Shipment Visibility Is Not Cargo Coverage

The most persuasive screen in logistics is the map. A blue dot leaves Ningbo, reaches Vancouver, waits, moves east, and finally turns into a delivered event. It gives a management team the feeling that the shipment has become knowable. It has become observable to us.

Not the same thing.

The container’s GPS pin tells us where the goods are. It does not tell us when the insurance promise began, ended, or changed hands.

A carrier feed tells us a box entered a terminal. It cannot establish whether buyer or seller bore risk there. Whether the certificate was on ICC (A), (B), or (C). Whether the goods had entered an elected distribution warehouse. Or whether the policy somebody assumes exists has a relevant exclusion. This is an authority-boundary problem, not a visibility problem.

Our logistics platform is authoritative for movement events.

The policy, certificate, sale contract, and transit clause are authoritative for coverage. Ask one system to answer the other’s question and an operationally sophisticated company gets surprised by an ordinary cargo loss. So the useful architecture is less fashionable than a control tower. We put the operational timeline beside a coverage timeline. We reconcile them at each material handoff.

Keep the evidence of what each event did to cover.

The interface can be very modern. The answer still comes from the contract.

A shipment has more than one clock running

The first clock is commercial, and we do not own it.

Incoterms 2020 allocate cost, risk, and specified obligations between buyer and seller. They don’t transfer title, and they don’t settle every question of insurable interest.

They decide something our map cannot see.

Who bears the loss the moment something goes wrong. CIF and CIP are the two rules requiring seller-procured insurance for the buyer’s benefit.

Their similar names hide a consequential difference.

CIF keeps Institute Cargo Clauses (C) as the minimum.

CIP requires ICC (A).

The conventional minimum sum insured is 110% of contract value, in the contract currency. The sale term makes that visible in the commercial record we hold.

The shipment record does not.

“The supplier arranged insurance” is therefore a poor control statement. A CIF certificate can meet the trade-term requirement and still provide named-perils cover. For a container of manufactured goods, that is a very different promise from the all-risks cover the importer assumed it bought. Under FOB, meanwhile, risk passes when goods are on board.

A damage event on the quay is the seller’s loss, even where the tracking history records the same vessel departure.

The second clock is ours.

The Institute Cargo Clauses are model wordings maintained by the Joint Cargo Committee of the Lloyd’s Market Association and the International Underwriting Association.

Policies and certificates incorporate them by reference.

They are not a general label for “marine insurance,” and both the 1982 and 2009 editions remain in circulation.

A record saying only “ICC (A)” left out the part that says what it means.

The third clock belongs to the carrier. Hague-Visby, CMR, COTIF-CIM, Montreal: not insurance rules.

They govern what a carrier owes, on what terms.

That matters most after payment, when we exercise subrogation and find a carrier defence, a package limit, or a time bar standing between us and a full recovery.

So one shipment moves smoothly through a carrier’s network while three legal clocks run.

Our record has to preserve the relationship. “In transit” correctly describes the least important fact.

A, B, and C are three different promises

The clause letter is not a product tier we chose for marketing.

It determines how a claim reaches us.

ICC (A) is all-risks, subject to its exclusions.

The assured establishes a fortuitous loss, and we bring the loss within an exclusion.

ICC (B) and (C) are named-perils forms.

The claimant has to show the loss came from a peril on the list.

ICC (C) is largely a major-casualty list.

Fire or explosion. Stranding, grounding, sinking, capsizing. Collision. Overturning or derailment. Discharge at a port of distress. Jettison. General average sacrifice.

ICC (B) adds earthquake, volcanic eruption, lightning, washing overboard, entry of sea, lake, or river water, and total loss of a package lost overboard during loading.

Which makes the ordinary container incident unusually clear.

Rain entering through a corroded roof seam is not sea, lake, or river water.

A loss from it has no listed peril under ICC (B), and ICC (C) is narrower again.

Under ICC (A) it may be covered, absent an exclusion.

Insufficiency or unsuitability of packing, for instance.

“Water damage” is a useful alert category.

It is not a conclusion we can act on.

War and strikes are excluded from all three clause sets, and bought back under separate Institute War and Strikes Clauses.

Delay is excluded even when an insured peril caused it.

Insolvency or financial default of the carrier. Inherent vice. Ordinary wear and tear. Packing issues within the stated conditions.

Coverage questions an accurate arrival estimate cannot answer.

I would stop the model right here.

AI can classify documents, compare the 1982 and 2009 wordings, and flag a shipment moving into a risk-sensitive state.

It should not infer that an event is covered because it resembles a covered event.

“The model thought rain was close enough to sea water” will not improve the claim file.

Warehouse to warehouse is an instruction, not a slogan

“Warehouse to warehouse” sounds like a comfort phrase.

It is a transit clause with one attachment rule and several terminations.

Under the 2009 wording, cover attaches when goods first move within the origin warehouse for immediate loading.

Earlier than 1982, which attached when goods left.

It continues through the ordinary course of transit.

Inland movement, port handling, sea carriage, discharge, onward.

It ends at the first relevant event, whichever we reach.

Delivery to the named final warehouse ends it.

So does delivery to another warehouse the assured elects to use for storage outside the ordinary course of transit, or for allocation or distribution.

For marine transit, the outer limit is sixty days after discharge at the final port.

The air equivalent is thirty.

Those facts deserve event codes in our system.

They should not become automatic coverage declarations.

A scan at a transload facility may be routine transit, an elected distribution stop, or a carrier disruption.

The contract of carriage may end short of destination.

Cover can continue for a limited period in that abnormal case, and prompt notice and additional premium matter.

A change of voyage has similar notice-dependent protection.

Each of those events needs a legible rule beside it.

A practical Canadian operating problem, this.

A container discharged in Vancouver can dwell while duty, inspection, rail capacity, or an inland move is resolved.

The carrier reports every day of that dwell to us.

It will not report that the sixty-day marine limit has become the coverage issue.

The federal Marine Insurance Act governs marine cargo.

A land-only inland-transit policy falls under provincial law.

A door-to-door movement isn’t a door-to-door legal route.

Carrier liability is the recovery, not the cover

When an uninsured buyer hears that the carrier caused the damage, the obvious thought is that the carrier will pay.

That thought begins an expensive education.

Under Hague-Visby, sea-carrier liability is limited per package or unit, or per kilogram of gross weight, whichever is higher.

The bill of lading’s treatment of the package matters as much as the apparent damage.

A container described as one package produces a different ceiling from a bill enumerating its cartons.

The rule governs the carrier’s obligation.

Not the cargo’s value.

The inland record sets its own limits.

CMR limits road per kilogram, COTIF-CIM limits rail.

In most Canadian provinces, domestic road carriage caps liability at a few dollars a kilogram unless a higher value was declared.

Montreal Convention cargo limits are per kilogram, and change on a five-yearly inflation review.

No sensible control treats those ceilings as a substitute for the insured value of the goods.

The clocks are short too.

Under the Hague-Visby regime applied here, apparent damage requires notice at or before removal into the consignee’s custody.

Non-apparent damage requires notice within three days.

Suit against the ocean carrier, or a written extension, comes within a year of delivery or expected delivery.

We pay the cargo claim, then subrogate.

The carrier’s event history, delivery receipt, notice, and survey evidence are valuable because they preserve that later route.

They do not create it for us.

General average makes the missing record visible

General average is where a tidy tracking story collapses.

A container can be intact, visible, and delayed, and still unavailable, because a casualty elsewhere on the ship turned delivery into a security exercise.

The York-Antwerp Rules apply where the bill incorporates them.

The Comité Maritime International promulgates them.

No legislature does.

The chosen edition matters.

The 1994, 2004, and 2016 editions remain in circulation, and they treat salvage and port-of-refuge matters differently.

The coverage record keeps the incorporated edition.

Not a generic web link labelled “general average.”

The rule shares an intentional and reasonable sacrifice or expenditure for the common safety among ship, cargo, and freight, in proportion to value.

Cargo actually burnt is particular average.

Cargo soaked by water used to fight the fire can become general average, because the water damage followed an intentional act for common safety.

Not an exotic distinction.

It decides which claim path opens.

The shipowner has a lien on cargo for its contribution, and can withhold delivery.

An insured consignee commonly supplies an average bond and an average guarantee from us.

An uninsured consignee posts a deposit the adjuster controls.

Salvage security gets demanded separately.

After a severe container ship fire in 2018, security demands ran to more than two-fifths of cargo value for salvage alone.

Those demands are deliberately conservative early estimates.

Not a tracking exception to close with an apology.

The average adjustment then takes time.

Two to five years to publication is unremarkable.

Logistics needs the invoice and valuation now.

The claims and insurance records need the policy response, guarantee, and security evidence now.

The final allocation arrives much later.

So the operational timeline can’t be the only one.

The system of record is a reconciliation

The answer isn’t a more panoramic carrier feed.

It is making the evidence beside it more exact.

Per shipment: the sale contract and Incoterms rule.

The open-cover declaration or shipment certificate. The Institute clause letter and edition. The insured value. The attachment and termination facts. The bill of lading. Delivery records. And the notices and survey evidence we need if something goes wrong.

Where a general-average event occurs, add the invoice, valuation, average bond, average guarantee, and salvage-security demand.

Each answers a question no GPS coordinate can.

The board framing is then straightforward.

Can we reconcile shipment visibility to the insurance timeline, the buyer-seller transfer of risk, and the evidence a cargo or general-average claim needs?

If we cannot, we have visibility of a shipment and no control of the promise attached to it.

Visibility follows the shipment.

Coverage follows the contract.