Trade Data vs. Maritime Intelligence: What Customs Agencies Actually Need

At a Glance

• Trade data was built to answer commercial questions about flows, volumes, and counterparties, not enforcement questions about concealment.
• The vessel a customs agency needs to stop is the one that declared inaccurately and transmitted dishonestly, which is precisely the vessel both trade data and AIS describe least well.
• Satellite imagery, RF, behavioral analytics, and ownership resolution are generated independently of what a vessel or operator chooses to declare.
• Container targeting is where the gap is sharpest, because a clean declaration says nothing about the three voyage legs that preceded it.
• The test for any enforcement data source is what it still shows when the vessel switches off.

Trade Data Was Built for Commercial Visibility

Cargo manifests, customs declarations, shipping records, and import and export filings exist to describe commerce. They do that well.

How much crude moved through Yanbu last month? Which companies are shipping dry bulk into Rotterdam? What do the dominant trade flows look like across the Red Sea corridor this quarter? These are real questions with real answers, and the platforms built to answer them serve commodity traders, banks, and supply chain managers properly.

Those are not the questions customs enforcement asks.

Enforcement Asks a Different Set of Questions

Which of the 15,000 containers arriving at this port today holds something that was not declared? Which vessel took a detour through a sanctioned port before arriving here? Which manifest is fraudulent? Which operator has a behavioral history that signals concealment?

Trade data does not answer any of these, and the reason is structural rather than a matter of coverage or refresh rate. Trade data is a record of what parties declared. The enforcement problem begins at the point where the declaration stops being true.

AIS does not close the gap either. A vessel running dark produces no AIS record to check. A vessel broadcasting manipulated positions produces a record that is worse than nothing, because it looks complete. In both cases, the vessel most worth examining is the one generating the least usable data.

That is the shape of the enforcement problem. The subject of the investigation controls the inputs.

Intelligence Starts Where Declarations Stop

All-source maritime intelligence fuses the layers that exist independently of what a vessel or operator chooses to declare.

EO and SAR satellite imagery sees a vessel whether or not it is transmitting. SAR works through cloud and darkness, so coverage does not depend on conditions. RF detection identifies a vessel’s emissions signature when AIS is off, because a ship running dark is still operating equipment that radiates.

Behavioral analytics built on years of vessel activity flag deviations no single manifest could reveal. A route that does not match the stated voyage. A loitering pattern in a corridor associated with transfer activity. A dark period that begins and ends in places that make the gap itself the signal.

Ownership and entity resolution surfaces the beneficial owner behind layers of flag hopping and shell structures, which is how an operator with a history of concealment appears as the same operator rather than as a new company with a clean record.

Fused, these produce an intelligence picture rather than a data feed. The distinction matters operationally. A feed tells an officer what was reported. A picture tells them what happened, with enough specificity to justify an inspection.

Container Targeting Shows the Gap Most Clearly

Container targeting is the sharpest illustration.

Trade data can tell you that a container is declared as machine parts from a legitimate exporter with a clean filing history. Every field checks out. On a declaration-based targeting model, it ranks low and passes.

Trade data can tell you that a container is declared as machine parts from a legitimate exporter with a clean filing history. Every field checks out. On a declaration-based targeting model, it ranks low and passes.

When UK authorities seized roughly $100 million of cocaine from a container ship at London Gateway, the public reporting named almost nothing. A container vessel, a Panama port call, a June arrival. Those three details were enough to run a search, returning three vessels that matched the profile during the relevant window. One carried a documented history, surfaced automatically through adverse media screening, of a 2016 case in which approximately 200 pounds of cocaine were found concealed in a container loaded at Santos.

The routing is the part a declaration cannot show. Panama in early June, across the Atlantic, Antwerp on June 21, Rotterdam on June 23, London Gateway on June 25, then onward to Germany. Each leg generated its own paperwork, and each set of paperwork was accurate. The shape of the voyage, and the vessel’s own history, are what carried the signal.

All-source intelligence can tell you that the vessel carrying it called at a sanctioned port three legs ago with AIS switched off, that the declared exporter shares ownership infrastructure with entities already under sanctions, and that the vessel’s behavioral pattern in the week before loading matches a known concealment sequence.

None of that is visible in the declaration. All of it is available before arrival.

That is the difference between a declaration and an intelligence product. A declaration describes an intention. An intelligence product describes a history. Customs officers allocating finite inspection capacity need the second one, because the first was written by the party they are assessing.

The timing compounds the advantage. Declarations arrive close to arrival, leaving a narrow window to act. Vessel behavior accumulates across the entire voyage, which means the risk picture can be built while the ship is still at sea.

What Do You See When the Vessel Switches Off?

That is the test, and it is worth applying to any source an agency relies on.

Trade data sees nothing. The declaration describes the leg that was filed, not the voyage that happened, and it arrives regardless of what the vessel did in between.

AIS sees nothing. That is the point of switching it off.

All-source intelligence still has a picture. Satellite detection places the vessel. RF confirms it. Behavioral models read the gap itself as an indicator rather than a blank. Ownership intelligence connects the operator to everything else it controls. With all-source operational intelligence, whatever an agency needs to complete the picture comes in alongside the rest, whether that is national holdings, partner reporting, industry data, or a source specific to the mission.

Concealment is not an edge case in customs enforcement. It is the case. A targeting model that performs well on cooperative vessels and loses sight of uncooperative ones is optimized against the wrong population.

The vessels that matter most are the ones that stopped reporting. What an agency can see during those windows determines what it can stop.

Written by Maya Romi, Windward Senior Content Manager. Reviewed by Inbar Tropen, Windward Product Marketing Manager.