Shipment visibility platforms: aggregating other people's data

Coverage differs by mode, and so does the meaning of a milestone

Ocean movements are tracked through carrier and terminal events plus vessel position, giving good visibility of the sea leg and much weaker visibility of the inland legs at either end. Air moves through consignment status messages tied to the transport document. Road relies on telematics from the vehicle, which requires the carrier's consent and a working integration per carrier. Because each mode has its own event vocabulary, an aggregated milestone such as departed means different things in different lanes, and any cross-mode report needs a translation layer that somebody must own.

The matching problem nobody demonstrates

An order references a purchase order line; a warehouse references a delivery; a carrier references a booking, a container, a consignment or a vehicle. Building a chain from the customer commitment to the moving asset requires all those identifiers to be captured and passed along, and gaps appear wherever a party re-references a shipment, splits a load or transships. The most common failure of a visibility programme is not missing carrier data but unmatched carrier data, and the honest measure to track from the first week is the share of shipments successfully linked end to end.

Predicted arrival is a model, not an observation

Estimated arrival combines observed position and events with historical performance for the lane, terminal and carrier. Prediction quality therefore depends on history volume, and predictions are weakest exactly where they matter most: unusual lanes, disrupted periods and the final leg. Buyers should ask how estimates are produced, how often they update, and whether accuracy is measured against actual arrival and published. Anyone using predictions to trigger labour plans or customer promises needs that accuracy figure to be a monitored operating metric rather than a sales claim.

Visibility shows problems; it does not resolve them

Knowing a container will be late earns nothing on its own. Value appears when the information reaches a decision, expedite the balance, re-plan the receiving labour, warn the customer, hold the promotion, and that requires the platform to be wired into the systems and people who can act. Programmes that stop at dashboards produce a well-informed organisation with the same service outcomes, which is why the mature pattern pairs visibility with defined exception ownership rather than treating the screen as the deliverable.

Data rights and the carrier relationship

Much of the data belongs to somebody else. Carriers grant access under agreements that may limit onward sharing, and telemetry from a haulier's vehicles is also information about that haulier's operation. Sorting out permissions is part of implementation, not a legal formality, and it explains why carrier onboarding stretches over months. Intergovernmental work on trade digitalisation has pushed towards standardised data sharing, but in practice each connection still carries its own commercial conversation.

Frequently asked questions

How is this different from tracking on a carrier's own site?

A carrier shows its own movements in its own format. An aggregator normalises many carriers into one vocabulary and links their events to your orders, which is the part that takes effort and the part that fails when identifiers do not line up.

Can a visibility platform replace the transport system?

No. It holds no plan, no rates and no settlement, and it cannot instruct a carrier. It observes execution, so it complements a planning and execution system rather than substituting for one.

What should be measured after go-live?

Match rate between shipments and tracked assets, event coverage by carrier and lane, and prediction accuracy against actual arrival. Those three explain almost every complaint that the platform is not working.