Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- By Friday, pull 90 days of lanes and flag every one at 700 miles or more carrying 20-plus FEU a month. Those are the only real candidates; everything else stays on the road.
- Before comparing rates, get the destination ramp's appointment window, free time in hours, and chassis pool size in writing, and build three days of dwell into the rail number.
- Run a two-week pilot on 10% of one lane and judge it on door-to-door days and accessorials per box, not on the quote.
- Contract drayage separately: a named carrier, a committed number of weekly pulls, and a service penalty. The rail rate stops at the ramp gate, and the thirty miles decide the outcome.
- Set the exit trigger now. If ramp dwell passes 48 hours for two straight weeks, move that lane back to the road within seven days and review at the 30-day mark.
With truck capacity tightening and diesel costs volatile, North American intermodal conversion accelerated in July: CSX reported intermodal volume up 9%, Norfolk Southern intermodal revenue up 22% on 5% higher volume, and Union Pacific domestic intermodal up about 19%. The economics favor lanes beyond roughly 700 miles.
The Analysis
The number worth staring at in the July intermodal reports is not the 9% or the 19%. It is 22% against 5%. Norfolk Southern moved 5% more intermodal volume and took 22% more intermodal revenue doing it. Divide one by the other and revenue per unit is up about 16%. Somebody is paying for that, and it is not the railroad.
Truck capacity tightened, diesel kept moving, and shippers did what shippers always do when the road gets expensive. They called the ramp. CSX intermodal volume up 9%, Union Pacific domestic intermodal up about 19%. The reports put the crossover at roughly 700 miles. That is a fine number for a rate table. It is not the number that decides whether your box gets on the train.
What decides it is thirty miles of drayage, a chassis, and a gate appointment. I have watched a shipper win the rate argument and lose the whole month to a ramp that only takes empties until noon.
Start with what a conversion physically is. A truck moves your box door to door, one move, one bill. Rail breaks that into four steps: a truck pulls it to the origin ramp, a crane lifts it onto a well car, a crane lifts it off at the destination ramp, and another truck pulls it to the door. The line-haul gets cheaper per mile. The handling gets more expensive per box, and it is fixed, because it does not care whether you are going 400 miles or 1,400. That is the entire reason the 700-mile line exists. Below it, two lifts and two drayage legs eat the line-haul saving. Above it, they do not. It is arithmetic, not policy.
The chain that produced July's numbers runs one way only. Road capacity tightens first, and it tightens where you can least see it. Not in the rate, in the availability. A shipper who cannot get a truck on Thursday at any price stops arguing about Thursday and starts asking about the ramp. Volume moves. Rail takes it, reports it, and then the constraint relocates rather than disappears. A 40-box-a-week importer that converts still needs 40 drayage pulls at each end, and the drayage fleet in that metro did not grow because the railroad had a good quarter. The queue moves from the highway to the gate, and the gate has fewer open hours in a day than the highway does.
Who feels it first depends on size, and the middle gets squeezed hardest. A big importer running a hundred boxes a week already has rail in the mix, a drayage contract, and somebody whose job is appointments. It picks up a modest unit-cost gain and a scheduling problem it can absorb. A shipper at ten to forty boxes a month sits in the worst seat in the house: too big for the spot van market to be kind, too small to hold a ramp slot or a chassis pool, and nobody on staff whose full-time job is the gate. Under five boxes a month, none of this changes and it should not, because the two lifts will eat you alive on a short lane. Staying on the road is the honest answer there.
Direction matters as much as size. Importers running into a fixed inland DC on a steady weekly flow convert cleanly, because rail rewards repetition and nothing else. Exporters with seasonal peaks, odd dimensions, or anything that needs a flatbed are in a different business. A ramp clears a narrower envelope than a highway does, and a load a step-deck would take without a conversation can get turned away at the gate for the way it sits in the well. That rejection costs a day and a rebooking, and it lands at the worst possible hour.
The timing is not what the rate sheet suggests. You will not see this as a price first. You will see it as appointment availability, then as dwell, then as a missed pickup, and weeks later as a rate. In my experience that sequence runs four to six weeks behind the volume move, and it bites hardest in the six to eight weeks before your retail customers start asking for firm delivery dates, which is exactly when you have the least room to absorb a day.
The headlines leave out the half that matters. That 22% against 5% is not only rail taking share. It is rail pricing the share it takes. Revenue per unit up roughly 16% means the saving from converting is being split at the ramp, not handed to you. Your model at the 700-mile line shows a gross saving. What you keep is the gross saving minus ramp storage, minus chassis per diem, minus reconsignment, minus the day you lose when the gate turns you around. The headlines also miss where the ceiling sits. A railroad can add a train faster than a city can add drayage drivers. So conversion volume does not run into a line-haul limit. It runs into a thirty-mile truck limit, and that limit is priced by the hour, not by the mile.
Run a small version of it. Say I move 40 FEU a month on a 900-mile lane, and the all-in road rate and the all-in rail rate differ by 400 dollars a box. Both of those are my placeholders, not figures from the report, chosen so the shape is visible. Gross, that is 16,000 a month. Now put ten of those boxes on the destination ramp for two days at 75 dollars a day of chassis per diem, and reconsign four at 150 dollars a lift. That is 1,500 plus 600, so 2,100 off the top, before anybody misses a pickup. Push dwell to four days on half the boxes and the per diem alone takes 3,000. The trade is still positive. It is just not the trade in the quote, and the difference lives entirely in the thirty miles at each end.
The mistake I see most is treating a ramp like a dock. It is not a warehouse. Free time is measured in hours, the appointment is the product, and nobody at the gate is moved by the news that your container is late because a customer changed a PO. The second mistake is assuming chassis are available because they were available last year. Pools are sized on last year's volume, and last year's volume was on the road. The third is mixing equipment. A 53-foot domestic container and an international 40 are different services on different networks, and the 19% at Union Pacific is domestic intermodal, which may have nothing to do with the box you are trying to move.
Ask what happens if you do not convert. Nothing dramatic. You stay on the road, you pay the tight-market rate, and you keep the flexibility rail takes away, which on a lane where your customer changes orders weekly is worth more than the spread. That is a legitimate answer. The illegitimate one is converting on the quote and discovering the accessorials in month two. So before you sign anything, call the destination ramp and ask what the appointment window looks like for the week you plan to start, not for this week. If they cannot tell you, you already have your answer.
↑ Back to the key points