Fleet Utilization and the Cost of Idle Time: A CFO Framework for Fresh Produce Peak Shipping

Why idle time is the most expensive line you do not track

Fleet cost reporting is organised around things you buy: vehicles, fuel, maintenance, insurance, drivers. Idle time is not a purchase, so it has no line, and it is consequently the largest unexamined cost in most fleets.

The reason it is expensive is structural. 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.

Read that composition against an idle vehicle. Driver compensation accrues. Equipment cost accrues. Insurance, financing, and licensing accrue. Only fuel pauses, and fuel is the smallest of the four. Roughly seven tenths of the cost base continues while the asset produces nothing, which makes an idle hour close to a full-cost hour with zero output against it.

The same logic applies to deadhead, with the difference that deadhead also burns fuel. 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 repeated claim that a third of trucks run empty, which measures something else.

For a CFO the useful framing is that fleet utilization is not an efficiency metric. It is the ratio between the capacity you are paying for and the capacity you receive, and idle time is the gap.

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

What changes during a produce surge

Idle and deadhead do not simply increase at peak. Their composition changes, which matters because the response differs.

Steady state idle is mostly gaps between assignments: a vehicle finishing early, waiting for the next load, or held because the plan was built for a different volume than arrived.

Peak idle is queuing. Packhouses run at capacity and dock queues lengthen at origin, while receiving facilities run at capacity and queues lengthen at destination. The vehicle is committed, the driver is on the clock, and neither is moving. This is the category that grows fastest during harvest and the one least visible in a utilization percentage, because the vehicle is technically assigned.

Facility-side dwell is well documented as a cost. ATRI found drivers were detained at 39.3 percent of all stops in 2023, losing between 117 and 209 hours per year depending on sector, at a cost of 3.6 billion dollars in direct expenses and 11.5 billion dollars in lost productivity. During harvest those hours concentrate, and in perishable freight they carry a second cost against the product.

Steady state deadhead is largely directional imbalance, which is structural and only partly addressable.

Peak deadhead is repositioning under time pressure. Volume appears where the plan did not expect it, so vehicles move empty to reach it. That deadhead is a planning artefact rather than a network characteristic, which makes it more addressable than the steady-state kind, and it is created by exactly the volatility harvest produces.

The planning failure behind both is documented. McKinsey has found that static planning models can leave as much as 60 percent of operating hours either understaffed or overstaffed, which describes a harvest operation planning daily against volume that revises hourly.

Why utilization is the only lever at peak

Three of the four ways to add capacity are unavailable once the season starts.

Buying vehicles takes longer than the surge lasts, and leaves you with capacity for eleven months you did not need it.

Hiring drivers faces the same timing problem and a market that does not cooperate. The ATA estimates a US driver shortage of roughly 60,000, projected above 170,000 by 2030, with annual turnover of 90 to 95 percent at large truckload carriers.

Buying spot capacity works and is priced for the moment. DAT data indicates spot rates typically run 15 to 30 percent above contract rates in normal markets and widen at peak, which is when you need them.

Recovering capacity you already own is the fourth, and it is the only one available on a timescale shorter than the season. That is what makes utilization a peak-season lever rather than an annual improvement programme.

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

Building the model

Four steps, using data most operations hold and few segment. The output is vehicle-days rather than dollars, for a reason covered below.

Step 1: separate idle into categories

Total idle is not actionable. Four categories are, and they have different owners.

Excluding the fourth explicitly matters, because including unavoidable idle in an addressable total is the fastest way to have the model discounted in review.

Step 2: separate deadhead into structural and planning-driven

Structural deadhead follows from network geography and directional imbalance. Planning-driven deadhead follows from repositioning to volume the plan did not anticipate.

Only the second is meaningfully addressable in-season, and at peak it is usually the larger of the two. Segment by whether the empty movement was scheduled in the original plan or created by a re-plan.

Step 3: cost the addressable portion honestly

Apply fully loaded driver and asset rates to idle hours, and fuel plus maintenance to deadhead miles. Use your own rates rather than ATRI’s, which are a US average and a composition reference rather than a cost input for your operation.

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

Then apply a recovery assumption and label it clearly as an assumption. No credible published figure states what proportion of idle or deadhead is recoverable through better allocation, and circulating percentages trace to software vendors rather than research. Model a conservative share, show the sensitivity, and let the range carry the case.

Step 4: express the result as vehicle-days released

This is the step that makes the model useful. Dollars saved from idle reduction is a soft number, because the cost did not disappear, it was redeployed. Vehicle-days released is a hard number, and it converts directly into either volume absorbed without additional capacity or spot purchases avoided.

For a peak-season case, the second conversion is the strongest, because a vehicle-day recovered in harvest week is a spot purchase not made at a 15 to 30 percent premium.

What to measure, and what not to

Measure utilization as a share of available shift time, not as a share of assigned time. A vehicle queuing at a dock is assigned and not producing, and the second measure hides exactly the category that grows at peak.

Measure idle by category, per the table above. An aggregate figure cannot be assigned to an owner.

Measure deadhead by whether it was planned, which separates structural from addressable.

Measure cost per delivered unit, not only cost per mile. Cost per mile falls when you drive more efficiently and says nothing about whether the asset was full. Cost per delivered unit is the only measure that moves when utilization improves.

Do not benchmark against published utilization figures by vertical. They do not exist at research grade. Private fleet survey data exists behind paywalls and public vendor figures should not appear in a model finance will scrutinise. Your own pre-season baseline is the comparator.

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

Where cross-fleet orchestration changes the arithmetic

Most produce operations run owned capacity alongside contracted carriers and, increasingly, spot. Utilization is usually measured and managed per pool.

That produces a specific and expensive pattern at peak. Owned capacity is filled first because it is already paid for, the remainder passes to contracted carriers, and what they reject goes to spot. The consequence is that the most expensive marginal capacity is used on the highest-volume days, which is exactly backwards, and it happens because the three pools are allocated in sequence rather than in one decision.

Allocating across all pools simultaneously changes two things. Owned capacity is used where it is genuinely the best fit rather than by default, and spot is purchased for the loads where it is cheapest overall rather than for whatever is left. For a CFO the modelling implication is that cost-mix improvement and utilization improvement are separate numbers with separate mechanisms, and combining them in one line invites double counting.

Where Locus fits

Locus, the world’s first Decision-Intelligent, Agentic TMS, operates on the decisions that determine utilization: how work is allocated, how routes are sequenced, and how the plan changes when volume revises.

Within DiSCO, the Capacity agent forecasts demand and right-sizes available capacity, the Dispatch agent plans and re-sequences against 250+ real-world constraints, the Carrier agent allocates across owned, contracted, and spot capacity in one decision rather than in sequence, and the Hub agent manages facility-side constraints where peak queuing accumulates.

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.

Two deployments quantify recovered capacity rather than cost saved, which is the form this model uses.

A Fortune 50 parcel and logistics provider ran a single-site capacity analysis that surfaced 565,000 dollars in unused capacity, including premium-tier service given away on cheaper classes, scaling to 14 million dollars-plus annualised across 25 sites. The capacity already existed and was invisible until allocation was instrumented, which is the argument for measuring idle by category rather than in aggregate.

A global FMCG leader operating across ten countries eliminated 12,000+ trips each month through demand-matched capacity and fuller loads, alongside 15 percent less distance travelled and plan run time falling from three hours to five minutes. The trip figure is capacity released; the distance figure is a variable cost saving. Keeping them separate in a model is the discipline that survives review.

Also Read: Predictive Capacity Planning for Peak Season: Building the Cost Model and Business Case in 2026

The pre-season calculation

Before harvest, pull last season’s peak weeks and segment idle hours into the four categories, and deadhead miles into planned and unplanned.

Cost the addressable portion at your own rates, then convert it to vehicle-days and compare that figure against your spot spend for the same weeks. In most produce operations the recovered vehicle-days are worth more than the operation expects, precisely because they are valued at peak spot rates rather than at annual average cost.

That comparison is the business case, and it is available from data you already hold before any platform decision is made.