Trade roles
A company appearing in the export record for the first time is a different proposition from one with a decade of history — for a buyer assessing risk, and for a service provider looking for customers at the moment they need help..
A company appearing in the export record for the first time is a different proposition from one with a decade of history — for a buyer assessing risk, and for a service provider looking for customers at the moment they need help.
| Signal | What it tells you |
|---|---|
| First appearance | The first cleared shipment is a datable event |
| Ramp shape | Whether volume builds or stalls after the first consignment |
| Single destination | Most start with one market, which is where support is needed |
| Small consignments | Trial-sized shipments, often LCL or air |
| Documentation gaps | Early filings are where classification errors cluster |
Method
Four steps that turn a record set into a shortlist.
Not as absent, but not as proven either.
A repeat consignment is the real signal, not the first one.
Early exporters frequently file under an approximate code.
Ramp shape tells you what they can realistically commit to.
Every other way of assembling this list starts from something a company chose to publish about itself. A directory entry exists because someone filled in a form. A marketplace listing exists because someone paid for visibility. A search result exists because someone invested in being found. None of those are evidence of trading; they are evidence of marketing, and the two correlate far more weakly than anyone would like.
A customs record exists because a consignment physically crossed a border and somebody filed a declaration for it. It carries a date, a classification, a quantity and a value, and it was filed under legal obligation rather than for promotional purposes. That is a different class of evidence, and it changes what you can conclude from a list.
Of the signals in the table above, first appearance is the one people misread most often, because it is the easiest to observe and the easiest to over-interpret. A single period is not a pattern. Establish what normal looks like across at least two to three years before treating any deviation as meaningful, and remember that the most recent one or two periods are still filling in — an apparent collapse at the end of a series is usually reporting lag rather than a business event.
The other signals are best read together rather than in isolation. Any one of them can be explained away; three of them pointing the same direction is a finding. Where they conflict — strong volume with erratic frequency, say, or a broad product mix with a single origin — the conflict itself is the interesting part, and it is normally the first question worth asking on a call.
The steps above are deliberately in order, because the order does most of the work. Filtering to a serviceable market before ranking on volume prevents you from building an impressive list you cannot act on. Ranking on behaviour before value prevents you from chasing one large occasional buyer while ignoring five steady ones. Checking the price band before outreach prevents the most demoralising outcome of all, which is three good conversations that end on price.
| Shortlist size | What it is good for | What it costs you |
|---|---|---|
| 5–15 names | Deep qualification, tailored outreach, high conversion | Slow to build, and one bad assumption removes a large share of it |
| 50–100 names | A working pipeline with room for attrition | Qualification has to be partly automated or it will not get done |
| 500+ names | Market structure analysis, not outreach | Almost nobody works a list this size; it becomes an artefact |
Two minutes spent on a counterparty’s history before the first message saves far more than it costs. Confirm they have shipped recently rather than historically; confirm the product genuinely matches rather than merely sitting in the same chapter; confirm their typical consignment size is compatible with what you can supply or absorb; and confirm the declared values sit in a band you can work in. Anything that fails those four is not a lead, it is a distraction with a name.
This is also where the most expensive mistakes get caught. Extending credit, committing production capacity or wiring a deposit against a counterparty whose trading history nobody checked is the failure that ends programmes rather than merely costing a quarter. The check is available, it is quick, and there is no good reason to skip it.
Generic outreach is ignored, and it is ignored at a rate that makes list quality almost irrelevant if the message is wrong. Referencing the specific product, the volume band and the origin a counterparty already trades demonstrates that the approach is considered rather than broadcast. It also, usefully, filters out the recipients for whom you have misread the record, because they will tell you.
Most categories have a buying calendar. Outreach that lands in the quarter budgets are set converts very differently from outreach that lands at random, and the calendar is visible in the seasonal shape of the record itself.
It is worth being explicit. Customs data covers goods that crossed a border. It does not cover domestic trade, it does not cover services, and it says nothing about margin, contract terms, payment behaviour or intent. A company that trades mainly within its own market will look small here and can be large in reality. Treat the record as a strong, dated, quantified observation to be corroborated — not as a conclusion that arrives pre-formed.
It is also not a compliance screen. Shipment history is useful evidence of what a counterparty actually does, and it supports due diligence, but it is not a sanctions check and it is not advice on your obligations. Where those obligations apply, they have to be discharged separately and documented at the time.
The teams that get the most out of this do not run one large project; they run a small recurring one. A weekly pass over a saved query — new counterparties appearing, existing ones changing origin, volumes moving outside their normal band — surfaces the handful of events worth acting on and ignores the rest. That is a half-hour habit rather than a research programme, and it compounds in a way that one-off analyses never do.
The second habit worth forming is writing down what you concluded and when. Trade records get revised, markets move, and six months later nobody remembers whether a counterparty was rejected because the volumes were wrong or because the timing was. A short note against each name turns a list into an institutional memory, and it is the difference between a pipeline and a spreadsheet somebody once made.
| Cadence | What to look at | Why |
|---|---|---|
| Weekly | New names in your chapters and markets | First appearances are the most actionable events in the record |
| Monthly | Volume and origin changes among names you track | A sustained shift usually precedes anything announced publicly |
| Quarterly | Price band across the tariff lines you trade | Bands move, and a stale benchmark is worse than none at all |
| Annually | Full market structure — concentration, entrants, exits | The slow changes are invisible at any shorter interval |
Related
A company appearing in the export record for the first time is a different proposition from one with a decade of history — for a buyer assessing risk, and for a service provider looking for customers at the moment they need help.
Markets refresh on their customs authority's own release cycle — monthly for most, 45 to 60 days for a few. A list is only as useful as its most recent period, so check the latest date before you work it.
Yes. Country, HS code, date range, port and value band are all filters, and combining them is what turns a large record set into a workable shortlist.
Search results export to Excel and CSV. Scheduled extracts and API delivery are available for recurring use.
At least two to three years. One period is a snapshot, two make a line, and three years make a pattern you can distinguish from seasonality. Exclude the most recent one or two periods entirely — they are still filling in, and reading that as a decline is the most common false signal in trade data.
Frequently. A declaration records the name as typed, so one business becomes several strings across legal-suffix variants, abbreviations, transliteration, branch filings and simple keying errors. Unresolved names inflate counterparty counts and deflate individual volumes, which makes a concentrated market look fragmented. Check which variants have been folded into a profile before you rely on it.
Yes, and it is usually where the value compounds — matching a shipment history against accounts you already hold shows which of them are growing, which have started buying from a new origin, and which have quietly stopped. Plan for entity resolution on the join, because company names will not match cleanly on either side.
No. A customs record exists because goods physically crossed a border. Services do not appear at all, and domestic trade does not appear either, so a business that sells mainly inside its own market will look small here and can be large in reality.
Read value and quantity separately. Declared value can rise because volume rose, because unit price rose, or because the mix inside the tariff line changed — and those lead to different decisions. Value alone cannot distinguish them, which is why it is the most over-interpreted number in trade statistics.
Give us the HS code and the market and we will send back a sample of the records this page describes.