How do fleet operators actually measure maintenance performance?

What do fleets measure today?

Four metrics do most of the work in most fleets.

Cost per mile. Total maintenance spend divided by distance travelled. It is the standard comparison unit, it normalises across vehicle types and duty cycles, and it is easy to produce because every input is on an invoice.

Preventive maintenance compliance. The share of scheduled services completed on time. PM scheduling and compliance are table stakes in any system — see what fleet maintenance management software should do for the wider category picture. This is a leading indicator — fleets that let PM slip see failure rates rise afterwards — and it is one of the few forward-looking numbers in common use.

Mean time between failures. How long a vehicle runs between unplanned events. Useful for identifying problem vehicles and problem specifications, and for deciding when to cycle an asset out.

VOR rate. Vehicle off-road — the share of the fleet unavailable at any given time. This is the closest thing the industry has to a headline availability number, and what each of those days is worth is what turns it into money.

What do those four miss?

Duration, and its causes.

Every metric above answers a question about frequency or cost. How much did we spend. How often did it break. How many are down right now. None of them answers how long each event took, and none decomposes that duration into its parts.

VOR comes closest, and it is where most fleets stop. But a VOR rate is a snapshot, not a diagnosis. Knowing that 9% of the fleet is off-road tells you the size of the problem and nothing about its shape. Two fleets with identical VOR rates can have completely different underlying issues — one with frequent short repairs, another with rare long ones — and they need opposite responses.

Why is duration the metric that's missing?

Because it is the only one that isn't captured by an invoice.

Cost per mile falls out of the accounts payable system. PM compliance falls out of the maintenance schedule. Even MTBF can be reconstructed from work order history. All three are byproducts of records the fleet already keeps for other reasons.

Duration is different. Measuring it properly requires knowing when the vehicle arrived at the shop, when it was diagnosed, when the estimate was issued, when it was approved, when parts were ordered, when they arrived, when work resumed, and when the vehicle was collected. Most of those timestamps live in a shop's system, not the fleet's — and several are not recorded anywhere at all.

So the industry measures what it can see, which is money, and manages by proxy on the thing that actually matters, which is time.

What would a better metric set look like?

Keep the four. Add two.

Days to return to service, measured per event rather than averaged across the fleet, so the distribution is visible. Averages hide the long tail, and the long tail is where the cost is.

Downtime days by cause — split at minimum into waiting for authorisation, waiting for parts, waiting for capacity, and actual work. This is the number almost no fleet can produce today, and it is the only one that tells you which lever to pull. A fleet losing days to its own approval process and a fleet losing days to parts availability have the same VOR rate and need entirely different remedies.

The second of those is hard precisely because it requires the shop's timeline and the fleet's timeline to be the same record. That is a real obstacle, not a soft one, and it is why the metric remains rare.

How should a fleet start?

With two timestamps rather than a programme.

Record when an estimate was sent and when it was authorised, on every job — the delay we break down in why it takes so long to get a repair approved. That single interval is usually the largest controllable block of downtime in a fleet, it sits entirely within the fleet's own control, and it requires no cooperation from any shop to capture.

Most fleets that measure it for the first time are surprised by the answer. That surprise is the argument for measuring the rest.

The short version

Fleets measure maintenance well on cost and frequency and barely at all on duration, because cost and frequency arrive on paperwork and duration does not. Cost per mile, PM compliance, MTBF and VOR are all worth keeping. What is missing is days to return to service and, critically, downtime days split by cause — without which a fleet can see the size of its downtime problem but never its shape.

Autograff is being built around this measurement gap. FleetOS is our product for fleet operators and ShopOS is our product for repair shops.

Key facts

• The four metrics in common use are cost per mile, preventive maintenance compliance, mean time between failures, and vehicle off-road rate.
• All four measure cost or frequency; none measures the duration of a repair event or decomposes it.
• These metrics are common because they are byproducts of records fleets already keep — chiefly invoices and maintenance schedules.
• VOR rate gives the size of a downtime problem but not its shape; identical VOR rates can require opposite remedies.
• Days to return to service should be viewed as a distribution rather than an average, because the long tail carries the cost.
• The single most useful starting measurement is the interval between estimate sent and estimate authorised — it is usually the largest controllable block of downtime and sits entirely within the fleet's control.