Port congestion intelligence: a brief on how to monitor what matters

At Mahaya we track port congestion as a routine input to client conversations, and we have opinions about how to do it without drowning. The category is noisy. Half a dozen indices, more than that number of free dashboards, and at any given moment one of the major ports is either "the next Long Beach" in some commentator's view or "back to normal" in some carrier's. The brief below is the working operational frame we apply.

What congestion actually is

"Congestion" as a single word covers four operationally distinct conditions. Mid-market exporters routinely conflate them, which is the first source of bad decisions.

The first is berth wait: vessels at anchor, waiting to dock. This is the headline number, the picture most often used in trade press: ships visible on AIS, sitting outside the breakwater. It is genuinely important when it is severe, but it is also the lagging signal. By the time vessel anchoring is in the headlines, the cargo currently on those vessels is already late and the upstream booking flow has already corrected.

The second is yard density: how full the container yard is, relative to design capacity. A yard running above 85% density tends to lose efficiency on truck turn times and gate throughput; above 95% it becomes operationally fragile in ways that won't show up in the berth wait number until the situation is already bad. Yard density is the leading signal that berth wait will deteriorate in 7–18 days.

The third is gate throughput: how fast trucks are getting in and out, measured in turn time minutes. This is the variable that directly hits the inland leg and connects to the last-mile cost picture. A doubling of average truck turn time at a major port (we've seen 22 minutes go to 48 in bad weeks) translates fairly directly into a 9–14% rise in drayage spot rates within a week.

The fourth is rail availability. At ports with significant intermodal cargo, the availability of rail slots and rail-ready chassis is often the actual binding constraint, not the dock. This one is structurally invisible to anyone monitoring vessel signals only.

The first two are upstream signals for the second two. The second two are what actually costs the exporter money. The buyer who confuses them ends up scrambling on the wrong end of the chain.

The data sources, ranked

For mid-market operators we recommend a hierarchy roughly like this:

• Port authority own data. For the ports that publish it cleanly (Singapore's MPA, Rotterdam, Hamburg, Los Angeles' Port Optimizer, Antwerp-Bruges), this is the most reliable baseline, because it's the operator measuring its own facility. Not all ports publish equally; the gap between best-in-class and lower-tier disclosure is significant and worth knowing.
• Carrier omissions and blank-sailing announcements. When the alliances start cutting voyages, they have already seen something a week or two earlier than the public signals. We treat alliance schedule changes as a high-quality early indicator.
• Vessel AIS analysis from the major commercial providers. Most visibility tools have a port congestion module; the substance behind these is similar across vendors because the underlying AIS data is similar. Treat as a useful aggregator, not as a primary source.
• Trade press. Useful as confirmation that something has become commonly known. Not useful for early action, because by the time it's in the press the lane is already mispriced.

The buying frame to apply here is the one we wrote about in our visibility note: most of these signals are not the vendor's data, they're the underlying source data. The vendor adds value in normalization and alerting, not in measurement.

What's a signal, what's noise

A few patterns we've trained ourselves on:

A single bad week at a major port is noise. The base rate is one in eight or nine weeks at any given major hub will look anomalous for stochastic reasons. Reacting to a single week reads the situation wrong.

Two consecutive weeks of degradation in the same direction across berth wait AND yard density is the threshold at which we start talking to clients about rerouting. Anything less than that and we hold position.

A sudden alliance announcement of three or more blank sailings on a lane is, in our experience, a more reliable indicator that something structural has changed than any AIS-derived index. The carriers see the yard data before the public does.

Sustained weather events near a major hub (a typhoon track through the Pearl River Delta, a windstorm at Northern European hubs, a hurricane in the Gulf) produce 5–11 day disruption windows that are very predictable in shape. The mistake is overreacting. The cargo that was in the system 72 hours before the event is the cargo that's at risk. Cargo bookings made after the event lands are usually pricing in the disruption already.

Strikes, both threatened and actual, are the signal class we take most seriously, because the operational shape is hard to forecast and the resolution window is political rather than meteorological. We watch the named ports (U.S. East Coast under USMX/ILA, the major German ports, the French Mediterranean ports) and we engage with clients differently depending on whether the dispute is in negotiating phase or escalating phase.

Acting on the signal

The operational levers, in roughly the order we'd reach for them:

First, accept a longer transit on the existing routing. For a transient event (one to two weeks), absorbing a delay is almost always cheaper than rerouting.

Second, switch port of discharge within the same region. If Hamburg congests, Rotterdam may not, and the truck cost differential into the destination market may be acceptable. This decision is lane-specific and customer-specific. Some receiving warehouses will refuse a port switch contractually; others won't notice.

Third, accelerate inland on cargo already at the discharge port. If your container is one of the first off the vessel and the yard is filling, paying for an expedited drayage slot may be 2–4x normal but it gets the box out of the danger zone. The math here often pencils when the alternative is a 10-day yard sit. This is where the last-mile picture intersects with the port picture directly.

Fourth, reroute future bookings. This is the most expensive lever and the one to reach for last. The cost is not just the alternative freight rate; it's the relationship cost with the existing carrier and forwarder, who will remember being switched off in a bad week.

The single most useful framework we've seen at the regulatory level is the IMO's safety of navigation reporting, which captures incidents and approach conditions that often presage congestion in the surrounding port system. It's not a congestion index. But the operators who read it tend to know about the disruption a few days earlier than the operators who don't.

The summary, working: most port congestion is not your problem; the small share that is your problem is identifiable in advance with the right two or three signals; the rest of the dashboard is noise dressed as data.