North American gateways: LA/LB 8-day rail dwell, Savannah 6 vessels waiting

The Analysis

Zencargo's 8 September update shows a North American gateway system where the vessel queue has largely cleared but the inland leg has not. Los Angeles and Long Beach report just one vessel waiting, yet rail-terminal dwell is running at 8 days. Oakland sits at 6 days, Seattle-Tacoma at 5, New York/New Jersey at 4. Savannah carries 6 vessels waiting. New York chassis shortages are easing, and the Canadian rails remain the weak link, with Montreal, Prince Rupert and Saint John all dwelling above 7 days.

Read the shape of it. Congestion has moved off the water and onto the rail ramp, which means a berth is no longer the constraint. The rail leg behind the gate is, and that is exactly where delivery promises are breaking. An 8-day rail dwell at Los Angeles and Long Beach is not a port delay to wait out; it is five extra days against a normal 3-day dwell, and those five days land directly in the inland transit time you have already quoted.

Every group below works from the same figures: an 8-day rail dwell at LA/LB, 6 days at Oakland, 5 at Seattle-Tacoma, 4 at New York/New Jersey, 6 vessels waiting at Savannah, and Canadian rail dwells above 7 days. Assumptions are stated wherever a calculation is used.

For Exporters

• By 22 September, split the US book into IPI and local cargo and quote IPI on a stated 8-day LA/LB rail dwell
• Size the exposure: 40 IPI boxes at 4 days over a 4-day free-time allowance is $600 a container, $24,000 a month and $288,000 a year
• Put a rail-dwell trigger in every DDP and CIF offer, moving the delivery date day for day beyond 5 days of dwell
• By 30 September, move IPI volume that can take another gateway to New York/New Jersey or Seattle-Tacoma
• Assign one named person to track free time daily from discharge and flag containers at day three
• Never treat rail dwell as a port delay, and keep origin and shipping documents matched to the declared gateway

Your delivery promise into the US interior was built on a 3-day rail dwell that no longer exists. Los Angeles and Long Beach are running 8 days of rail-terminal dwell with only one vessel waiting at anchor, so the delay is not at the berth you are watching, it is on the ramp behind the gate. Oakland is at 6 days, Seattle-Tacoma at 5, New York/New Jersey at 4. On a DDP quote the container sits in the terminal while your clock runs, and demurrage starts on day five of a standard 4-day free-time allowance.

Assumptions: 60 FEU a month to the United States, 40 of them IPI cargo moving inland by rail and 20 local door cargo, a 4-day free-time allowance and a demurrage charge of $150 a container a day. An 8-day dwell uses four days beyond free time, so $600 a container, $24,000 a month across those 40 IPI boxes, and $288,000 a year. Against a normal 3-day dwell you are also carrying five extra days of inland transit, so a quoted 30-day door-to-door with a 5-day buffer now needs a 10-day buffer, or the claims start.

By 22 September, split your US book into IPI and local cargo and quote them on different lead times, with the IPI lead time carrying a stated 8-day rail dwell at LA/LB. Put a rail-dwell trigger in every DDP and CIF offer: beyond 5 days of dwell the delivery date moves day for day. By 30 September, move the IPI volume that can accept a different gateway onto one, and write the demurrage exposure by container into the booking sheet so the account team sees the number before the box sails. Assign one named person to track free time daily from discharge, not weekly.

Alternatives, priced. Routing IPI cargo through New York/New Jersey at a 4-day dwell saves 4 days against LA/LB and removes the $600 a container demurrage exposure, but the ocean leg and the drayage both change, so the rate has to be re-quoted rather than assumed. Seattle-Tacoma at 5 days saves 3. Air freight covers a broken promise, not a routing problem, and at current rates it does not pay on a full container. The traps: reading an 8-day rail dwell as a port delay, so the wrong team gets called; assuming free time starts at discharge when it starts at the gate; and re-routing without matching origin and shipping documents to the declared gateway, which turns a five-day problem into a customs problem.

For Cross-Border E-commerce

• By 25 September, add five days of rail dwell to the inbound lead time of every LA/LB inland SKU and raise hero SKU cover from 20 to 25 days
• Quantify the choice: five extra days at 2,000 units a day is 10,000 units at risk, $12,000 a cycle at a $6 margin, against $7,200 a year to hold the same stock
• By 10 October, move the fastest-moving SKUs to a shorter-dwell gateway and leave slow movers on the cheapest rate
• Reset reorder points on actual published dwell rather than a port-to-door average, reviewed weekly
• Cap the winter buy on any SKU below 15% contribution margin once five extra days of carrying cost is included
• Do not price or promise on a 3-day rail dwell while the ramp is running 8 days

An 8-day rail dwell at Los Angeles and Long Beach is a replenishment problem long before it is a freight cost problem. The vessel queue is down to one ship, so your container arrives on time and then sits on the ramp for 8 days against a 3-day norm. Oakland is at 6 days and Seattle-Tacoma at 5, and Savannah still has 6 vessels waiting on the berth. If you run a 20-day safety stock on a two-week restock cycle, five extra days eats a quarter of your cover, and the SKU that runs out is rarely the one you planned for.

Assumptions: a container of 3,000 units, monthly sales of 60,000 units so 20 containers a month, a normal 3-day rail dwell and a 20-day safety stock. Five extra days on a daily run rate of 2,000 units puts 10,000 units of demand at risk per cycle. At a $6 contribution margin a unit, losing 20% of that to stockouts costs $12,000 a cycle. Holding the same 10,000 units instead costs $120,000 at a $12 landed cost, about $7,200 a year at a 6% cost of capital. The buffer is the cheaper of the two by a wide margin.

By 25 September, add five days of rail dwell to the inbound lead time of every SKU that lands at LA/LB and moves inland, and raise safety stock from 20 days to 25 on hero SKUs only. By 10 October, move the fastest-moving SKUs onto a gateway with a shorter dwell and leave slow movers where the rate is cheapest. Set reorder points on actual dwell rather than a port-to-door average, and review them weekly against the published Zencargo figures. Cap the winter buy on any SKU whose contribution margin falls below 15% once the extra five days of carrying cost is in the number.

Alternatives. Rerouting to New York/New Jersey at a 4-day dwell recovers 4 days, but the inland leg and the rate both change, so run the landed cost again rather than assuming it is better. Splitting a buy across two gateways hedges the dwell risk and doubles the inbound flows to manage. Air freight recovers the five days at many times the ocean cost, and it only makes sense on a hero SKU with a stockout in front of it. The traps: reading an 8-day rail dwell as a port delay and missing the inland leg entirely; setting safety stock on last quarter's transit time; and letting the tail of the range carry the same cover as the top sellers.

For Manufacturing Plants

• By 22 September, put the published 8-day rail dwell into every imported part's transit time and reset reorder points from 3 days
• Raise critical-path safety stock from 7 to 12 days by 30 September, and leave non-critical parts unchanged
• Size the trade: five extra days at $80,000 a day holds $400,000 of pipeline stock, $24,000 a year at 6%, against $400,000 of lost output if a line stops
• Qualify an alternative gateway or transload-plus-truck leg by 30 September for every part where dwell exceeds 20% of inbound transit
• Pull long-lead spares and equipment parts to a 90-day horizon by 15 October
• Never run one safety stock across a 3-day and an 8-day route, and never treat IPI and local cargo as one flow

Your inbound rail leg has become the longest part of the journey. Los Angeles and Long Beach are running an 8-day rail-terminal dwell with a single vessel waiting, Oakland is at 6 days, Seattle-Tacoma at 5 and New York/New Jersey at 4. For a plant pulling materials on a four-week cycle, five days against a 3-day norm is not a rounding error: it is a change in the reorder point, and it is already sitting inside the standard cost and the schedule your planners are working to.

Assumptions: three inbound containers a week, 13 a month, an 8-day dwell against a 3-day norm, and daily material consumption of $80,000. Five extra days of transit holds $400,000 of extra material in the pipeline, which at a 6% cost of capital is $24,000 a year. That is cheap against the alternative: a line stopped for five days on $80,000 a day of consumption is $400,000 of lost output before overtime and expediting. Bringing safety stock on critical parts from 7 days to 12 days covers the dwell for a small fraction of that exposure.

By 22 September, put the actual rail dwell into the transit time of every imported part number and re-run reorder points at 8 days rather than 3. By 30 September, raise safety stock on critical-path parts from 7 days to 12 and leave non-critical parts alone, because blanket cover is cash you do not need to spend. Qualify one alternative gateway or a transload-plus-truck leg for every part where the dwell exceeds 20% of its total inbound transit. By 15 October, pull long-lead spares and equipment parts to a 90-day horizon, since a ramp delay on a spare is a line stop with a buying lead time behind it.

Alternatives, ranked. New York/New Jersey at a 4-day dwell removes 4 days and the demurrage exposure, at the cost of a different ocean rate and inland leg. A West Coast transload into a truck leg bypasses the rail ramp entirely but adds cost per container and does not scale. Domestic sourcing removes the rail leg and changes the piece price. The traps: running the same safety stock on a 3-day and an 8-day route; ordering on the planning system's nominal transit time while the ramp runs five days behind; and treating IPI and local cargo as one flow, which hides which containers are actually exposed.

For Brand Owners

• By 25 September, publish the delivery promise by gateway with the rail dwell behind it instead of one national date
• Quantify the miss: 15% of 120,000 monthly orders is 18,000 exposed, and a 10% miss rate is 1,800 late orders, $5,400 in credits and $9,000 of lost repeat
• By 10 October, move volumes that can switch onto New York/New Jersey or Seattle-Tacoma and shorten the promise by the days recovered
• Set written channel and region inventory priority so the tightest promise gets the first container off the ramp
• Put a dwell-based date-adjustment clause into every annual customer agreement with a trigger and 14 days' notice
• Do not market a 28-day promise while LA/LB rail dwell runs at 8 days

A delivery promise is the one thing a brand sells that it does not manufacture, and the North American ramps are currently rewriting it. Los Angeles and Long Beach show 8 days of rail dwell with a single vessel waiting, Oakland 6, Seattle-Tacoma 5, New York/New Jersey 4, and Savannah 6 vessels waiting. The vessel queue is no longer the story, which means the port-congestion messaging you built is pointing at the wrong part of the journey. What the customer experiences is five extra days against a normal 3-day ramp, and they experience it as a broken date.

Assumptions: 120,000 orders a month at a $25 gross profit an order, a 28-day delivery promise, 15% of orders exposed to the extra five days, a $3 credit per late order and a 20% churn rate among delayed customers. 18,000 orders are exposed, a 10% miss rate is 1,800 late orders a month, $5,400 in credits and $9,000 of lost repeat gross profit at $25 an order. That is $14,400 a month, $172,800 a year. Holding the same order book five days later in the pipeline is a working-capital question; a missed date is a repeat-purchase question, and the two are not the same size.

By 25 September, publish the delivery promise by gateway with the rail dwell behind it, instead of one national date. By 10 October, move the volume that can take a different gateway onto New York/New Jersey or Seattle-Tacoma and shorten the promise by the days you actually recover. Put the channel and region inventory priority in writing, so the buyer with the tightest promise gets the first container off the ramp. Add a dwell-based date-adjustment clause to every annual customer agreement, with a stated trigger and 14 days' notice, so a rail delay becomes a documented event rather than a credit conversation.

Alternatives and the trade-offs. A shorter promise on a shorter-dwell gateway is a real competitive advantage and it costs a re-rate plus an inland leg. Absorbing the five days upstream with more stock protects the date and ties up cash. Using air on the exposed orders keeps the promise at a cost that only pays on the highest-margin SKUs. The traps: marketing a 28-day promise while the ramp runs 8 days; giving the fastest promise to the channel with the slowest inventory; and letting the credit line absorb the cost invisibly, because $3 a late order looks small until it is 1,800 orders a month.

For Procurement Teams

• By 22 September, put a 5-day rail-dwell threshold and a demurrage-allocation clause into every ocean and intermodal contract
• Size it: 120 IPI boxes a year, 4 days over free time at $150 a container, is $600 a box and $72,000 a year
• By 30 September, split every tender into local-cargo and IPI lanes with a gateway-substitution clause covering 30% of volume on 14 days' notice
• By 15 October, qualify New York/New Jersey and Norfolk for Chicago-bound IPI volume, with Canadian rails above 7 days as the reason
• Price the alternative before choosing it: 4 days saved is about $26 a container in capital cost but $600 in demurrage
• Do not sign a contract that names a port and not a ramp, and name one owner for daily free-time tracking

Procurement's question is not whether the ramps are slow, it is which volume sits behind them and what the contract says when they are. Los Angeles and Long Beach are at 8 days of rail dwell with 1 vessel waiting, New York/New Jersey at 4, Seattle-Tacoma at 5, Oakland at 6, Savannah with 6 vessels at anchor, and the Canadian rails above 7 days. A contract that names a port but not a ramp, and sets no dwell threshold, leaves the demurrage and the delay with you by default.

Assumptions: 480 FEU a year into the United States, 120 of them IPI cargo through LA/LB, a 4-day free-time allowance and $150 a container a day of demurrage. An 8-day dwell is four days beyond free time, $600 a container, $72,000 a year on those 120 boxes. Now price the alternative: moving that volume to a gateway at a 4-day dwell removes the $600 entirely, while 4 days of transit on a $40,000 container is about $26 of capital cost a box, roughly $3,200 a year across 120. The demurrage, not the inland leg, is where the money is.

By 22 September, put a dwell threshold and a demurrage-allocation clause into every ocean and intermodal contract: beyond 5 days of rail dwell the carrier or ramp operator carries the charge, and delivery windows move day for day. By 30 September, split every tender into a local-cargo lane and an IPI lane with separate volumes and rates, and add a gateway-substitution clause allowing up to 30% of volume to move on 14 days' notice. By 15 October, qualify New York/New Jersey and Norfolk as the primary alternatives for Chicago-bound IPI volume, since the Canadian rails above 7 days are the weakest link in that corridor. Fix one owner for daily free-time tracking and put the exposure into the monthly procurement report.

Alternatives and their cost. A West Coast transload into a domestic truck leg bypasses the ramp but adds cost per container and depends on truck capacity that is tight. Routing through New York/New Jersey removes 4 days of dwell and the demurrage but changes the ocean rate, so price it rather than assume it. Norfolk is a genuine alternative for the Chicago corridor. The traps: assuming free time starts at discharge when it starts at the gate; signing a contract that names a port and no ramp and sets no threshold; and treating the tariff dwell figure as the inland transit when the ramp behind the gate runs longer.

↑ Back to the key points