One way of working across every plant, with room for local reality.
A practical framework for a central maintenance team: what to standardise, how to measure it, and how to roll it out across several plants without breaking what already works.
This is a general framework based on common industry practice (for example ISO 55000 asset management, EN 15341 maintenance indicators and SMRP metrics). It is not the internal standard of any company. Targets shown are indicative and should be calibrated for each plant.
Each pillar defines what must be true in every plant, and one or two measures that prove it.
People know who owns what, and the team has the skills the plant needs.
Measures: skills-matrix coverage, training hours per technician
Work is planned before it is done, and effort goes where failure costs most.
Measures: PM compliance, planned-work ratio, backlog in weeks
Every serious failure teaches the plant something, and the lesson is not lost.
Measures: MTBF, MTTR, breakdown frequency, repeat-failure rate
The right part is on the shelf when needed, without cash tied up in stock that never moves.
Measures: stock-out rate, inventory turnover, slow-moving share
Clean, consistent data is what makes comparing plants possible at all.
Measures: master-data completeness, orders with correct failure codes
Maintenance is done safely, and each plant is checked against the same yardstick.
Measures: audit score, findings closed on time, safety incidents
A KPI only works across plants if every plant calculates it the same way. These are the core definitions.
| KPI | Definition | Indicative target | Review |
|---|---|---|---|
| OEE | Availability × Performance × Quality | Set per line against its own baseline | Daily / weekly |
| MTBF | Operating time ÷ number of failures | Rising trend per asset | Monthly |
| MTTR | Total repair time ÷ number of repairs | Falling trend per asset | Monthly |
| PM compliance | PM orders completed on time ÷ PM orders scheduled × 100 | 90% or higher | Weekly |
| Planned-work ratio | Planned maintenance hours ÷ total maintenance hours × 100 | 80% or higher | Monthly |
| Breakdown frequency | Number of unplanned stops per month, per line | Falling trend | Weekly |
| Stock-out rate | Requests for a part not in stock ÷ total requests × 100 | Below 3% for critical parts | Monthly |
| Maintenance cost ratio | Annual maintenance cost ÷ replacement value of assets × 100 | Compare with industry range | Quarterly |
The same ladder is used to assess every plant, so the central team can see where to help first. You can score your own plant in five minutes.
Work starts when something breaks. Little planning, few records, and heavy dependence on a few individuals.
PM schedules exist and most are done, but task lists are generic and spare parts are managed by habit.
Work is planned and scheduled weekly, criticality is defined, KPIs are measured, and data in the CMMS can be trusted.
Root cause analysis is routine, condition monitoring supports PM decisions, and the plant improves from its own data.
Maintenance is managed as part of total asset cost. Strategy is reviewed regularly, and the plant shares practice with others.
Standards fail when they are announced instead of built with the plants. This is the sequence I would follow.
Score every plant against the maturity levels to set a baseline. Try the assessment.
Write the standard with plant representatives, not only for them.
Test it in one plant and fix what does not work in practice.
Move plant by plant in waves, with training and a named local owner.
Re-assess regularly, share the best practice, and update the standard.
A short list of rules every plant must follow: definitions, data, criticality, safety and reporting. This is what makes plants comparable.
Everything else adapts to each plant's products, equipment and people, as long as the core is respected.
I am glad to talk about assessments, standards and training for maintenance teams. See the free downloads or contact me.