people waiting, production not completed and sometimes penalties. This cost often exceeds all the others combined.
Preventive maintenance is not judged by its completion rate
Almost every organisation that manages maintenance tracks the same indicator: the percentage of planned preventive interventions completed.
That is an activity indicator, not an outcome indicator. It says that the organisation did what it planned. It does not say whether what it planned was useful.
A preventive plan inherited from the supplier, never revised and applied scrupulously produces an excellent completion rate—and sometimes many needless interventions while the actual causes of failure remain untouched.
The indicator that matters is the proportion of corrective interventions on equipment already covered by preventive maintenance. If a machine fails three weeks after its preventive visit, the preventive work did not fail—its content was not the right one.
This cross-check requires very little: every corrective intervention must be attached to a piece of equipment, and its history must be visible. That is all. It makes it possible to revise a plan based on facts rather than the manufacturer’s manual.
Diagnosis without history: the invisible cost
A technician attends a fault. The same fault on the same machine occurred eight months earlier. A colleague resolved it.
If there is no searchable history, the technician starts the investigation again from the beginning. The lost time appears nowhere: it is absorbed into the duration of the intervention and treated as normal.
This is probably the most immediate and least promoted benefit of tool-supported maintenance management. Not planning, not indicators: memory. What was done, by whom, with which part, and whether it lasted.
For that memory to exist, the intervention report must be entered. And for it to be entered, it must be possible to enter it where the technician is—that is, in front of the machine, not in an office at the end of the day. A report written from memory three days later is an empty report.
Internal, external, or both: write down the boundary
Almost every organisation combines internal teams and service providers. Few have written down the boundary.
What happens then is that the provider intervenes, but its report stays with the provider. The equipment has a history in two parts, only one of which can be consulted. When the organisation wants to assess the provider—response times, recurrence and quality—it has only the provider’s own invoices on which to judge it.
What can be required contractually, at no cost when negotiated at signature, is for the report to return to your system, in your format, within a set time. After signature, that becomes a favour.
The question of property assets
In addition to machinery, many organisations manage buildings and technical installations: energy, air conditioning, lifts and fire safety.
These assets follow the same logic—an equipment hierarchy, interventions, periodic obligations and costs to allocate. But they have one particular feature: a significant share of their maintenance is mandatory and scheduled. Regulatory inspections are not negotiable, and missing one has consequences beyond availability.
The essential minimum for this scope is an inventory of obligations by installation, with their frequency and next due date, and an alert that does not depend on one person’s memory.
What not to expect from software
Three promises circulate and should be put back in perspective.
Predictive maintenance. It requires sensors, accessible measurement data and enough failure history to establish correlations. If you have neither sensors nor history, no software will predict anything. Start by building the history—it is a prerequisite and takes months.
Replacing technical supervision. Industrial controllers, building management systems and sensors measure and raise alarms. Maintenance software uses their alarms to trigger interventions. It does not replace them.
Managing the parts store like a warehouse. Maintenance software manages parts needed for interventions: inventory, consumption and thresholds. It is not a warehouse management system, and it should not be asked to become one.
Where to start
The order that works, in our experience:
The equipment hierarchy—the structured list of everything you maintain, with its location and criticality. It is the foundation for everything else and can be built without software.
Recording interventions, including corrective work, and attaching each to a piece of equipment. This creates the memory.
Allocating time and parts—the only way to approach cost by asset.
Preventive maintenance, once you know what fails and why.
Renewal decisions, once the first three years of history exist.
Many organisations start with point 4 because it is the most visible and most rewarding. It is also the one that produces the least result when the first three are missing.
