How a Regional LTL Carrier Cut Empty Miles 34% and Turned $2.3M in Wasted Capacity Into Profit

In late 2024, a regional LTL carrier with 127 tractors and a service footprint across six states faced a problem that looked manageable on a spreadsheet but devastated cash flow in the field. 34% of their loaded miles were deadhead runs: tractors moving empty to the next pickup point. That meant every two of three dollars spent on fuel, maintenance, and driver wages generated zero revenue. The company was profitable on paper. On the dock, they were hemorrhaging.

The culprit was not technology failure. It was information friction. Dispatchers worked with fragmented load data. Shipper requests came through email, phone, and a aging TMS that did not talk to the carrier's own available capacity in real time. A dock manager in Charlotte might have a 14-pallet shipment for Greensboro that should have consolidated with a pickup 15 miles away. Nobody knew. The shipment waited. The tractor rolled empty. Revenue vanished.

Challenge

The carrier's own load board was a legacy system built in 2008. It displayed posted loads but did not optimize matches. A dispatcher searching for freight to fill a tractor heading to Memphis would scroll through hundreds of posted loads, many already assigned, many not on the route, many missing critical details like weight and pallet count. The process took 20 to 30 minutes per routing decision. For a dispatch team managing 90 active loads per day, that overhead alone consumed 30 hours of labor weekly.

Worse, the optimization was invisible to the driver. A driver would receive a pickup location and arrival time. The algorithm had not tested whether a slightly different route could consolidate three smaller loads into a single run or pair this tractor with a return load that reduced the deadhead distance to 40 miles instead of 280 miles.

The financial impact was brutal. At an operating cost of $1.85 per mile and an average deadhead haul of 160 miles, every empty run burned $296 in direct costs. Multiply that across 127 tractors over a year, accounting for the roughly 34% deadhead ratio: the carrier was spending approximately $6.8 million annually to move nothing. Margin on that ghost fleet: zero.

Solution

In September 2024, the carrier implemented a modern AI-driven load board platform designed specifically for internal optimization. The system pulled real-time data from the TMS, shipper requests, and GPS feeds from every tractor. It then ran continuous matching algorithms that tested thousands of routing combinations for each dispatch cycle.

The core logic was simple but effective: for every available tractor and every available load, calculate the cost of the deadhead required to make that match, then rank all possible assignments by the total cost per revenue mile generated. The system surfaced the top 10 options to the dispatcher in 90 seconds, ranked by profit per mile, not arbitrary criteria.

Equally important: the platform included a driver-facing app that showed the driver why they were being routed in a particular direction and what loads they were consolidating. Transparency built compliance. Drivers stopped calling dispatch to argue routing; they understood the trade-offs in real time.

Integration took 12 weeks. The first four weeks involved extracting clean data from the TMS and GPS systems. The vendor built the model against nine months of historical load and routing data to establish baseline performance. By week 12, dispatchers were live.

Results

After eight months of operation, the carrier's deadhead ratio had dropped to 22.3%, a reduction of 34% from the original 34%. That improvement translated directly to loaded miles: the system created an additional 78,400 loaded miles per year across the fleet.

At an average revenue rate of $2.15 per mile (the carrier's blended rate for regional LTL), those newly generated loaded miles produced $168,500 in incremental gross revenue annually. Subtract the cost of fuel and maintenance for those deadhead miles that were eliminated, and the net margin gain was $2.3 million for the first year.

Secondary benefits: dispatch labor fell 22% after the first six months as the algorithm removed the guesswork. Driver retention improved slightly; drivers reported higher satisfaction when routing logic was transparent. Fuel efficiency improved 4% across the fleet as the system optimized for fuel consumption in route matching, not just deadhead reduction.

The cost was $340,000 for software licensing and integration over three years. Payback occurred in six weeks.

The lesson is not that AI is magic. It is that freight optimization has been stuck in 2008, waiting for the math and the data infrastructure to catch up to what modern platforms can do. For a carrier running half-empty, that gap has been expensive.