1. Corrective-to-preventive ratio
What it measures: the health of the operation, in a single figure.
How to calculate it: corrective hours divided by total maintenance hours.
Why it comes first: it tells you whether you are running maintenance or reacting to it.
The trap: there is no universal good number. What matters is the trend. If corrective work rises three months in a row, the plan is falling apart even if nobody has said so.
2. Preventive plan compliance
What it measures: whether the plan really exists.
How to calculate it: preventive work orders completed on time divided by those scheduled.
The trap: it can be dressed up by closing work orders that were never done, or by moving them before they fall due. Always read it next to the number of inspections postponed more than once. A 95% compliance rate with sixty inspections pushed back twice is not 95%.
3. Maintenance cost per asset
What it measures: which equipment is costing you money.
How to calculate it: hours times cost per hour, plus parts and travel, grouped by asset.
What it is for: it is the argument that wins a capital spending discussion. “This compressor cost $9,000 in repairs this year, and a new one costs $6,800” ends the conversation (example figures).
The trap: if parts are not charged to the asset, the cost comes out at half and the argument falls apart. And parts are exactly what people forget to record.
4. MTBF by equipment family
What it measures: reliability and, above all, how it changes over time.
The trap: comparing it across different families tells you nothing. Its value lies in comparing an asset with itself over time, or twin assets with each other. Two identical pumps with very different MTBF point to an installation or usage problem, not a pump problem. The details are in how to calculate MTBF and MTTR.
5. MTTR breakdown
What it measures: where the time goes in a repair.
How to calculate it: split MTTR into detection, travel, waiting for parts and actual repair.
Why it is on the list: the biggest stage is rarely the repair. Without the breakdown, meetings end up discussing technician productivity when the problem is in the parts room or in how requests come in.
The ones you can usually drop
- Maintenance cost as a share of replacement value. Useful in heavy industry, not very practical for a service company.
- Work orders closed per technician. It rewards closing fast, not closing well.
- Maintenance overtime. It is a symptom, not a management KPI.
- Composite indexes that blend five things into one number. When it drops, nobody knows why.
The rule before adding one more
A KPI that does not change any decision is decoration. Before adding one, answer a single question: if this number gets worse, what would I do differently tomorrow? If there is no answer, leave it out.
How to set them up without wasting time
- One single source. If each KPI comes from a different spreadsheet, the figures never match and the meeting is spent arguing about them.
- Every month, on the same day. You only see the trend if the cutoff is regular.
- With the detail at hand. Behind every figure there must be the work orders that make it up. Otherwise nobody can act on it.
Five KPIs reviewed every month are worth more than thirty in a report nobody opens. And if you can only look at one, look at the corrective-to-preventive ratio.
Last updated: