Mode selection on inbound lanes is usually made once, at sourcing, and then inherited for a decade. That was defensible when truckload capacity was abundant and rates moved gently. It is harder to defend in 2026, when higher insurance, maintenance and labour costs are anchoring a structural cost floor under trucking that sits well above fuel, and intermodal is gaining ground across a widening set of lanes specifically because cost stability and network efficiency now outweigh transit-time sensitivity for a lot of freight. But the argument that actually matters for a plant is not the rate. It is variance. A lane that averages three days and sometimes takes six will generate premium freight regardless of what the rate card says, and a lane that reliably takes five will not. Ask for the deployment brief to weight your own lanes on reliability rather than on average transit.
ENTERPRISE GUIDE · INBOUND MODE
Inbound Transport Mode Optimisation
Road, rail and sea trade-offs weighted on reliability rather than average transit — plus the lane characteristics that disqualify a mode outright, and how mode choice made early avoids premium freight later.
Forecast rate movement, 2026
Truckload contract
+8% to 15%
Intermodal
+3% to 5%
This does not automatically make rail cheaper. It makes it more predictable — and for a plant, predictable transport cost is frequently worth as much as low transport cost.
Weight Variance, Not Average Transit
Average transit time is the number on every mode comparison and the wrong one to optimise. A plant does not buy transit — it buys arrival inside a window, and what breaks the window is variability.
What the average hides
Two lanes with identical mean transit can produce completely different expedite bills. The one with a long tail generates shortages; the one with a tight distribution does not, even if it is slower on paper.
What planners should be weighing
Total transit time, terminal dwell and lane reliability together — not rate against transit alone. Terminal dwell in particular is invisible on a rate sheet and is where much of intermodal's variability actually sits.
Why variance drives cost
Every hour of unexplained variability is absorbed either as buffer stock or as premium freight. Both are paid for; only one of them appears in the transport budget.
The mitigation that changes the maths
Knowing the status and location of in-transit shipments materially mitigates the impact of longer transit times and travel time variability on rail. Visibility does not make a lane faster; it makes a slower lane usable.
Reliability is a live concern, not a theoretical one
Shippers have consistently raised rail service reliability as a factor in mode choice — at Surface Transportation Board hearings in 2022 and 2024, witnesses testified that deteriorating service reliability was one reason customers chose truck even where a rail or intermodal option existed. From September 2024 the STB began collecting weekly local service metrics from Class I carriers, which shippers can use where reliability falls below a threshold. Treat published reliability data as an input to the model rather than taking a carrier's word or a competitor's anecdote.
Mode Fit by Lane Characteristic
The comparison that matters is not mode against mode in the abstract — it is mode against the specific lane. Four characteristics decide most cases.
← Swipe to see all columns →
| Mode | Distance | Volume and cadence | Transit tolerance |
|---|---|---|---|
| Road, truckload | Any, and the only realistic option under long-haul thresholds | Handles variable and irregular flow | Tightest windows, first and last mile, exception management |
| Intermodal | Enough distance to absorb the additional handoffs | Consistent weekly volume | Flexibility of one or two additional days |
| Rail, carload | Corridors above roughly 750 miles | Heavy, high-volume — one railcar can carry the equivalent of three or four truckloads | Transit measured in days rather than hours |
| Sea | International and port-to-inland | Bulk and containerised manufactured goods | Non-urgent, with port congestion risk priced in |
A useful capacity reference when sizing a shift: one intermodal shipment is typically the same size as one truckload, while one railcar can move the same cargo as three or four truckloads depending on product dimensions and car size. That ratio decides whether carload is even a candidate — most inbound part flows do not generate railcar-sized quantities per destination.
Mode is a lane-level decision, and most plants make it once at sourcing.
The deployment brief covers lane-level mode scoring on distance, volume, cadence and measured reliability, and how in-transit visibility changes which lanes become viable.
What Disqualifies a Lane
Worth stating plainly, because the enthusiasm for modal shift tends to skip it. Intermodal is not a universal replacement for trucking and requires realistic transit expectations.
1Just-in-sequence freightSequenced parts feeding a frozen window cannot absorb an additional day of transit or a terminal dwell surprise. This is a hard disqualification rather than a preference.
2Highly variable flowsIntermodal is most effective on manufacturing replenishment lanes with consistent volumes. Irregular demand undermines the economics and the scheduling both.
3Genuinely time-sensitive shipmentsWhere the window is measured in hours, the additional handoffs are the whole problem. Intermodal is almost always slower door to door than a direct truck.
4Insufficient distance or volumeThe strongest candidates have enough of both to absorb the extra handoffs. Below that threshold the drayage on either end eats the line-haul saving.
The structural answer for a plant network
Let rail carry the predictable spine — core lanes, recurring port flows, campus-to-campus movements — while trucks handle first mile, final mile and exception management. That division frees scarce trucking capacity for the time-critical and specialised work where it is genuinely irreplaceable, rather than spending it on routine replenishment that could have moved on rail.
Scoring Reliability Into the Decision
Five inputs, weighted deliberately. The point is to make reliability a number in the model rather than a reservation someone voices in a meeting.
On-time arrival distributionNot the mean — the spread, and specifically the tail. Score the 90th percentile transit alongside the average, because the tail is what generates expedites.
Terminal dwellTime at ramps and ports, which never appears on a rate sheet and often accounts for most of the variability on an intermodal lane.
Published carrier service metricsWhere regulators collect and publish service data, use it. It is more objective than either the carrier's account or your own small sample.
Visibility available on the laneStatus and location in transit materially offsets longer and more variable transit. A lane with good visibility can carry a tighter buffer than its raw variance suggests.
Disruption exposureIntermodal infrastructure is vulnerable to weather events, accidents and labour action that can sharply degrade a mode's capacity. Score the exposure, and know the fallback.
Mode Choice as Premium Freight Avoidance
The cheapest expedite is the one that never becomes necessary, and mode decisions taken early are one of the few structural ways to prevent them.
Decided early
Mode selected before capacity becomes a problem, at lane design rather than at booking
Buffer sized against measured variance, so normal variation is absorbed
Contract structures aligned across legs, so the savings are not diluted
Truckload capacity preserved for genuinely time-critical work
Premium freight becomes an exception with a cause, rather than a monthly line item.
Decided late
Alternative modes considered only after failing to cover a truckload
Long-haul truckload becoming more expensive to protect as rates firm
Premium capacity spent on routine freight it cannot be justified for
Misaligned contract structures quietly diluting whatever savings remain
Expedites become the mechanism by which the mode decision gets made.
The Review Sequence
Five steps, and the first two require no negotiation with anyone.
1Measure actual transit distribution per laneFrom your own arrival data rather than the carrier's quoted transit. Mean plus 90th percentile, per lane, for a full seasonal cycle.
2Classify lanes by toleranceSequenced and time-critical, replenishment with consistent volume, and irregular or low-volume. Only the middle group is a modal shift candidate.
3Score the candidates on all four characteristicsDistance, volume and cadence, transit tolerance, and measured reliability. A lane failing any one of them is not a candidate regardless of the rate delta.
4Check the contract structures alignMisalignment between the line-haul and drayage arrangements can dilute intermodal savings if it is not managed deliberately.
5Pilot one lane and measure the same waySame distribution measurement, same seasonal window. Expand only where the pilot proves the variance assumption rather than the rate assumption.
Score your lanes on variance, then on rate
Fleet Rabbit captures actual arrival distribution per lane, tracks terminal dwell and window adherence, and holds mode, tolerance class and reliability score per lane — so the annual review starts from measured performance rather than from quoted transit.
Frequently Asked Questions
Is rail cheaper than truck?
Sometimes, and that is not the strongest argument for it. For corridors above roughly 750 miles rail typically offers lower cost per ton-mile and markedly better fuel efficiency than highway-only moves. But the more durable case in 2026 is predictability: truckload contract rates are forecast to rise materially more than intermodal pricing this year, which does not automatically make rail cheaper — it makes it more predictable. For a plant carrying buffer against transport variability, predictable cost is often worth as much as low cost.
Which inbound lanes should never shift to intermodal?
Just-in-sequence freight, highly variable flows, and genuinely time-sensitive shipments. Intermodal is not a universal replacement for trucking and requires realistic transit expectations — it is most effective on manufacturing replenishment lanes with consistent volumes. It is also almost always slower door to door than a direct truck, because of the drayage, the lift, the rail journey and the whole process in reverse. Lanes feeding a frozen window cannot absorb that.
How should we weight reliability against rate?
Make it a scored input rather than a caveat. Weigh total transit time, terminal dwell and lane reliability together, using your own measured arrival distribution — mean plus 90th percentile — rather than quoted transit. Where regulators publish carrier service metrics, use them: they are more objective than a carrier's account or your own small sample. Shipper concern about rail reliability is well documented, including testimony that deteriorating service was one reason customers chose truck even where rail was available.
Does visibility change which modes are viable?
Yes, and it is the most underrated lever in the decision. Knowing the status and location of in-transit shipments mitigates the impact of longer transit times and travel time variability on rail. It does not make the lane faster — it makes a slower lane usable, because a buffer sized against known position can be smaller than one sized against uncertainty. Lanes that fail a mode assessment on variance alone sometimes pass once visibility is added.
When should mode be decided?
At lane design, before capacity becomes a problem. Searching for an alternative mode only after failing to cover a truckload is already late in the conversation, and it is how routine freight ends up consuming premium capacity that is becoming harder to justify. Review mode on a schedule as well — rate relationships, service reliability and your own volumes all move, and a decision inherited from sourcing five years ago is running on obsolete inputs.
What quietly erodes intermodal savings?
Misalignment between contract structures across the legs, which can dilute the saving if it is not managed carefully. Terminal dwell is the other one — it sits outside the rate and inside your variance, so a lane that looked competitive on paper underperforms in practice. Measure both explicitly during a pilot rather than assuming the line-haul rate delta is the whole picture. Start a free trial to capture arrival distribution and dwell per lane.
Buy the Distribution, Not the Average
Reliability scored as a number rather than raised as a doubt, disqualifying characteristics applied before rates are compared, rail carrying the predictable spine while trucks keep the exceptions, and the decision made at lane design rather than at the point of failure.
Rate forecasts and service metrics change — confirm current figures before committing · Site-level configuration
August 20, 2026By Emily Davis
All Blogs
