Skip to content

25 Essential Fleet KPIs, With the Formula for Each

Most fleet KPI lists tell you what to measure and leave out how it is worked out, which is where the arguments start. This sheet gives the formula alongside every metric, so two people reading the same number are calculating it the same way.

The problem with a fleet KPI is rarely the KPI. It is that the depot, the workshop and finance each compute it slightly differently and none of them know it. Utilisation counted against owned vehicles is a different number from utilisation counted against available vehicles, and the gap between those two figures is exactly the conversation nobody is having.

So every row here carries its arithmetic. Where a metric can be computed two defensible ways, the sheet says which one it means. Pick a definition, write it down, and use the same one every month — a mediocre metric measured consistently beats a better metric that changes shape whenever someone new builds the report.

Nothing here is a target. There are no benchmark figures on this page, because a benchmark drawn from someone else's duty cycle is worse than no benchmark at all. Compare against your own last quarter, or against your own other depot. That comparison you can defend.

Cost and spend

KPIWhat it tells youFormula
Cost per kilometreWhat a kilometre of work actually costs once every line is counted. Only comparable between vehicles doing similar work.Total operating cost ÷ Total distance run
Cost per vehicle per monthThe figure that survives outside the fleet department, because everyone has an instinct for what it should be.Total annual fleet cost ÷ Vehicles ÷ 12
Fuel cost per kilometreIsolates the largest variable line. Moves with fuel price as much as with driving, so track both.Fuel and energy spend ÷ Distance run
Fuel consumptionThe engineering figure underneath fuel cost, unaffected by price movement. Reported per hundred kilometres.(Litres burned ÷ Kilometres run) × 100
Maintenance cost per kilometreRises as a vehicle ages. The trend against a vehicle's own history matters far more than the level.Maintenance, tyre and repair spend ÷ Distance run
Total cost of ownershipThe whole life cost of an asset. Purchase price is the part everybody negotiates and the part that matters least.Depreciation + Fuel + Maintenance + Insurance and tax + Downtime cost
Fixed cost ratioHow much of your spend continues when the work stops. Decides whether cutting routes can save anything.(Fixed cost ÷ Total fleet cost) × 100

Utilisation and productivity

KPIWhat it tells youFormula
Fleet utilisation rateHow hard the vehicles that could work actually worked. Say out loud whether the denominator is owned or available.(Hours vehicles were in use ÷ Hours they were available) × 100
Vehicle availabilityThe share of the fleet fit to go out this morning. The number people guess at and get flatteringly wrong.(Vehicles fit to work ÷ Vehicles owned or leased) × 100
Distance per vehicle per dayA blunt productivity check per asset. Useful within one duty cycle, meaningless across a mixed fleet.Total distance ÷ (Vehicles × Days in period)
Payload utilisationWhether you are moving air. A fleet that cubes out before it weighs out needs the volume version instead.(Weight actually carried ÷ Rated payload) × 100
Empty running ratioKilometres you paid for that earned nothing. The clearest planning failure a fleet can measure.(Empty kilometres ÷ Total kilometres) × 100
Idle time ratioFuel and engine hours consumed while stationary. Often a loading bay problem rather than a driver problem.(Idle engine hours ÷ Total engine hours) × 100

Maintenance and reliability

KPIWhat it tells youFormula
Preventive maintenance complianceWhether the servicing you planned actually happened when it was due. The leading indicator for everything below it.(PM services completed on time ÷ PM services due) × 100
Mean time between failuresAverage run between things going wrong. Needs a firm definition of failure or it drifts every quarter.Operating hours in period ÷ Number of failures
Mean time to repairHow long a fix takes once it starts. Excludes waiting for parts, which is why it can look good while vehicles sit.Total repair hours ÷ Number of repairs completed
Vehicle off-road rateThe share of fleet days lost to the workshop, the parts queue and the damage nobody has quoted yet.(Vehicle days off road ÷ (Vehicles × Days in period)) × 100
Roadside breakdown rateFailures that happened in front of a customer rather than in a bay. Normalised per million kilometres.(Roadside breakdowns ÷ Kilometres run) × 1,000,000
Planned to unplanned ratioWhether the workshop sets its own week or has it set for it. Falls when preventive maintenance slips.Planned maintenance cost ÷ Unplanned repair cost

Safety, service and compliance

KPIWhat it tells youFormula
Accident rateIncidents normalised for exposure. A depot that drives twice as far will have more accidents at identical safety.(Recorded accidents ÷ Kilometres run) × 1,000,000
Harsh event rateBraking, acceleration and cornering events per unit of distance. Sensitive to the threshold you set, so freeze it.Harsh braking, acceleration and cornering events ÷ 100 kilometres
Speeding exposureDistance rather than count. One driver at nine over for an hour is a bigger risk than nine brief spikes.(Distance driven above the posted limit ÷ Total distance) × 100
On-time delivery rateService as the customer experiences it. Define the window before you measure, not after a bad month.(Jobs completed inside the promised window ÷ Jobs completed) × 100
Defect close-out rateWhether reported defects get fixed or just recorded. A high reporting rate with low close-out is a liability.(Defects closed within the agreed period ÷ Defects raised) × 100
Carbon per tonne-kilometreEmissions against work done, not against distance. Supply your own published fuel emission factor.(Litres burned × fuel emission factor) ÷ Tonne-kilometres carried

Frequently asked questions

Four: cost per kilometre, vehicle availability, preventive maintenance compliance and on-time delivery. Between them they cover money, capacity, the thing that protects capacity, and whether the customer noticed. Adding the other twenty-one before those four are trustworthy produces a dashboard nobody reads. The discipline is not collecting more metrics — it is being able to explain last month's movement in the ones you already have.

Because we would have to make them up. A published benchmark is an average across fleets with different vehicles, routes, load profiles, accounting policies and definitions of the word failure. Quoting one at your workshop manager invites an argument you deserve to lose. Your own previous quarter is a real baseline, measured with your definitions on your work, and it is the only comparison that supports a decision.

Neither is wrong, and that is the trouble. Utilisation against owned vehicles answers whether you are overfleeted. Utilisation against available vehicles answers whether the vehicles you can actually deploy are being worked. Both are legitimate; reporting them as one number is not. Pick one for the monthly pack, name it in the report, and keep the other for the fleet-size conversation.

Routinely. Mean time to repair improves if the workshop closes jobs that are still waiting for a part. Cost per kilometre improves if you send vehicles further. Harsh event rates fall if somebody quietly raises the threshold. Any metric attached to a bonus will be optimised, including in ways you did not intend. Pair each number with one that moves the opposite way — cost per kilometre with utilisation, repair time with breakdown rate.

The arithmetic is the easy half. The hard half is the inputs: distance from odometers nobody photographed, maintenance spend from invoices that arrive three weeks late, availability from a count someone did in the yard on Thursday. Fleet Analytics computes these from work orders, trips and fuel records as they land. The formulas on this page are the same ones. What changes is that the number is current and everyone is looking at the same one.

See these metrics on your own fleet

Fleet Analytics Software reports every one of these numbers from live operational data.