Every fleet says it does preventive maintenance. Most do scheduled maintenance that is frequently late, occasionally early, and rarely measured. The difference shows up in the unplanned maintenance ratio, and that number is the clearest single indicator of a fleet's operational health.
The objective
A preventive maintenance programme exists to move cost and downtime from unplanned to planned. Planned work is cheaper per hour, happens when the vehicle is not needed, uses parts you have in stock, and does not strand a driver at the roadside with a customer waiting.
The target is not zero failures. It is a defensible balance between the cost of servicing and the cost of failure.
Under-servicing is more visible. Over-servicing is more common. Both are corrected by evidence, not by instinct.
Interval design
Base the interval on the failure mode, not on habit. Different components degrade against different measures:
| Measure | Suits |
|---|---|
| Distance | Tyres, brakes, driveline, most consumables |
| Engine hours | Plant, refrigeration units, PTO-heavy vehicles, urban stop-start |
| Calendar time | Fluids that degrade with age, statutory inspections, certifications |
| Fuel consumed | A good proxy for engine work done |
| Duty cycle severity | Anything on a vehicle in extreme service |
Most schedules should be whichever comes first across at least two measures. A van doing 8,000 km a year still needs its brake fluid changed on time; a refuse vehicle doing modest distance at high engine hours needs servicing on hours.
Start from the manufacturer's schedule, then adjust for your duty cycle. Manufacturer intervals assume typical use. Urban stop-start, heavy loads, extreme temperatures, dusty environments and short journeys all shorten real component life.
Use your own failure data to refine. After 18 months of clean work order history, look at what actually fails between services. Components failing consistently before the interval need a shorter one; components never failing at the interval may support extending it. This is the difference between a programme and a copied table.
Tolerance windows
An interval without a tolerance is a schedule nobody can meet. Define, per service type:
- Target — the ideal point
- Tolerance — the acceptable window (often ±10% of the interval)
- Overdue — beyond tolerance, triggering escalation
- Critical — beyond which the vehicle should not operate
Then measure PM compliance: services completed within tolerance, as a percentage. Not services eventually completed — that number is always near 100% and tells you nothing.
Fleets moving from unmeasured to measured PM compliance typically find their real figure much lower than expected. That discovery is the first genuine step in the programme.
Scheduling that works in practice
Forecast, do not react. Use telematics odometer data to project when each vehicle will reach its interval, and book workshop capacity weeks ahead. A programme driven by "overdue" alerts is permanently late by construction.
Batch by route and depot. Servicing vehicles from the same route on the same day reduces the disruption to operations.
Use natural downtime. Vehicles idle at weekends, during seasonal troughs, or between shifts. Servicing then costs no operational availability at all.
Combine services. Bundling a due inspection with a service already booked saves a whole downtime event.
Reserve capacity for the unplanned. A workshop booked to 100% planned capacity cannot absorb a breakdown without displacing planned work, which is how programmes decay.
Measuring the programme
| Metric | What it tells you | Direction |
|---|---|---|
| PM compliance % | Whether the schedule is real | Up |
| Unplanned : planned cost ratio | Programme effectiveness | Down |
| Roadside breakdowns per 100k km | Failure prevention | Down |
| Mean time between failures | Reliability trend | Up |
| Repeat repairs within 30 days | Repair quality | Down |
| Availability % | The operational outcome | Up |
| Average PM cost per vehicle | Efficiency | Watch alongside failures |
The unplanned-to-planned ratio is the headline. A fleet where unplanned work represents a large share of maintenance spend is firefighting; one where planned work dominates is managing.
Transition from reactive to planned
Fleets in a reactive state cannot jump straight to a mature programme. A workable sequence:
- Get accurate mileage. Telematics-fed odometer, not manual entry. Nothing else works without this.
- Build the asset register properly, with correct classes and specifications.
- Apply manufacturer schedules as a starting point, unadjusted.
- Measure compliance and publish it, without blame, for three months.
- Book ahead rather than reacting to overdue alerts.
- Capture work orders properly — every job, with parts, labour and cause.
- Analyse failures after 12–18 months and adjust intervals on evidence.
- Introduce condition-based triggers from telematics fault codes and wear data.
Expect the transition to take a year, and expect maintenance cost to rise slightly before it falls, as deferred work is caught up.
Frequently asked questions
How often should vehicles be serviced?
Start with the manufacturer's schedule adjusted for your duty cycle, then refine using your own failure data. There is no universal interval — a van doing motorway miles and one doing urban multi-drop wear out on completely different timescales.
What PM compliance percentage should we target?
Aim for a high figure within tolerance, but the more useful goal is a rising trend and a falling unplanned maintenance ratio. A fleet reporting near-perfect compliance with frequent roadside failures is measuring the wrong thing or the tolerance is too wide.
Should we service on distance or on time?
Both, with whichever occurs first. Low-mileage vehicles still degrade — fluids age, seals dry, brakes corrode — and a distance-only schedule leaves them unserviced for years.
How do we handle vehicles that are always needed?
Schedule around natural downtime, use spare or hire vehicles to cover, and book well ahead. A vehicle that can never be released is a vehicle that will eventually fail unexpectedly, which costs more availability than the service would have.
Is condition-based maintenance better than interval-based?
It is more efficient where you have reliable condition data — oil analysis, fault codes, wear sensors. Most fleets should run interval-based schedules with condition-based triggers layered on top, rather than replacing one with the other.
