Skip to content
    Supply Chain Technology

    How to tell a quiet shipment from a broken data feed

    Oct 7, 20266 min read

    A shipment with no news and a dead data feed look identical on screen. How to tell a quiet shipment from a broken data feed, and catch the silent ones early.

    Sixty containers out of Shanghai and Ningbo, and not one has produced a tracking event since 30 September. Your planner raises it on Tuesday morning. Golden Week ran from 1 to 7 October, so the obvious reading is that nothing happened because nothing was open: factories closed, documentation desks empty, feeds with nothing to publish. That reading is correct for fifty-eight of those containers. For the other two, the feed itself stopped nine days ago, and nothing in your system said so.

    This is the hardest week of the year to read silence correctly. Chinese factories restart from 8 October and the backlog clears into the third week of the month. Carriers have announced blank sailings through week 42, pulling capacity out just as the rush arrives. Every one of those events creates a legitimate gap in your data, which is precisely what makes the illegitimate gaps invisible.

    The plumbing is also less dependable than most teams assume. The Port of Los Angeles blocked more than 120 million cyberattack attempts in August, up from roughly 40 million a month in 2022. Almost none of that reaches an individual shipper, but it is a useful reminder of what your visibility actually rests on. Carrier APIs, port community systems and forwarder EDI connections are ordinary software. They have outages, credentials expire, and schemas change without notice. When one stops working it does not raise a flag. It just stops saying anything.

    Silence is not a status

    Every operational screen you use displays last known state. The container is "departed Ningbo" because that was the last event anyone published, and the screen will keep saying "departed Ningbo" forever, whether the box is mid-ocean, sitting in a yard, or invisible because the connection broke. The record looks calm. Calm is not the same as current.

    This is why dashboards mislead so reliably. A missing value does not render as a gap, it renders as the previous value, and most supply chain dashboards flatter the business for exactly this reason. The failure mode is quiet by construction.

    Compare it with a bad ETA. A wrong arrival date at least announces itself: the number moves, somebody notices, and the work of improving ETA accuracy across carriers starts with a signal you can actually see. Silence gives you nothing to react to. It is the reason 62% of supply chain exceptions are still detected only after they have reached the customer. The exception was there the whole time. Nothing surfaced it.

    Monitor the source, not just the shipment

    The fix is a change of subject. Stop asking only "what is happening to this shipment" and start asking "when did I last hear from this source."

    Every feed has a rhythm. An ocean carrier milestone API publishes something every six to twelve hours on an active booking. A terminal publishes at gate moves and not between them. A forwarder EDI connection lands as a nightly batch at a predictable hour. Those rhythms are measurable, and once you have measured them, a source that has gone quiet for three times its normal interval is a fact you can act on rather than a feeling somebody has on a Tuesday.

    Most teams cannot run that check, because the evidence lives in twelve different portals with twelve different notions of a timestamp. That is the gap a normalized data layer closes. Orkestra connects carriers alongside ERP, TMS and WMS systems through more than 200 pre-built integrations and normalizes what comes back, so every source lands in one model with one comparable last-updated time. The question becomes answerable in a single place: which sources have gone dark, and which shipments are sitting behind them.

    Give every feed a heartbeat

    Set an expected update interval per source and per lane, then alert on the absence rather than the event. Three details decide whether this works in practice.

    Baseline by lane, not globally. A transpacific booking and a European road movement have different natural cadences, and a single global threshold will either flood you or catch nothing.

    Load the holiday calendars. Golden Week, Tet and Chinese New Year are expected-silence windows. A heartbeat monitor that does not know about them will cry wolf for a week every year, and a monitor that cries wolf gets muted, which leaves you worse off than before.

    Separate the feed from the booking. When two sources cover the same shipment, one can fail while the other keeps reporting, which looks like agreement until you check. Deciding how to unify carrier and freight forwarder tracking into one record is what makes a half-silent shipment visible at all.

    Keep the history. Once every gap is logged, feed reliability stops being an impression and becomes a number you can put in front of a carrier or a forwarder at the next review: this source met its publishing interval 94% of the time last quarter, that one 61%. Partners will argue with a complaint. They rarely argue with their own uptime.

    Three kinds of silence, three different owners

    Once you can see silence, sort it. There are only three kinds, and conflating them is what keeps teams busy without making them faster.

    Nothing happened. The box is genuinely sitting still, the holiday is real, the vessel is mid-ocean. No action beyond confirming the next expected milestone.

    Something happened and nobody published it. The container was rolled at a transshipment hub, or it is accruing dwell at origin while your ERP shows the order as shipped. This is the expensive one, and it is why tracking cargo before it leaves the origin port pays for itself. The event exists. The publication does not.

    The pipe is broken. An integration failed, a credential expired, a partner changed a format. This one belongs to IT or to your provider, not to the planner, and routing it to the planner wastes both.

    Rank silence by what it is hiding

    Not every quiet shipment deserves a phone call. A silent container three weeks from a regional warehouse with four weeks of cover is a note. A silent container feeding a production line next Tuesday is today's work. Triage by consequence, the same way you would prioritize any other exception when everything is late at once.

    This is ordinary exception management applied to a signal most teams never feed into it: detect the gap automatically, rank it by what it threatens, and give it an owner and a deadline. OIA Global cut exception resolution time by 60% in a quarter working this way. The gain does not come from finding more problems. It comes from spending the week on the four that matter instead of re-checking sixty portals to find them.

    Where to start this week

    Pick your top twenty lanes and list every data source behind them. For each one, write down when you last received anything, and how often it normally publishes. You will usually find one or two that have been dead longer than anyone realized. That cadence is worth keeping once you have it: how fast a lane tells you something is wrong is one of the inputs to a lane risk score.

    Then set a heartbeat on those sources, load this month's holiday calendar so the restart from 8 October does not drown you, and route a broken feed to the person who can fix the feed rather than the person chasing the box.

    The teams who come through a congested October well are not the ones watching more shipments. They are the ones who notice, on the day it happens, that a source stopped talking. Getting there means treating an absent update as an event in its own right, which is what exception management is for: the gap is detected as it opens, ranked against everything else in play, and handed to someone with a deadline, before a customer calls to tell you about it.

    Trusted by leading supply chain teams
    MatalcoOIA GlobalKingsdownDBWMatalcoOIA GlobalKingsdownDBW
    DLAIllyLevitt SafetySquareInstoreDLAIllyLevitt SafetySquareInstore
    TAKE THE NEXT STEP

    Action breaks down when operations are disconnected.

    Orkestra gives teams one place to see what matters, coordinate the response, and move operations forward.