How to calculate cost per vehicle per day
Sum every cost attributable to a unit and divide by the days it was in your custody, then break the total down by phase. The arithmetic is trivial. Trustworthy day counts and honest phase attribution are the real job, and where most fleets go wrong.
1. Decide what goes in the numerator
Include what accrues while the vehicle is handled or waiting: yard rent, transport legs, recon labor and parts, wash, charging or fuel, title and registration admin, and handling staff time. Exclude depreciation, insurance, and financing unless finance asks for a fully loaded figure. Mixing operating and ownership costs gives operations a number it cannot improve.
2. Fix the denominator before you trust anything
Most calculations fail here. Custody days, calendar days, and billable days are three different counts, and mixing them swings the result widely. Pick custody days, from the recorded intake to the recorded release, and apply that definition to every unit.
3. Attribute costs to the VIN, not the site
A site invoice split evenly across units is an average in disguise. Allocate on recorded presence and phase. A unit that sat 26 days and one that sat two did not cost the same, and treating them as equal hides the behavior you are trying to find.
4. Break the total into phases
A blended figure tells you the number is too high. A phase breakdown tells you where. Split it across intake, inspection, readiness, storage, transport, and redeploy, and the fix usually shows itself, often one vendor-dependent phase absorbing too much dwell.
5. Sanity-check against the outliers
Before you present the figure, check the longest-dwelling 5% of units and confirm they are real, not records nobody closed. Open aged units inflate the mean and are the top reason a figure gets challenged. If they are real, they belong in the number and are your first improvement target.
Four errors that make the number indefensible
Averaging across programs. Long-term storage and fast-turn rental should not share a denominator. Report per program and client, then roll up.
Calendar-month boundaries. Units that span month ends get counted twice or not at all. Count custody days from event timestamps, not the accounting calendar.
Excluding the units you would rather not discuss. Removing aged or problem units to get a cleaner figure is the fastest way to lose credibility when someone recomputes it from raw data.
Estimating dwell after the fact. If dwell is inferred from arrival and departure dates rather than recorded per phase, the phase breakdown in step four is fiction, and the phase breakdown is the only part of the exercise that changes behavior.
Making it repeatable
A figure that takes a week to assemble gets built once and argued about for a quarter. The one that survives is computed continuously from events already recorded, so phase timestamps must be a byproduct of the work. See cost per vehicle per day for how it is enforced, and the glossary definition.
Frequently asked questions
Not in the operating figure. Depreciation is a function of ownership and time rather than of handling, so including it obscures the part of the number operations can actually change. Report it separately, and combine the two only when finance explicitly wants a fully-loaded per-unit-per-day cost.
There is no credible cross-industry benchmark - the figure depends on market, program type, vehicle class, and how much handling the contract includes. The useful comparison is your own number across sites and programs, and the trend over quarters. Treat any vendor quoting a universal target with suspicion.
It stays in your custody days if the unit is your responsibility while it is out - which it usually is. The shop time becomes its own phase, which is precisely the attribution that makes vendor turnaround visible instead of buried in a blended average.
Thirty minutes, screen-shared. We take one of your vehicles through every gate in the live product.