It’s not glamorous work. But it’s where a lot of avoidable delays, angry customer emails and late invoices come from, so it deserves a proper plan.

Here’s how to think about it, from the data formats you’ll run into to the questions worth asking a provider before anyone writes code.

Why Shipment Data Gets Stuck

A single international shipment can involve a forwarder, one or more carriers, a customs broker, a warehouse and a trucking company. Each one produces status updates on its own schedule and, often, in its own format.

So the typical setup looks like this: an operations coordinator checks a tracking portal, copies the ETA into a spreadsheet, and pings sales when something slips. It works at low volume.

As shipment volumes grow, manual tracking becomes harder to maintain. Teams may spend more time reconciling updates, checking conflicting information and following up on missing statuses. People stop checking, or they check different sources and get different answers.

The fix isn’t a fancier spreadsheet. It’s deciding which system is the source of truth for shipment status and then feeding it automatically.

Three Ways Logistics Data Reaches Your Systems

Logistics data can reach business systems through several methods, depending on the provider’s technology and the services involved. Knowing which one you’re dealing with shapes the whole build.

Customer Portals

This is the default. You log into a web portal, search by booking or container number, and see milestones, documents and sometimes alerts. Portals are fine for people. They’re less useful for software, unless the portal offers exports, scheduled reports or email notifications you can parse.

Parsing emails is fragile, though. Anyone who’s maintained a scraper knows how that ends. A scheduled CSV export is a more reasonable stopgap while a proper feed gets set up.

EDI

Electronic data interchange is the older, heavier option, and it’s still common in freight and trade. Formats such as ANSI X12 and UN/EDIFACT define structured messages for things like shipment status, bookings and invoices. EDI tends to be dependable once it’s running, but onboarding a new trading partner can take weeks of mapping and testing.

If your provider mainly works in EDI, budget for that mapping time. Don’t assume a JSON feed is hiding behind it.

APIs and Webhooks

Some providers expose REST APIs for tracking, documents or rates, and a few push events through webhooks. APIs and webhooks can fit well into modern software architectures, although their usefulness depends on the provider’s documentation, data quality and integration requirements. Integration details also vary between providers. One provider’s “departed” event might be another’s “loaded on vessel,” and the field names may differ.

Ask for documentation, a sandbox and sample payloads before you commit to a design. A short call with the provider’s tech contact can save days.

Model the Events Before You Write Code

The most useful thing a tech team can do early is agree on an internal event model. That means a short list of shipment milestones your business cares about, what each one means, and which source system is allowed to set it.

You don’t have to invent the format from scratch. GS1, the supply chain standards body, maintains EPCIS, which it describes as its flagship data sharing standard for enabling visibility within organizations and across trading partners. Version 2.0 added JSON and JSON-LD syntax plus a REST API for capturing and querying event data, which GS1 pitches as developer-friendly.

Even if you never adopt EPCIS formally, a similar shape works well internally. For each event, record what happened, when, where and the business reason behind it. A container reaching a port, a customs release and a pallet received at a warehouse all fit the same pattern.

A few habits tend to save pain later:

  • Store the raw payload from every source next to your normalized event, so you can reprocess when a mapping turns out to be wrong.
  • Keep provider timestamps and your own received-at timestamps separate. Time zones in freight data can be a headache.
  • Treat ETAs as a history, not a single field. The way an ETA drifts over a week tells you more than the latest value.
  • Label every event with its source, so nobody has to guess whether a status came from the carrier, the broker or a person.

Plan for Messy Data from Day One

Freight data is rarely clean. The same shipment might show up under a booking number, a bill of lading number, a container number and your own purchase order number, depending on who’s sending the update. Your integration needs a way to link them, and that usually means a reference table you maintain yourself.

Duplicates happen too. Two sources may report the same vessel departure a few hours apart with slightly different timestamps, and a carrier may resend an event after correcting it. Decide early whether the newest message wins, whether a specific source wins, or whether a person reviews conflicts.

Then there are units and codes. Weights in kilograms or pounds, ports written as names or as UN/LOCODE codes, dates with or without time zones. None of this is hard on its own, but it adds up, and normalizing at intake is far cheaper than cleaning up reports later.

Customs Data Is Its Own Stream

Customs often gets bolted on last, and that’s a mistake. For U.S. imports, the filings go through CBP’s Automated Commercial Environment, which the agency calls the system through which the trade community reports imports and exports. CBP says it also uses ACE to determine whether goods are admissible.

Plenty of importers never touch ACE themselves. A licensed customs broker often files on their behalf, and the status comes back through the broker or forwarder. Your software needs to distinguish between filing status, CBP cargo release, and any outstanding holds or requirements from partner government agencies.

Customs holds can create particularly difficult visibility problems. A shipment can sit at the port, technically arrived, while a hold from another government agency keeps it there. If your dashboard only shows “arrived,” your sales team will promise dates the warehouse can’t meet.

What to Ask a Provider Before You Integrate

This is where the build either goes smoothly or stalls. Before you pick a provider, or before you start connecting to the one you already use, get clear answers on a handful of points:

  • Which delivery methods do you support: portal exports, EDI, API, webhooks?
  • Which milestones do you publish, and how do you define each one?
  • How fresh is the data? Is it pushed as events happen, or polled on a schedule?
  • Can we pull documents, such as commercial invoices and release notices, programmatically?
  • Is customs status visible in the same place as freight status, or kept separately?

It also helps to look at how a provider has packaged its own visibility tools, because that shows where the data already lives. Some providers of worldwide logistics services combine freight operations with customer-facing visibility platforms. One example is the Worldview platform, which has separate views for international freight, warehousing, trucking and customs. Its customs view also provides CBP release status and notifications about releases and holds. This gives businesses an example of how different logistics data streams can be organized within a customer-facing platform.

Build the Dashboard People Will Open

Plenty of logistics dashboards get built, demoed once and ignored. The ones that stick tend to be built around exceptions rather than lists.

A sales rep doesn’t need to see 300 shipments that are on time. They need the six that slipped, why they slipped, and how the ETA changed. Finance wants landed-cost data and invoice status. Operations wants holds, missing documents and anything sitting at a port longer than expected.

So build views by role, and keep them small. Exception-based views also make it easier to wire up alerts in Slack or Teams later, since you’ve already defined what counts as a problem.

And put the shipment reference everywhere. The order system, the warehouse system and the freight data should share an ID, or someone will spend half their week matching records by hand.

Security and Data Ownership

Shipment data includes supplier names, product descriptions, declared values and sometimes personal details for consignees. Treat API keys and EDI credentials like any other production secret: scope them tightly and rotate them. If an outside developer or agency builds the integration, keep the credentials in your own accounts rather than theirs.

Clarify ownership, too. Ask what happens to your shipment history if you change providers, whether you can export it in bulk, and how long the provider keeps documents. Those questions rarely come up in a sales call, and they matter a lot two years later.

Start with One Lane

You don’t need to connect every carrier and every country at once. Pick a single trade lane or one high-volume supplier, connect it end to end, and see where the data breaks. Then expand.

The goal isn’t necessarily perfect real-time visibility across every shipment. It’s fewer surprises, a team that trusts the dates on its screen, and less copy-pasting between tabs. For most businesses, that’s enough to justify the build.