A tractor goes down at hour four of a 900-mile lane. Within minutes there is a driver on the shoulder, a diagnosis to chase and a decision about who gets dispatched. Within a day there is an invoice. That invoice is the number the business will remember, and it is the wrong one.

Ask a fleet what a breakdown costs and you get the repair total. Ask the shipper waiting on that load and you get an entirely different figure. The gap between the two is one of the most persistent accounting failures in freight, and it is why unplanned downtime keeps getting solved in the wrong department.

Spending on roadside assistance for semi trucks is visible, budgeted and argued over line by line. The cost of the load that never made its appointment sits nowhere at all.

The number everyone measures

Maintenance is tracked as a cost per mile, and the trend is not encouraging. According to ATRI’s Analysis of the Operational Costs of Trucking, the industry-average cost of operating a truck reached $2.336 per mile in 2025, the highest figure in the report’s history. Repair and maintenance was among the fastest-rising line items, up 8.6 percent year over year to roughly 21.5 cents per mile.

That is useful for budgeting. It says almost nothing about what happens when a truck actually stops.

The number nobody owns

Go back to the tractor on the shoulder. What follows is not a maintenance event. It is a network event.

The load misses its delivery appointment. The receiver reassigns the door and the next available slot is 24 to 48 hours out. Detention and redelivery charges attach. On-time performance takes a hit that will surface in the next scorecard review. A recovery tractor gets pulled off a scheduled load, creating a second disruption to cover the first. The driver’s hours-of-service clock burns down in a truck stop parking lot, which means the reload is late even after the repair is finished.

None of that reaches the repair invoice. All of it reaches the supply chain.

This is the same blind spot distribution operations spent the last decade correcting on the technology side. The sector has already worked through what IT downtime really costs a distribution operation, and the reframe changed the investment case entirely. Once teams stopped measuring the cost of fixing the server and started measuring the cost of the outage, redundancy suddenly looked cheap. Vehicle maintenance has not had that moment.

Why the data never crosses the boundary

Part of the problem is architectural. Telematics platforms are excellent detection layers. A fault code fires, a dashboard lights up, an alert reaches the maintenance inbox. What happens next is almost entirely analogue: phone calls, a shop selected on proximity rather than capability, and an estimate nobody 800 miles away can validate.

That sequence generates no structured data. The failure is logged as a repair line item, the delay is logged by planning as a late load, and the two records never meet. Nobody can see that a maintenance decision made 18 months ago is why a lane keeps missing its window.

The scale is easy to underestimate. ATRI has tracked average mileage between unscheduled repairs at just over 38,000 miles, against annual utilisation approaching 86,000 miles per truck. For a 100-unit fleet that is roughly 200 unplanned interruptions a year, almost none captured in a form the planning function can use.

Moving the metric

Three changes matter more than any software purchase.

Measure downtime in service-level terms rather than repair cost. Hours out of service, loads affected, appointments missed and recovery miles driven are the metrics that connect a wrench to a customer.

Treat breakdown response as a coordination capability rather than a transaction. Whether an event resolves in four hours or two days is decided almost entirely in the first thirty minutes: accurate triage, a provider with the right capability for the failure, and pricing agreed before work begins. Fleets running that through a coordinated network consistently compress the window compared with those calling whoever answers first at 2am.

Feed repair events back into planning. Recurring failures by unit, system and lane are forecasting inputs. Treated that way, maintenance history becomes a reliability model instead of a cost archive.

The garage will always be where the wrench turns. But the decision about what a failure costs, and what it is worth spending to prevent it, belongs upstream with the people accountable for service levels. For as long as downtime is priced as a repair, it will be underfunded, and freight will keep arriving late for reasons the maintenance budget never explains.