Preventive Maintenance Scheduling: The Interval Is Not the Problem

Most fleets do not have a preventive maintenance problem. They have a preventive maintenance scheduling problem, and the two are easy to confuse.

The PM program itself is usually sound. Someone wrote it; someone defends it, keeping trucks legal and running. What preventive maintenance scheduling rarely examines is the number underneath it – the interval. Every 15,000 miles (about 24140.16 km). Every 250 engine hours. Every 90 days (about 3 months), whichever comes first.

That number was always set as a proxy. It stands in for a question nobody could answer directly at the time it was written: how much life is left in this component, on this asset, in this duty cycle? A fixed interval is a reasonable answer to that question when you have no visibility into the asset between service events. It is a much weaker answer when the asset is transmitting condition data continuously.

A fixed interval is an average applied to a fleet that is not average

The interval assumes your assets are interchangeable. They are not.

Two tractors with sequential VINs, spec identically, delivered the same week, will not wear the same way. One runs regional line-haul on interstate at steady load. The other runs urban stop-and-go with a heavy PTO cycle and a driver who idles through every dock to wait. By month eighteen, those two assets are not the same machine, and servicing them on the same calendar is a decision to overserve one and underserve the other.

Multiply across a mixed fleet – different vocations, terrains, seasons, drivers, and body types – and the fixed interval stops being a maintenance strategy. It becomes an averaging error that you found twice.

Where the money leaks

The cost of a mistuned interval does not show up as a line item. It shows up in three places that are usually owned by three different people.

Servicing too early

An asset comes into the bay because the calendar said so, not because anything is wrong. You spend the parts, you spend the labor, and – more expensively – you spend the bay hour and the vehicle's availability. Fluid and filter life gets thrown away with usable service left in it.

This is the leak that is hardest to see, because nothing goes wrong. A too-early PM looks exactly like a well-run maintenance program on a report. Failure is silent and it repeats every cycle.

Servicing too late

The mirror image is the asset whose duty cycle is harsher than the interval assumed. It reaches the PM already degraded, or it fails between services. Now you are paying for the repair, the tow, the missed delivery, the driver's downtime, and the emergency-rate parts – and the repair is unplanned, which means it lands in the schedule of a shop that was already full.

Servicing the wrong thing

The third leak is the subtlest. The PM happens on time, the checklist is completed, and the component that actually fails three weeks later was never on the checklist – because the checklist reflects the OEM's assumptions about the asset, not your observed failure history for that asset class in that application.

If your work-order history is not coded consistently, you cannot see this pattern at all. Which is why the fix has less to do with scheduling software than most vendors will tell you.

What condition-based preventive maintenance scheduling requires

"Move to condition-based maintenance" is easy to say and frequently oversold. It is worth being specific about what must be true before the interval can safely flex.

A live condition signal. Telematics and ECM data must reach one place, continuously, for every asset in scope - not just the newest of the fleet. Mixed-vintage and mixed-OEM fleets are where this usually breaks down, and it is a real constraint, not a rounding error.

Service history can be analyzed. This is the one that stops most programs. If three technicians describe the same repair in three ways in free text, you have records but not data. Consistent coding of the work – VMRS or an equivalent structured scheme applied the same way every time – is what turns history into a failure curve you can schedule against.

Honest capacity math. Dynamic scheduling moves to work around. If your bays and technicians are already at capacity, a smarter schedule mostly reveals the constraint rather than relieving it. Know which problem you are solving.

A compliance floor that does not move. Some inspection and recordkeeping obligations are set by regulation and are not candidates for optimization at all. Condition-based scheduling applies to the discretionary layer above that floor.

A practical sequence, in the order that works

Fleets that get this right rarely do it in one move. The sequence below is deliberately unglamorous.

1. Segment the fleet by duty cycle, not by asset type. Group by how assets are used - measured, not assumed. Idle percentage, load profile, terrain, stop density, PTO hours. You will usually find three to five real groups where the org chart claimed one.
2. Fix the coding before you touch the intervals. Standardize how work is captured and requires it. This is a change-management project with a technology component, not the reverse. Expect it to take a quarter to stick.
3. Test one segment. Take the group with the clearest data and the lowest consequence of error. Adjust the interval in one direction, hold everything else constant, and run it long enough to see a real signal.
4. Let condition data override the calendar – with guardrails. Set the floor and the ceiling. The interval flexes inside that band based on observed condition; it does not flex outside it. This is what makes the change defensible to your safety group and your insurer.
5. Re-baseline on a fixed cadence. Duty cycles drift. Contracts, routes, and the segment definitions you built in step one will be staled within a year if nobody owns them.

What to measure

If you cannot show the change worked, it will be reversed the first time an asset fails. Track a small set of measures, and track them by segment rather than fleet-wide:

• Maintenance cost per mile, split into planned and unplanned. The whole thesis is the ratio between those two shifts.
• Unplanned events per asset per period – the number that justifies the program to operations.
• Bay utilization and PM completion timeliness – to confirm you have not simply moved the bottleneck.
• Parts and fluid consumption per asset-mile – where the too-early leak becomes visible.
• Roadside and out-of-service events – the safety-side counterweight that keeps the program honest.

Set the baseline before you change anything. A baseline reconstructed after the fact is an argument, not a measurement.