Cost per mile is the most quoted and least consistently calculated figure in fleet management. Two fleets quoting the same number can be measuring completely different things, which makes benchmarking against others largely pointless and internal trend analysis essential.
The components
Fixed costs — incurred whether the vehicle moves or not:
- Depreciation, or lease and finance charges
- Insurance
- Road tax and licensing
- Operator licence and compliance costs
- Permits and zone charges where paid annually
- Allocated overhead (workshop, fleet administration, telematics subscriptions)
Variable costs — driven by distance and use:
- Fuel or energy
- Tyres
- Maintenance and repair, parts and labour
- Consumables and fluids
- Tolls and per-use zone charges
- AdBlue or equivalent
Sometimes included, sometimes not — and this is where comparisons break:
- Driver wages and on-costs
- Downtime cost
- Accident and claim costs
- Corporate overhead
'Cost per mile' without a definition is not a number. Publish yours, then hold it constant.
The calculation
`` Cost per mile = (fixed costs + variable costs) ÷ miles travelled ``
Both over the same period, and per vehicle or per vehicle class.
Depreciation is the component most often mishandled. Options:
- Straight line: (purchase price − expected residual) ÷ expected life. Simple, stable, understates early-life cost.
- Market value based: actual value decline each year. More accurate, more volatile, requires valuation data.
- Accounting depreciation: whatever finance uses. Consistent with the accounts, sometimes unrelated to economic reality.
Use market-value depreciation for operational decisions such as replacement timing, and accounting depreciation for financial reporting. Say which you are using.
Overhead allocation should be simple and defensible. Total fleet overhead divided by fleet miles, or by vehicle count for genuinely fixed items. Elaborate allocation models generate argument and rarely change decisions.
Segment or the number is useless
A fleet-wide cost per mile hides everything actionable. Segment by:
- Vehicle class — a 3.5-tonne van and a 26-tonne rigid are not comparable
- Age band — old vehicles cost more, and the profile tells you when
- Duty cycle — urban multi-drop versus motorway trunking
- Depot — variation reveals process differences
- Individual vehicle — the outliers are where the money is
The individual vehicle view is the most useful. Ranking every vehicle by cost per mile within its class and investigating the top decile is the highest-return use of the whole calculation.
Using it
| Question | How cost per mile helps |
|---|---|
| Which vehicles to replace | Rising trend versus class average |
| Which specification to buy next | Compare classes on lifetime cost per mile |
| Whether a depot has a problem | Compare like-for-like across sites |
| What to charge for a contract | Cost base plus margin, with confidence |
| Whether an efficiency programme worked | Trend against a frozen baseline |
| Lease versus buy | Compare total cost per mile under each |
Common errors
- Mixing periods. Annual fixed costs against a month of mileage.
- Excluding downtime. The vehicle that is off the road half the year has a wonderful cost per mile and is a terrible asset.
- Including driver cost in some comparisons and not others. Especially when benchmarking externally.
- Using odometer readings of uncertain provenance. Manual entry decays; use telematics where possible. See telematics data quality.
- Ignoring seasonality. Compare like periods, or use rolling twelve months.
- Allocating shared costs arbitrarily and then treating the result as precise.
- Benchmarking against published figures whose definitions you do not know.
Cost per hour
For vehicles where distance is a poor proxy for use — refuse collection, plant, urban service vehicles that spend hours stationary with the engine running — cost per engine hour is more meaningful. Compute both; the divergence between them is itself informative about duty cycle.
Common questions
Should driver cost be in cost per mile?
For operational fleet management, usually not — it obscures the vehicle's own cost. For pricing a service or comparing transport modes, yes. Report both, clearly labelled, rather than arguing about which is correct.
How often should we calculate it?
Monthly for trend monitoring, with a rolling twelve-month view to smooth seasonality. Annual calculation is too infrequent to detect a deteriorating vehicle before it becomes expensive.
Can we benchmark against other fleets?
Only with extreme caution, because definitions, duty cycles and vehicle specifications differ so much. Internal comparison — across depots, classes and time — is far more actionable and far less likely to mislead.
What is a good cost per mile?
There is no universal answer; it depends on vehicle type, duty cycle, geography and what you include. The useful question is whether yours is falling relative to your own baseline, and which vehicles sit above their class average.
How do we handle electric vehicles in the same calculation?
Use the same structure with energy replacing fuel, and be careful with depreciation, since used-EV residual values are less established. Include charging infrastructure cost allocated across the electric fleet, or the comparison flatters the electric vehicles.
