For years the rule was simple: if the West is jammed, send it East. On the planning side, the coast was never the variable. One terminal was.
Published June 20, 2026 · The Loaded Box · Status Operator-facing · A companion read to Vizion's East-vs-West port-congestion analysis
1. Opening
Layer 1 · Headline · Opening Frame
The instinct, when one coast slows down, is to reroute to the other. It held for a long time. Vizion's recent East-vs-West analysis showed why even the physical version of that rule has quietly frayed: the two coasts no longer lose time the same way. The West loses it at the berth, getting vessels worked. The East loses it in the yard, in container dwell. Kyle Henderson put it well - the map in our heads updates a lot slower than the ports actually do.
That work measures the import cycle: the box after it lands. There is an earlier question every exporter answers first, weeks before any of that. When you commit drayage, rail, and gate appointments against a published receiving date, can you trust that date to still be there when the booking rolls?
We looked at that question across roughly 2,800 US export port-calls on all three coasts, from March to May 2026 - the same window VIZION studied - through the cargo receiving window (the terminal's Earliest Return Date) and the CY cutoff.
The coast does not answer it. The East coast holds both the steadiest and the least steady receiving windows in the country.
Share of port-calls where the receiving date moved late, March to May 2026. Ten of the eleven ports we measured, across all three coasts, sit inside one narrow band. One terminal sits far outside it.
Two things are true at once. As a whole, the East coast looks shakier than the West - 25.6% of East port-calls saw a late receiving-date move, against 11.2% in the West and 12.1% in the Gulf. And yet almost every individual port - East, West, and Gulf - behaves the same. The contradiction is the story.
2. The Mechanism
Layer 2 · The terminal is the unit
The coast-level number is real, but it points at the wrong object.
Take one terminal - Savannah - out of the East coast, and the East (13.3%) becomes indistinguishable from the West (11.2%). A single terminal accounts for the entire coast-to-coast gap. The "East coast" did not produce the pattern. One place inside it did, and it is large enough to color the whole coast.
Late receiving-date moves by coast. Remove a single terminal and the two coasts converge on the same number.
The mechanism is straightforward once the coast frame is dropped. Receiving-window stability is set at the terminal, not the coast. The receiving date is the terminal's instrument - it decides when its yard opens to take export cargo. Some terminals set that date and hold it. Some keep moving it through the weeks when exporters are already committing against it. That choice does not travel by coastline.
The East coast does not have a receiving-window problem. One terminal does - and it is large enough to make the whole coast look like it does.
3. Bridge to Evidence
Layer 2 to Layer 3 · From claim to patterns
If the terminal is the unit, three patterns must appear in the data:
-
One terminal carries the coast's signal, while the rest cluster together.
-
That cluster crosses coasts - most East ports and most West ports look alike.
-
Where the instability lives - in the terminal's date or the carrier's cutoff - varies terminal by terminal, not coast by coast.
Here is what March to May showed.
4. Quantified Proof
Layer 3 · Three findings that anchor the mechanism
One terminal moved its receiving date late on more than half of its port-calls. Savannah sat at 56%. The next highest, Norfolk, was 20%; no other port reached even that. Savannah is not the leader of a coast-wide trend. It is an outlier the coast average is built around.
Take that terminal out and the coasts are even. East without Savannah lands at 13.3%, against the West's 11.2% and the Gulf's 12.1% (Houston). Ten of the eleven ports - Los Angeles, Long Beach, Oakland, Seattle, Tacoma, New York, Newark, Charleston, Norfolk, and Houston - sit between 5% and 20%. Coast explains almost nothing once the outlier is set aside.
The two coasts fail differently in shape, not just in rate. When an East receiving date moves late, it usually moves by about a day. When a West receiving date moves late, it moves rarely - but the worst one in ten runs past a week. East exporters get nibbled. West exporters get hit hard and seldom.
Two coasts, two failure modes. The East moves often but small; the West moves rarely but, in the tail, by more than a week.
Methodology note. This edition reads the cargo receiving window and CY cutoff for container export voyages, anchored to those gates rather than to vessel sailing. The classification methodology is maintained internally and held consistent across publications.
5. The Illustrative Case
Layer 3 · Where the instability lives
Not every late move is the same kind of late. The receiving date is the terminal's to set. The CY cutoff is the carrier's. Place every terminal on those two axes and the picture sorts itself - not by coast, but by who is moving the date.
Where each terminal's instability lives. Far right means the terminal is moving the receiving date. Higher up means the carrier is moving the cutoff. The colors - East and West - do not cluster. The behavior does.
Savannah sits alone on the right: its receiving date moves a great deal, while its cutoff barely does. This is a terminal-side pattern. Norfolk, Oakland, and Tacoma sit higher and to the left - more of their motion is carrier-side, in the cutoff. And Charleston sits low on both axes. Why these terminals behave so differently comes down to one thing the data lets us see directly - when, if ever, each terminal lets the receiving date stop moving.
6. The lock, in the data
Where the receiving date is allowed to settle
You do not need a terminal to tell you its date is locked. Track the receiving date closely enough and the data shows you when it stops moving - and the pattern lines up with each terminal's own policy.
How often the receiving date moves at all, by terminal. The shaded cluster at the bottom - Charleston, New York, Newark - rarely moves. Savannah moves on nearly nine port-calls in ten.
Charleston publishes a policy: the receiving date is locked two days before the window opens, and will not change. The data agrees. 83% of Charleston receiving dates never move at all. The policy is visible in the numbers.
Savannah publishes a lock too, but a different kind. Its own vessel schedule marks each date locked or almost-locked - and on the live schedule, every vessel whose receiving date is still in the future reads almost-locked. The date does not lock until the window has already opened. So the lock confirms the date only once it is too late to plan against it - which is exactly what the movement shows: Savannah's receiving date keeps moving, on 88% of port-calls, right up to the day it opens. (Norfolk runs a lock as well, but on the vessel's arrival and departure, not the receiving date - a secured sailing is not a secured gate.)
You do not need a terminal's lock flag to know whether a date is firm. Track it closely and the date tells you itself. Charleston's holds. Savannah's does not settle until the window has opened.
The reputation an exporter routes on - "the East is calmer," "the West is the risky coast" - is not what the receiving windows say. The risk did not pick a coast. It picked a few terminals - and in each terminal's own data you can see, before you commit, whether its date can be trusted.
7. Decision and Cost
Layer 4 · The decision shift
The unit of the routing decision is wrong.
Routing by coast reputation answers a question the data no longer supports. The planning question is one level down: which terminal, and does its receiving window hold there?
The decision moves from the coast to the terminal.
The cost lands either way, just in different shapes. An exporter who sends cargo East expecting calm inherits, at the wrong terminal, a receiving date that keeps nudging - a replan, then another, until the commit deadline arrives against a date that is already stale. An exporter who stays West trades that for a rare event: a receiving window that holds for weeks and then jumps more than a week, with a truck already dispatched and a rail slot already booked. The first is a slow tax. The second is a single large bill. Neither shows up as a line item. Both come out of the same operation.
The shift is not "watch more." It is to commit against the terminal-day, and to set one recheck before dispatch where the window is known to move:
- Verified - the receiving window at this terminal has held. Commit normally.
- Watching - this terminal moves its date. Recheck before you dispatch.
- Unstable, use buffer - this terminal moves late and large. Build time and flexibility into the plan before the booking rolls.
That is a terminal-level posture, not a coast-level one. It is the same discipline whether the cargo moves East, West, or Gulf.
8. System Statement and Forward Hook
Layer 4 to Close
The coast was a usable proxy once. On the planning side, it has stopped sorting the risk. The receiving window an exporter commits against is set by the terminal, and the terminals do not agree with each other - not within a coast, and not in a way the coastline predicts.
If this holds, the routing instinct has to change shape. The question is no longer East or West. It is which terminal, and whether its receiving window can be trusted on the day you commit.
The picture now spans all three coasts, and the conclusion holds on every one: the receiving window is a terminal's to keep or to break. One thread stays open - whether the concentration holds. Is Savannah a permanent outlier, or does the pattern move to another terminal next quarter? A future edition will test it, terminal by terminal - the only unit that has held up.
Kyle Henderson, thank you for the prompt - this is the planning-side companion to the cycle you measured. The map updates slower than the ports. On this layer, it turns out the map was never drawn at the right scale.
TRADELANES · THE LOADED BOX · JUNE 2026
Leave a Comment