Choosing a maintenance strategy for your fleet looks like a technical decision. It’s actually a cost decision: each approach moves the spend to a different place — onto the spare part, the scheduled job, or the vehicle downtime. The right question isn’t “which maintenance is best”, but “where do I want the cost to land, and how much risk am I willing to carry”.
This article compares the three logics — corrective, preventive and predictive — from the point of view of someone running a fleet, not a plant. For the complete guide to the predictive approach alone, see Predictive maintenance for fleets: how it works and when it pays off.
Corrective: repair after the failure
Corrective maintenance is the simplest: you act when the component fails. No plans, no monitoring — until something breaks, you do nothing.
On paper it costs little, because you don’t spend until you have to. In a fleet’s reality it’s the most expensive, for one reason: the cost isn’t the repair, it’s the unplanned downtime. A van that dies mid-route doesn’t just cost the part — it costs the missed deliveries, the overtime to recover them, the possible rental, the call to the customer. The failure, by definition, arrives at the worst moment.
When it makes sense: on non-critical, low-cost components where the failure doesn’t strand the vehicle (a bulb, an accessory). As a fleet strategy, almost never.
Preventive: the manufacturer’s calendar
Preventive maintenance is what most fleets do: interventions at fixed intervals — every so many kilometres or months — following the manufacturer’s plan. Service, belt, brakes, filters: all on schedule.
It’s a huge step up from corrective, because it turns downtime from unexpected to planned. But it has two structural limits.
The first: you act even when it isn’t needed. The plan is calibrated on the average case, not your vehicle. A van running light, calm urban kilometres replaces components that were still good, just because time has passed.
The second, more insidious: sometimes you act too late. The calendar doesn’t know that one vehicle works twice as hard as the others, or that its driver brakes harshly. Between one service and the next, a stressed component can fail anyway — and you’re back to corrective, with the very downtime you wanted to avoid.
When it makes sense: it’s the reasonable baseline for any fleet, and for many components it remains the right choice. The point isn’t to abandon it, but to understand where the calendar isn’t enough.
Condition-based and predictive: from data, not the calendar
This is where data enters. Condition-based maintenance (CBM) acts when a real parameter crosses a threshold — not on a fixed date, but when the vehicle signals it. Predictive maintenance goes one step further: it uses the trend of the data over time to estimate when that parameter will reach the threshold, giving margin to schedule.
The difference between the two is maturity, not nature: condition-based reacts to the current state (“battery voltage is low, check it”), predictive projects the trend (“at this rate the battery fails within two weeks”). Both rely on the vehicle’s telematics data — diagnostic codes, real mileage, battery voltage, driving style.
The advantage is that you act at the right time on the right vehicle: neither too early like preventive, nor too late like corrective. The drawback is that it requires a data source and a minimum of logic to read it — it’s not something you do with a paper diary.
When it makes sense: on critical components and high-utilisation vehicles, where the divergence between calendar and real wear is greatest — and where downtime costs the most. For the detail on which data you need and what’s realistic today, see the predictive maintenance guide.
The comparison that counts for a fleet
Summing up on the four axes that really weigh on a fleet’s budget:
- Direct cost (parts and jobs): preventive is the most predictable; condition-based/predictive reduces it by eliminating needless work; corrective looks low but hides the downtime cost.
- Unplanned downtime risk: maximum in corrective, medium in preventive (the calendar doesn’t see everything), minimum in predictive.
- Data required: none for corrective and preventive; telematics and thresholds for condition-based and predictive.
- Management complexity: minimal for corrective (but costly in downtime), low for preventive (a diary), medium for predictive (you need a system that collects and reads the data).
A vehicle’s total cost of ownership isn’t decided on the price of the spare part, but on how many breakdowns you avoid and how much needless work you cut. That’s where the maintenance strategy moves the fleet TCO.
It’s not “pick one”: it’s a mix per fleet segment
The most common mistake is hunting for the strategy. Well-run fleets don’t use one — they combine them, based on how critical a vehicle is and how much data it produces.
A typical mix:
- Critical, high-utilisation vehicles (the ones that cost you the day if they die) → condition-based/predictive on the key components, using the telematics data.
- Standard post-2019 vehicles → preventive as the baseline, enriched by the diagnostic alerts that arrive from the connected vehicle anyway.
- Marginal or low-utilisation vehicles → pure preventive: the calendar is enough, more isn’t worth it.
- Non-critical components → conscious corrective: you repair when it fails, because the downtime doesn’t hurt.
The good news is the move isn’t a leap: a fleet that already has telematics for utilisation rate and fuel already has the data source to start shifting its critical vehicles towards predictive, with no new investment.
How a fleet moves from preventive to predictive
The realistic path is incremental, not a factory project:
- Collect the data the vehicle already produces — via Cloud OEM on post-2019 vehicles or OBD on older ones.
- Start from the components that strand you most — typically battery, brakes, tyres. Set sensible thresholds, not complex models.
- Keep preventive as a safety net on the rest: you don’t throw away the calendar, you flank it with data where it matters.
- Connect the alerts to delivery planning, so a “check me” vehicle doesn’t end up on a critical route.
Along this path, OptivoTrack provides the foundation: it reads diagnostic codes, manages service schedules on each vehicle’s real usage data (not a single date for everyone) and tracks the driving style that anticipates wear. It’s the concrete bridge between calendar-based maintenance and data-based maintenance — without promising a predictive model that guesses the failure to the kilometre, which on mid-sized fleets is still premature.
Frequently asked questions
What’s the difference between preventive and predictive maintenance?
Preventive follows fixed intervals, the same for the whole fleet (every so many km or months). Predictive is based on the real state of each vehicle, inferred from telematics data: you act when the signals indicate it, not when the calendar says so.
What about condition-based maintenance?
It’s the step between preventive and predictive: it acts when a real parameter crosses a threshold (“low battery voltage”). Predictive adds the projection over time, estimating when the threshold will be reached. Both rely on vehicle data.
Do I have to abandon preventive maintenance?
No. Preventive remains the reasonable baseline for most of the fleet and most components. Predictive doesn’t replace it: it flanks it on critical, high-utilisation vehicles, where the calendar isn’t enough and downtime costs more.
Where should I start?
From the vehicles that cost you the day if they die, and from the components most frequently subject to failure (battery, brakes, tyres). If you already have telematics for position and fuel, you already have the data source to start, with no new investment.
Go deeper: Predictive maintenance for fleets: how it works and when it pays off — the complete guide to which data you need, what’s realistic today and how to start.
In the glossary: Predictive maintenance · Telematics · Total cost of ownership