The CFO’s Fleet Management and Utilization TCO Model: Why Utilization Never Appears in Your Cost Per Mile

Why utilization is missing from your model

Ask a finance team for the fleet management and utilization TCO model and you will get a good model of the first half. Acquisition or lease, financing, fuel, maintenance, tyres, insurance, licensing, telematics, driver compensation, and depreciation, built per vehicle per year and rolled up to the fleet.

Utilization will not be in it, and the reason is structural rather than an oversight.

Every line in that model is a cost. Utilization is not a cost. It is a measure of how much output the fleet produced against the capacity it had available. So when utilization improves, no line in the TCO model moves. Fuel might rise slightly, since the vehicle did more work. The model, read literally, shows the improvement as neutral or marginally negative.

That is a denominator problem. TCO per vehicle per year answers what the fleet costs. It does not answer what a delivery costs, and utilization only shows up in the second question, because utilization is what determines how many deliveries the same cost base produced.

The practical result is that utilization investment is difficult to justify inside the model finance already runs, which is why it competes poorly against fuel and maintenance initiatives that show up as line reductions. The fix is not a better argument. It is a second denominator.

Also Read: Fleet Utilization Rate: How to Measure it, What Good Looks Like, and How AI Closes the Gap

What the model contains, and what each line is sensitive to

A fleet management and utilization model starts by classifying every cost line by whether utilization affects it. Most do not, and that is the finding.

The cost structure explains why. ATRI’s operational cost data puts average cost at 2.26 dollars per mile in 2024, with driver compensation at roughly 44 percent of operating cost, equipment at roughly 28 percent, and fuel at roughly 21 percent, and records non-fuel marginal costs at a record 1.779 dollars per mile. Roughly seven tenths of the cost base sits in categories that accrue regardless of whether the asset is productively loaded.

That is the case for utilization stated in cost-structure terms: you are already paying for the capacity, and utilization determines how much of it you receive.

The final row is the model change. Add cost per delivered unit alongside cost per vehicle, using deliveries or cases rather than miles, since miles are an input and deliveries are the output finance is funding.

The four levers worth quantifying

Each is modellable from data most operations hold, and each maps to a different part of the gap between capacity paid for and capacity used.

1. Capacity fill

The share of available vehicle capacity actually loaded. The clearest lever, and the one with a credible external anchor: Chalmers University of Technology research indicates that optimised consolidation can raise vehicle fill rates from approximately 45 percent to approximately 74 percent.

How to model it. Measure current fill by route type over a quarter, model an improvement, and express the result as fewer vehicle-days required for the same volume rather than as a cost reduction. Vehicle-days is the unit that converts cleanly into either capacity avoidance or absorbed growth.

2. Empty running

Miles run without load. ATRI puts deadhead at approximately 16.7 percent of all truck miles, and notes this share-of-miles figure should be used rather than the commonly quoted claim that a third of trucks travel empty, which measures something different.

How to model it. Empty miles carry fuel and maintenance cost with no offsetting output. Model a reduction against your own empty mile share, and be conservative: some empty running is structural in a network with directional imbalance and cannot be removed by better allocation.

3. Non-productive time

Shift hours in which the asset and driver are paid and not producing: dock waiting, facility dwell, idle between assignments, and time lost to plans that could not execute.

How to model it. This is where the largest number usually sits, because it consumes the two biggest cost lines simultaneously. McKinsey has found that static planning models can leave as much as 60 percent of operating hours either understaffed or overstaffed, which is the structural inefficiency this lever addresses.

Model it as hours recovered, costed at fully loaded driver and asset rates, then converted into additional output capacity rather than into headcount reduction.

4. Capacity avoidance

The vehicles and contracted capacity you do not add because existing capacity absorbed the growth.

How to model it. Against your actual growth plan rather than hypothetically. If the plan adds vehicles at a defined volume threshold, model the deferral in months and value it at the capital and operating cost of the deferred units. Present it as deferral rather than elimination, since growth eventually requires capacity regardless.

This is usually the strongest line in the business case, because it is the counterfactual finance is already fluent in: capital not spent. It is also the easiest to verify after the fact, which matters for credibility on the next cycle.

Also Read: How Fleet Utilization Impacts Last-Mile Delivery Costs: Five Economic Mechanisms Enterprise Logistics Leaders Should Understand in 2026

Modelling a mixed owned, contracted, and gig fleet

Fleet management and utilization economics differ by pool, and a single blended cost per mile across a mixed fleet hides the decision the model is supposed to inform.

The three pools have different cost structures. Owned capacity is largely fixed, so utilization is the only lever available on it. Contracted capacity is semi-variable, priced per movement or per day with commitment terms. Gig capacity is closest to variable, priced per task with no idle cost and a higher marginal rate.

Three modelling consequences follow.

Model each pool separately, then model the allocation. The saving from better utilization on owned capacity and the saving from better allocation between pools are different numbers with different mechanisms, and combining them invites double counting.

Fixed-cost absorption is the owned-fleet argument. Every additional delivery on an already-paid-for vehicle carries only marginal fuel and maintenance cost. That is the highest-margin output in the network and it is only available through utilization.

Overflow architecture is a modellable cost. Many operations fill owned capacity first and pass the remainder outward, which means the most expensive marginal capacity is used on the highest-volume days. Model the alternative, allocating across all pools in one decision, as a cost-mix improvement rather than as a utilization gain, and keep the two separate in the model.

What to leave out

Three exclusions that buy credibility for everything included.

Headcount reduction you will not take. If the intent is absorbing growth without hiring rather than reducing the team, model it that way. Finance can check a claimed reduction against next year’s plan.

Double counting between fill and empty running. A consolidated load raises fill and reduces empty miles, and counting the same trip in both lines inflates the total. Define the boundary and state that you have.

Utilization benchmarks from vendor sources. Circulating fleet utilization benchmarks by vertical do not trace to research firms. Private-fleet survey data exists but must be purchased, and public vendor figures should not appear in a model finance will scrutinise. Use your own baseline and label the external anchors that are research-grade, such as the ATRI and Chalmers figures above.

Also Read: Fleet Management and Utilization: How AI Architecture Improves Capacity, Cost, and Performance in 2026

Which assumptions carry the model

Run sensitivity on three, because these determine whether the case holds.

Output growth. Capacity avoidance is worth nothing if volume is flat, and it is the largest line if volume is growing. State the growth assumption explicitly and show the case at zero growth.

Fill improvement achievable. The Chalmers range describes optimised consolidation, not a guaranteed outcome for your network. Model a fraction of it in the base case.

Non-productive time recoverable. Some dwell is outside your control, particularly at customer facilities. Model the share you influence rather than the total.

For payback framing, a Gartner-commissioned analysis indicates more than 40 percent of TMS adopters break even within 6 to 12 months and a further 25 percent within 18. Use it as a category range and label it as such, with your own numbers carrying the base case.

Also Read: The Empty-Mile Problem: The Fleet Cost Hiding Behind Healthy Utilization in 2026

What the deployments show

Two cases give utilization value in the form a CFO can use.

A Fortune 50 parcel and logistics provider governing 4,500+ drivers across 51 active service-centre locations ran a single-site analysis that surfaced 565,000 dollars in unused capacity, including premium-tier service given away on cheaper classes. Scaled across 25 sites, that reached 14 million dollars-plus annualised, alongside weekly execution moving from 75 percent to 92 percent.

Two things make this the most useful reference in a utilization business case. The value was capacity that already existed and was not visible until allocation was instrumented, which is precisely the argument this model is built to express. And the finding was partly commercial rather than operational, since premium service on cheaper classes is a pricing leak that no fuel or maintenance line would ever have surfaced.

A global FMCG leader operating across ten countries with 1,000+ distributors eliminated 12,000+ trips each month through demand-matched capacity and fuller loads, alongside 15 percent less distance travelled and a reported 3X return on investment.

Note the distinction between those two numbers, because it matters in the model. Distance reduction is a variable cost saving that appears in fuel and maintenance. Trip elimination is capacity created, which appears only in cost per delivered unit. Tools that move only the first report the smaller number.

Where Locus fits

Locus, the world’s first Decision-Intelligent, Agentic TMS, operates on the decision that determines utilization rather than on the monitoring that reports it. Within DiSCO, the Capacity agent forecasts demand and right-sizes available resources, the Dispatch agent plans and re-sequences against 250+ real-world constraints, and the Carrier agent allocates across owned, contracted, and gig capacity in one decision rather than in sequence.

That single-decision allocation is the mechanism behind the mixed-fleet modelling above, since the overflow architecture it replaces is where cost-mix losses accumulate.

Locus has been recognized by Gartner for seven consecutive years, featured in the 2026 Hype Cycle for Supply Chain Execution and Logistics Technologies, named a Leader in TMS by QKS Group (SPARK Matrix), and ranked #1 in Route Planning on G2’s 2026 Best Software Awards. In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Locus continues to operate independently.

Also Read: Mixed EV-ICE Fleet Cost-Per-Mile: Why Utilization Decides EV Payback in 2026

The line to add before the next budget cycle

Add cost per delivered unit to the fleet management and utilization report, next to cost per vehicle and cost per mile.

Cost per vehicle tells you what the fleet costs. Cost per mile tells you what movement costs. Neither tells you what output costs, and output is what the business is funding. Until that line exists, utilization improvements are invisible in finance reporting and will lose every budget argument to a fuel initiative that shows up as a line reduction.

It takes one calculation from data you already hold, and it changes which investments look defensible.

See how Locus can improve your fleet utilization and management, schedule a demo here.