Your purchase order carries a must-arrive-by date of Thursday. The receiver's scheduling portal has nothing open before Monday at 06:00, and the freight is ready to leave your Columbus plant tomorrow morning.
Whichever slot your coordinator takes, the cost is already chosen. Book Monday and the order misses its MABD. Send the truck toward Thursday without a confirmed window and the detention line arrives about three weeks later, by which point nobody can reconstruct why the driver was sitting there.
Detention is usually decided before the truck moves
Detention gets handled as an execution problem. The driver sat too long, the dock ran behind, the lumper crew went home at four. By the time the charge reaches accounts payable, the only question anyone asks is whether the carrier billed it to the contract.
Much of it was settled earlier than that. A load with sixteen hours of transit, tendered against a window eleven hours out, will wait somewhere. No amount of effort at the gate changes that arithmetic, because the appointment was not achievable when it was accepted.
The American Transportation Research Institute's detention study, released on September 10, 2024 using 2023 data, defines detention as dwell beyond two hours. It found that 33.8 percent of stops on contract freight crossed that line. Contract freight is the planned kind: known lanes, committed volume, standing appointments. A third of those stops still ran long.
The cost does not stop at the carrier's invoice. The US DOT Office of Inspector General's 2018 report on driver detention estimated that each 15 minute increase in average dwell time raises the expected crash rate by 6.2 percent, and put collective driver earnings losses between $1.1 billion and $1.3 billion a year. Your scheduling decisions land on somebody's working day.
Six facts decide whether a window is real
A portal publishes availability. It does not publish feasibility, and those are different questions. Whether a given slot works for a given load turns on six inputs, and the portal holds one of them.
- The must-arrive-by date. The slot has to satisfy the order, not merely exist. A window two days past MABD is not an option, it is a decision to take the penalty.
- Ready time. When the freight is actually picked, staged and loaded, which is rarely the moment the ERP released the order.
- Transit. Drive time for the lane under current conditions, the driver's remaining hours, and the break the clock will force.
- Facility hours and dock capacity. Receiving hours, shift changes, how many doors take this commodity, and whether the site runs a thin crew on Mondays.
- Carrier lead time. How long this carrier needs between tender and pickup to put the right equipment under the load.
- Neighbouring arrivals. What is already booked on either side of the slot at the same door. A clean 09:00 sitting behind two overrunning 07:00 appointments is not clean.
Only the first of those lives in the purchase order, and only the fourth lives in the portal. The rest sit in the TMS, the warehouse system, the carrier's dispatch, and last quarter's gate records.
Availability is the input, not the answer
| Question | Booking against the portal | Computing a feasible window |
|---|---|---|
| Which slots are open? | The list the facility published | The same list, treated as the starting set |
| Can the freight physically be there? | Assumed | Ready time plus transit plus hours of service, tested against the slot |
| Does the slot satisfy the order? | Not checked | Slot tested against MABD and the receiver's early-delivery rule |
| Will the truck wait after it arrives? | Unknown | Door-level dwell history and what is booked either side |
| What if nothing fits? | Take the next available | Rank the bad options by what each one costs |
| When does the failure surface? | On the day, at the gate | At booking, while choices remain |
Nothing in the middle column is wrong. It is just thin. Confirming that a slot is free is clerical work, and your coordinator is doing it in place of the planning work that would have told them whether the slot was any good.
You are choosing between two different invoices
This gets booked badly because the two failure modes are paid by different people out of different budgets, and nobody holds both numbers at the same moment.
Miss the window and the penalty is a compliance charge. Talk Business & Politics reported in September 2020 on a Walmart supplier memo setting on-time and in-full at 98 percent effective September 15, 2020, with a fine of 3 percent of the cost of goods for suppliers below it. The requirement covered suppliers moving their own freight and those using Walmart's logistics, across all categories. That same report put supplier performance at roughly 70 percent when the memo went out.
Hit the window by parking a truck overnight and the penalty is detention: billed by the hour once free time runs out, disputed six weeks later, and absorbed by transportation rather than by the commercial team whose order forced the timing.
Three percent of cost of goods on a truckload of ambient grocery is not the same order of magnitude as four hours of waiting. Sometimes taking the detention deliberately is correct. That is a defensible call when somebody made it against both numbers, and it is not what happens when a coordinator grabs the last slot in a portal at 16:40 on a Friday.
What Chase does before the load is tendered
Chase runs execution: tendering through delivery, exception detection and root cause. Booking is the earliest point where execution is still cheap to change.
Chase computes the feasible window set rather than reading the portal and picking from it. Ready time from the warehouse, transit for the lane, the driver's clock, the receiver's hours, the carrier's lead time and the arrivals already stacked at that door all go in. What comes out is an ordered list with the cost of each option attached.
- A feasible slot inside MABD exists. Chase books it and tenders to the window rather than to the date, which are not the same instruction.
- Nothing inside MABD is feasible. Chase says so at booking, names the binding constraint, and puts the trade in front of a person: take the compliance penalty, expedite, or ask the receiver for an exception while there is still time to grant one.
- The slot is feasible but the door is a known offender. Chase books it and records the expected dwell against the shipment, so the accessorial arrives explained instead of contested.
That third case closes the loop. When the invoice lands, Penny already holds the appointment, the gate-in and the dwell Chase expected. Deciding whether the charge was earned becomes a comparison rather than an investigation, which is the same argument as catching the cause while the tractor is still in the yard.
Booking fast is not the same as booking well
This layer is automating quickly. C.H. Robinson's newsroom describes an appointments agent that sets over 3,000 appointments a day across 43,000 locations in under a minute, and the company said in April 2025 that the previous version had already set more than a million before it was replaced. Those counts are the company's own.
Throughput like that is worth having, and it answers a different question than the one that produces detention. An agent that books the first available slot in sixty seconds books the wrong slot in sixty seconds. The constraint was never how long it takes to fill in the form.
What to do next
Pull last quarter's detention invoices and join each one to the appointment it was billed against. Add two columns: how many hours before the window the booking was made, and whether that window was the first feasible option or the only one left.
The result will not spread evenly. You will find a short list of receivers, a short list of origin plants with unreliable ready times, and a set of orders where the MABD never permitted a clean window at all. Those are three separate fixes, and only one of them involves the carrier.
Book a walkthrough to see Chase compute feasible windows on your own lanes and name the appointments that were never achievable.