Skip to content
Hafenlogik

Cases

Five engagements, anonymised on purpose.

Our clients' throughput is commercially sensitive and usually contractually confidential. So we publish volume as bands, never name a client, and state only figures we measured ourselves, before and after.

How to read these

Every volume is a band, not an exact number. Every result is a measurement we took on the client's own systems, in the same way before and after, and re-measured at the 90-day review. Where a result depends on something outside our work — a volume drop, a new crane, a mild winter — we say so in the case. We will not name a client, confirm a client or deny a client.

Schedule reliability: planning around the deviation instead of arguing about it

The brief

The operator's berth planning team and its largest forwarding customer had spent a year disagreeing about arrival punctuality. Both had data. The data did not match, because one side measured against the published schedule and the other against the berth window agreed the previous evening.

What we found

Across a 12-week sample, 61% of calls arrived inside the agreed berth window. The mean deviation was 94 minutes, but the distribution was the point: two-thirds of the lateness came from one in eight calls. Those calls were predictable by 18:00 the previous day in almost every case, and nobody was publishing that prediction to the gate or the yard.

What changed

We defined one shared arrival measure — forecast ETA against the agreed berth window, in minutes, with a single owner — and built the reporting view that made it visible to berth planning, the yard, the gate team and the customer at the same time. The gate slot release was then moved to follow the forecast rather than the published schedule.

Measured result

Window adherence rose from 61% to 78% over two quarters, mostly because the forecast was acted on rather than because ships got faster. Truck turnaround time on the affected services fell by an average of 14 minutes. Demurrage disputes with that customer fell from eleven in a quarter to two.

Caveat

The last quarter of the measurement period had roughly 6% lower volume than the first. We estimate that accounts for two to three points of the adherence improvement, and said so in the report.

„The argument stopped the week we agreed which clock we were reading.“
Head of berth planning, container terminal operator

The queue at the gate that was actually a queue in the yard

The brief

Hauliers were queuing up to 70 minutes at peak on Tuesdays and Thursdays. The terminal had costed a fifth gate lane at a seven-figure capital sum and wanted a second opinion before committing.

What we found

The gate lanes were not the constraint. Median processing time in the lane was 3 minutes 40 and stable. The delay was inside: the reshuffle rate in three yard blocks had risen to 1.7 moves per productive move, because export containers with long dwell times had been allocated into the blocks nearest the gate. Every truck was waiting for a straddle carrier that was digging.

What changed

A revised block allocation moved long-dwell export boxes away from the gate-adjacent blocks, a dwell-based restack rule was introduced for the night shift, and the truck appointment system's peak-hour quota was cut by 12% on the two peak days.

Measured result

Peak-hour truck turnaround time fell by 21 minutes. The reshuffle rate in the three blocks dropped to 1.2. The fifth gate lane was not built.

Caveat

The night-shift restack rule costs additional straddle carrier hours. Net of that cost the annual saving was still clearly positive, but it is a real cost and it is in the report.

„We had approved the capital. The audit cost us less than the design study for the lane we did not build.“
Chief operating officer, container terminal operator

Eleven per cent rework in the declaration process

The brief

The forwarder's customs team was working steady overtime and shipments were being released late. The team believed the customs office was the bottleneck. The customs office, unsurprisingly, believed the data was.

What we found

11% of declarations required at least one correction after submission. Three error types accounted for two-thirds of them: incomplete consignee data taken from an outdated master record, incorrect gross mass where the VGM was available but not linked, and classification defaulting to a previous shipment's code on repeat customers. None of these was a customs office problem. Each correction cost an average of 47 minutes of clerk time and pushed release into the next working day about a third of the time.

What changed

Master data ownership was assigned to one named person with a monthly cleansing routine. The VGM field was linked to the declaration at booking rather than retyped. The classification default was removed and replaced with a mandatory confirmation on repeat consignments. Cut-off times were published and enforced.

Measured result

Rework fell from 11% to 4% within one quarter and held at 4% at the 90-day review. Overtime in the customs team fell by roughly a third. Same-day release rose from 72% to 89%.

Caveat

Removing the classification default increased average keying time per declaration by about 40 seconds. Everyone agreed that was a good trade; it is still a cost.

„We were sure it was the authority. It was our own address book.“
Customs manager, freight forwarder

Two temperature excursions, one handover

The brief

Two claims in one year, both on fruit, both investigated internally without reaching a cause. The provider wanted a risk map before its insurer's renewal review.

What we found

Both excursions happened in the same 40-minute window: the gap between a unit being unplugged in the yard and being plugged in at the cold store, with a road leg in between and no genset on the tractor unit for that particular service. Plug occupancy in the yard also ran above 95% for six hours during the Monday arrival peak, which meant units arriving in that window waited unplugged, and nothing in the record showed how long.

What changed

Gensets were made mandatory for that road leg. A plug capacity plan was built for the arrival peak and 14 additional plug positions were commissioned. The handover protocol was rewritten so that unplug and plug times are recorded as events, not inferred from the data logger afterwards.

Measured result

Excursions at handover fell by 63% over the following year. More usefully, every remaining excursion had an attributable cause within a day, because the handover times now exist as records.

Caveat

One year of post-measurement is a short window for an event that occurred twice a year. The improvement in traceability is the more solid result; the excursion count is suggestive rather than proven.

„The risk map told us where to look. The event log told us what happened.“
Quality manager, cold-chain logistics provider

Thirty-four minutes per train, hidden in the shunting

The brief

Shuttle trains between a Hamburg terminal and two inland hubs were consistently departing late. The terminal and the rail operator each had a spreadsheet showing it was the other party's fault. Both spreadsheets were correct about their own part of the process.

What we found

Neither party measured the shunting movement between the reception siding and the loading track, because it belonged to a third party under a separate contract. It averaged 34 minutes and varied by 25 minutes depending on which reception road the train arrived on — which was itself decided by whoever answered the phone first.

What changed

The shunting movement was made a measured, owned step with a target time. Reception road allocation was fixed by service rather than by call order. The rail-mounted gantry's working block was aligned with the yard plan so the loading track was served without waiting for a straddle carrier to clear.

Measured result

Train handling time fell by an average of 34 minutes per call. Departure punctuality on the two shuttle services rose from 58% to 81%. Wagon buffer at the terminal was reduced by eight wagons.

Caveat

This result required both parties and the shunting contractor to sign a joint operating protocol. Without that, the measurement alone would have changed nothing.

„The lost half hour belonged to nobody, so nobody had ever counted it.“
Operations director, rail operator

Your problem is probably in one of these five shapes.

If it is, we can usually tell you which one in the free call — and roughly what finding the answer would cost.

Book a 45-min scoping call