Hapag-Lloyd Week 36: Charleston Waits to 18h, LA Pier 400 to 12h; Savannah Eases

The Analysis

North American terminal congestion in Week 36 is real, but it is local rather than systemic. Charleston's Wando Welch Terminal is the worst point on the board: non-Gemini vessels wait up to 18 hours, against 6 hours for Gemini services. Los Angeles APM Pier 400 runs to 12 hours, while Yusen and Fenix report no waiting at all. Savannah has eased to about 6 hours from 18 hours a week earlier, Houston's Barbours Cut sits at 6 hours and Norfolk at 2 to 4 hours. New York, Miami, Seattle and Tacoma show no notable vessel queues.

That pattern is the useful part. A network-wide crisis raises everyone's cost, so there is little to decide. A terminal-level problem is a routing decision, and routing is where money is won or lost this month. In the same week that Charleston is running 18-hour waits, a berth one sailing up the coast is running at 3.

The five groups below work from one shared set of numbers: 18 hours at Charleston, 12 at Los Angeles Pier 400, 6 at Savannah, 2 to 4 at Norfolk, and an assumed $180 per container per day of combined storage, demurrage and trucker waiting time. Each calculation states its own assumptions where it is used.

For Exporters

• By 25 September, map every open US East Coast booking to its discharge terminal and flag Charleston Wando Welch and LA Pier 400 by name
• Size the wait: 12 Charleston containers at $180 per container per day is $2,160 a month and about $26,000 a year for one day of terminal delay
• Move eight Charleston containers to Norfolk (2 to 4 hours) by 30 September and recover $1,440 a month without renegotiating any rate
• From 1 October, add 1.5 days of buffer to East Coast delivery commitments and put the appointment window, not a single ETA, in the contract
• Get the berth name on the booking confirmation and have the forwarder confirm in writing which terminal the vessel actually calls
• Do not quote one port-to-door transit for a gateway where two terminals in the same port differ by 12 hours

The delivery date you publish is not set at the port, it is set at the gate, and Charleston just moved it. On the Week 36 board, non-Gemini vessels calling Wando Welch wait up to 18 hours, which is 12 hours more than Savannah's 6 and roughly 15 more than Norfolk's 2 to 4, while Gemini services at the same terminal clear in 6. Los Angeles is not uniform either: APM Pier 400 runs to 12 hours, Yusen and Fenix report none. Sell CIF or DDP and that wait lands in your cost and your delivery promise. Sell FOB and the buyer pays the trucker, then sends you the claim on the next order.

Size the gap before anyone argues about rates. Assumptions: 40 containers a month into the US East Coast, 12 of them booked to Charleston, and $180 per container per day of combined storage, demurrage and trucker waiting. An 18-hour wait is 0.75 of a day, but demurrage and detention are billed in whole days, so count one: $180 a box, $2,160 a month, about $26,000 a year on those 12 containers. Move eight of them to Norfolk, where the board shows 2 to 4 hours, and the wait cost on those boxes drops to almost nothing, recovering $1,440 a month without renegotiating a single freight rate. Reading the board at terminal level rather than port level is what produces that number.

By 25 September, pull the Week 36 board and map every open booking to its discharge terminal rather than its port, flagging Charleston Wando Welch and LA Pier 400 by name. By 30 September, shift every Charleston booking that can accept Norfolk, Savannah or Houston before the documentation cut-off, and get the forwarder to confirm in writing which berth the vessel actually calls. From 1 October, add 1.5 days of buffer to East Coast delivery commitments and write the appointment window into the contract instead of a single ETA. Name one person to own the weekly terminal board and the short exception list of bookings that genuinely cannot move.

Alternatives, with their costs. Norfolk shows 2 to 4 hours, but inland drayage runs longer and rail options are thinner, so a 15-hour berth saving can be eaten by an extra day on the road. Savannah is back to 6 hours from 18, which makes it a genuine second option, though it recovered once and can slip again. Air freight is not a plan for 12 containers. The traps: reading a port-level average when the problem is one terminal; booking by port code and letting the carrier choose the berth; assuming the free-time clock starts at vessel arrival when terminals often start it at discharge; and quoting a 35-day door-to-door when your own last three boxes ran 36.5.

For Cross-Border E-commerce

• By 28 September, set lane-level cover: 24 days for Charleston and LA Pier 400 lanes, 21 days for Savannah, Houston and Norfolk
• Quantify the buffer: three extra days on a 60,000-unit SKU costs $72,000 of working capital and $4,320 a year, against $136,000 of revenue lost in two days out of stock
• By 5 October, move high-revenue-per-kilo SKUs off the constrained terminals first and book truck appointments at the same time as vessel bookings
• From 1 October, widen the published last-mile promise by two days on affected lanes
• Set an automatic air-freight trigger at three days of cover instead of leaving it to a judgement call in the warehouse
• Do not measure safety stock in calendar days when a berth wait and a rebooked appointment consume two of them

Inbound replenishment is priced on a smooth transit and scheduled against 21 days of safety stock, and Week 36 moved the second number. Charleston's 18-hour wait at Wando Welch, LA Pier 400's 12 hours and Savannah's improvement to 6 from 18 do not change your freight rate by a dollar. They change how many days of stock you must hold so a hero SKU does not go out of stock at the warehouse while its container is still at anchor. On a marketplace that promises same-week delivery, one missed replenishment cycle is a week of lost sales plus the cancellations and returns that follow it.

Put a number on the buffer. Assumptions: a hero SKU selling 60,000 units a month, a landed cost of $12 a unit, 21 days of safety stock today, and a 6% annual cost of capital. Adding three days, enough to cover an 18-hour berth wait plus a rebooked truck appointment, lifts cover from 42,000 to 48,000 units, so $504,000 becomes $576,000 tied up: $72,000 more working capital, or $4,320 a year in capital cost. Compare that with the downside. Two days out of stock on the same SKU is 4,000 units at a $34 retail price, or $136,000 of revenue, and most of it does not come back. Three days of buffer costs about 3% of the revenue you would lose in two days of stock-out.

By 28 September, split the inbound plan by gateway and give each lane its own cover: 24 days for anything discharging at Charleston or LA Pier 400, 21 days for Savannah, Houston and Norfolk. By 5 October, move the SKUs with the highest revenue per kilo and the tightest replenishment cycle off the constrained terminals first, and leave the slow movers where they are. Book the truck appointment at the same time as the vessel booking, not after the arrival notice lands. From 1 October, widen the last-mile promise shown on your site by two days for SKUs on the affected lanes, and set an automatic air-freight trigger at three days of cover rather than leaving it to a judgement call in the warehouse.

Alternatives and where they go wrong. Splitting one shipment across two gateways keeps a flow moving but doubles the customs entries and the drayage minimums. Holding more stock is what the arithmetic above supports. Air-freighting a hero SKU protects the listing but costs far more than three days of capital at this volume. The traps: measuring safety stock in calendar days while a berth wait and a rebooked appointment eat two of them; assuming the terminal's free-time clock matches your warehouse calendar; and letting a slow mover occupy buffer that a hero SKU needs, which turns a 12-hour berth delay into a stock-out on the one line that pays the bills.

For Manufacturing Plants

• By 30 September, add every imported part with a named discharge terminal to the MRP exception report and review it every Monday
• Raise cover on Charleston and LA Pier 400 parts from 7 to 9 days by 15 October, and qualify a second gateway per part before the peak build
• Size the trade: $50,000 of daily material consumption means two extra days of buffer is $100,000 of inventory and $6,000 a year, against $96,000 for one lost eight-hour shift
• By 31 October, move non-critical packaging and consumables to Norfolk or Houston and keep constrained terminals for critical-path parts
• Pull long-lead spares and control boards to a 90-day horizon as your own suppliers reset their lead times
• Give the scheduler an ETA range rather than a single date, and never treat the terminal board as a delivery commitment

Your line runs on material a planner scheduled weeks ago, and Week 36 moved the front of that schedule. Charleston's non-Gemini wait of up to 18 hours, Los Angeles Pier 400 at 12 hours and Savannah's recovery to 6 from 18 add up to this: a container your plant expects on a Tuesday can reach the gate on Wednesday afternoon. Lines rarely stop because freight got expensive. They stop because a resin, a liner or a control board is not on the floor when the shift starts, and that cost is measured in output, not in freight.

Assumptions: a plant consuming $50,000 of imported material a day, three inbound containers a week with one discharging at Charleston, a 7-day safety stock, and a second shift running at $12,000 an hour of throughput. An 18-hour berth wait plus a rebooked truck appointment pushes gate-in by about 1.5 days. A 7-day buffer absorbs that, but only if it is counted from gate-in rather than from vessel arrival. Take the buffer to 9 days and you carry $100,000 of extra inventory, $6,000 a year at a 6% cost of capital, against $96,000 for a single lost eight-hour shift. The buffer is cheap; the shift is not.

By 30 September, add every imported part with a named discharge terminal to the MRP exception report and review it every Monday. By 15 October, raise cover on parts discharging at Charleston or LA Pier 400 from 7 days to 9, and qualify a second gateway per part before the peak build. By 31 October, move the packaging and consumables that can tolerate a longer inland leg to Norfolk or Houston, keeping the constrained terminals for parts on the critical path. Pull long-lead spares and control boards to a 90-day horizon, because your own suppliers are running the same exercise. Give the plant scheduler an ETA range rather than a single date, so shifts are not planned against a number the terminal has not confirmed.

Alternatives, ranked by cost. A second gateway is the cheapest fix and the fastest, but inland drayage from Norfolk adds a day and rail schedules are thinner. Transloading on the West Coast trades a berth wait for a handover, a second set of documents and roughly two days of extra handling. Pulling orders forward protects the build schedule and ties up cash. A domestic or nearshore supplier for the critical few parts is the durable answer and takes a qualification cycle. The traps: reading port-level transit as terminal-level; assuming free time starts at vessel arrival when it often starts at discharge; and telling the planner an ETA is a commitment when the terminal board is telling you it is a range.

For Brand Owners

• By 30 September, publish lane-level delivery ranges instead of one national date, with a three-day range on Charleston and LA Pier 400 lanes
• Quantify the promise failure: 6,000 affected orders at a 12% cancellation rate is 720 lost orders ($64,800) plus 5,280 support contacts ($21,120), against a few hundred dollars to rebook containers
• By 10 October, set a written channel priority for constrained stock and tell every channel which SKUs it covers
• Require the 3PL to issue an appointment-level exception report within 24 hours of any terminal wait over 12 hours
• From 1 October, keep the promotional calendar clear of every SKU on an affected lane
• Notify affected customers before the tracking page does, and treat the 3PL's ETA as an estimate rather than the terminal's real time

A customer never sees a berth queue. They see the delivery date on your site, and Week 36 made that date less reliable: 18 hours of waiting at Charleston Wando Welch, 12 at Los Angeles Pier 400, 6 at Savannah, 2 to 4 at Norfolk. This is a promise-management problem before it is a freight problem. Publish five days and deliver 6.5, and the difference turns up as cancellations, support contacts and returns, none of which appear in a freight budget. The carriers describe the pattern as local, which means it is fixable by routing rather than by waiting for the market to clear.

Assumptions: 40,000 orders a month, 15% of them on SKUs inbound through an affected terminal, and a $90 average order value. That is 6,000 orders. A 1.5-day delay pushes them past the published promise; assume 12% cancel and the rest generate one support contact at $4. You lose 720 orders at $90, or $64,800 of revenue, and field 5,280 contacts at $4, or $21,120, for $85,920 in a single month. Rebooking eight containers from Charleston to Norfolk costs a few hundred dollars of extra drayage. The routing decision is two orders of magnitude cheaper than the promise failure.

By 30 September, publish delivery as a range by lane instead of one national date, and show the range for SKUs on Charleston and LA Pier 400 lanes as three days wide. By 10 October, set a written channel priority for constrained stock, owned e-commerce first, wholesale second, marketplace last, and tell every channel which SKUs it covers. Set a 3PL service level that requires an appointment-level exception report within 24 hours of a terminal wait exceeding 12 hours. From 1 October, hold the promotional calendar clear of any SKU on an affected lane, because a discount window that opens while a container sits at anchor converts a delay into a permanent margin give-away.

Alternatives. Promise a range rather than a date and manage the expectation honestly; the cost is a slightly softer conversion rate and the benefit is no compensation claims. Position inventory closer to the customer, which is what the buffer arithmetic in the supplier sections supports. Split a critical replenishment across two gateways so one flow keeps moving. Prioritise one channel openly rather than letting every channel compete for the same boxes. The traps: one national promise date; running a promotional event into a 1.5-day lane delay; treating the 3PL's stated ETA as the terminal's real ETA; and leaving the customer to discover the delay from a tracking page rather than a notice, which is what turns a late delivery into a refund.

For Procurement Teams

• By 30 September, add a terminal-level annex to every ocean contract: named berths, a waiting-hour threshold, and who pays demurrage and appointment-miss charges above 12 hours
• Size the two levers: one day of wait on 12 Charleston containers is $2,160 a month and about $26,000 a year; moving eight to Norfolk removes $1,440 a month before any rate talk
• Price spot cover honestly: 12 containers at $400 above contract is $4,800, under a fifth of the annual wait cost, and it buys a 2 to 4 hour gateway
• By 15 October, cap any single terminal at 60% of a coast's contracted volume and name the second gateway
• By 31 October, secure peak space on a 14-day booking window and define waiting hours as anchorage or berth, since the two differ by hours
• From 1 November, attach a demurrage tariff schedule so the step from day one to day two beyond free time is priced, not estimated

Terminal-level congestion is a contract problem before it is a rate problem. Week 36 shows Charleston's non-Gemini wait at up to 18 hours against 6 for Gemini services and 2 to 4 at Norfolk; Los Angeles Pier 400 at 12 hours against none at Yusen and Fenix; Savannah back to 6 hours from 18. None of that moves the rate you negotiated. All of it moves who pays the drayage, the demurrage and the missed appointment, unless the contract already answers the question. The carriers describe this as local, which means it is addressable by routing and by wording, not by waiting.

Assumptions: 40 containers a month into the US East Coast, 12 discharging at Charleston, and $180 per container per day of combined storage, demurrage and trucker waiting. One full day of terminal wait across those 12 boxes is $2,160 a month, about $26,000 a year. Now price the alternative. Twelve containers bought on spot on a strained lane at, say, $400 a container above your contract rate is $4,800, which is under a fifth of that annual wait cost and buys you a gateway with 2 to 4 hours of waiting. Move eight boxes to Norfolk and you take $1,440 a month out of the cost line before any rate conversation starts. The routing lever is worth more than the rate lever this month.

By 30 September, add a terminal-level service annex to every ocean contract: named berths where the carrier can commit to them, a waiting-hour threshold, and an explicit statement of who pays demurrage and appointment-miss charges once the wait exceeds 12 hours. By 15 October, cap any single terminal at 60% of a coast's contracted volume and put the remainder on a named second gateway. By 31 October, secure peak space with a 14-day booking window rather than the usual seven, and write the waiting-hour definition into the agreement as arrival at anchorage or at berth, because the two differ by hours. From 1 November, attach a demurrage tariff schedule showing the step from the first day beyond free time to the second, so the exposure is priced rather than estimated.

Alternatives and the trade-offs. Paying a premium for a service that clears in 6 hours instead of 18 is a real option on one lane, but it pays only where volumes justify a dedicated allocation. Shifting contracted volume to Norfolk or Houston works until everyone does it, so build the second gateway early. Transloading on the West Coast trades a berth wait for handling, documents and two days. The traps: demurrage tariff steps, where the jump from day one to day two beyond free time usually costs more than the wait itself; free time measured in calendar days while appointments are only available on working days; storage and demurrage clocking separately, so both can run at once; and a contract that names the port but not the terminal, which leaves you paying for a berth you would never have chosen.

↑ Back to the key points