A shipment was due in Rotterdam on a Tuesday afternoon. The carrier’s system said so, with a precise time: 14:00. The forwarder booked the truck for 13:30, reserved the warehouse slot, and released the container for pickup. The ship arrived on Thursday. The ETA was never updated and never corrected. According to the carrier, it was never wrong. It was simply the wrong number to plan on.

Carrier ETAs are promises. Predictive arrival times are models. The gap between the two is the only number a supply chain team can actually plan around.

Why the Carrier Number Is a Promise

For decades, ocean freight planning ran on a single figure: the shipping line’s estimated time of arrival. It looks precise, a date and a time stamped by the carrier’s own system. But that number is built from schedules and voyage plans, not from what the vessel is actually doing. It does not know about the queue at the anchorage, the berth that freed up late, the storm that slowed the crossing, or the terminal that quietly changed its working hours.

Carriers manage ETAs as commercial instruments as much as operational ones. A schedule that slips quietly is preferable to one that announces itself, and schedule recovery is a constant, invisible game of shifting arrival times. The result is a number that is always present and rarely accurate. Studies of ocean reliability routinely show that a large share of vessels arrive outside their published windows, and the ones that land inside them often do so only on paper.

What the PTA Adds

Predictive time of arrival (PTA) signals work differently. They combine live AIS positions, historical terminal performance, weather and tide data, and machine learning models trained on years of actual port behavior. The output is not one point either. It is a range, a probability distribution that narrows as the vessel approaches the coast. A PTA for a ship still two weeks out carries wide uncertainty; the same model a day out can be startlingly precise.

The smartest teams stopped asking which number is right. They watch the gap between the carrier ETA and the PTA. When the gap is small and stable, planning proceeds normally. When it starts to widen, that is the earliest warning the operation has, often days before the carrier formally changes anything.

The gap changes decisions. A one-day gap means the truck does not idle at the gate; it can be rebooked. A two-day gap means the warehouse slot can be given to another shipment. A widening gap in a congested port means labor and yard space should be reassigned before the delay hardens into a detention invoice. None of this requires predicting the future. It requires watching the difference between two honest estimates instead of trusting one flattering number.

The Price of Planning on a Promise

Consider the planner who books everything on the carrier ETA. She releases the container, schedules the drayage, and promises her customer a delivery date. The ship slips. The truck shows up anyway, because nobody told it not to. The detention clock starts ticking at the terminal. The warehouse holds a slot for cargo that is not there, then pays overtime to process it when it finally lands. Every one of those costs is an ETA tax: the price of trusting a promise instead of a signal.

The Team That Switched

One freight team made the change after a single painful quarter. They had paid out six figures in detention and demurrage charges, most of it on shipments that arrived late but had been planned as if they were on time. They started logging the ETA-PTA gap daily on their top lanes. Within two months, they had rebooked more than a third of their drayage moves before the trucks left the yard. Detention spend fell by nearly half. The carriers did not change. The ships did not speed up. The team simply stopped planning on a promise.

Watch the Gap

The era of the single arrival number is ending, not because ETAs will disappear, but because they will finally have a rival. The gap between the carrier’s promise and the predictive signal is free information, already sitting inside the data your visibility tools collect. The question is whether your team looks at it.

Ask one thing tomorrow morning: what is the gap between ETA and PTA on your most important lane right now? If nobody can answer, you are still planning on a promise. And a promise, as that Rotterdam shipment proved, costs more than a model ever will.