Methodology

How a SeaLines report is built.

These are the rules written into every report. We publish them so you can check our numbers instead of taking them on trust.

The data

Every SeaLines report is built from real customs declarations and bills of lading — the paperwork filed when a shipment crosses a border. Nothing is a survey or an estimate. A "filing" is one declared line; one bill of lading can carry several lines. We count filings unless we say otherwise.

When a search returns a set of filings for your product and period, we read all of them. No sampling. Page two of every report says how many filings matched, how many we read, and how many bills of lading they belong to.

Values and prices

Values are the customs values declared at the border — the invoice amount. A unit price is that value divided by the declared net weight. It is the closest thing to a real market price that exists, and the best guide to what competitors are charging. It is not a wholesale or retail price, and we never present it as one.

We report medians and ranges, not averages, so one unusual shipment can't distort the picture. When prices cluster in two or three separate groups, the code contains more than one product, and we size each one on its own price level.

What we leave out of the price figures, and why

Filings with no declared weight — declared in packages, pieces or units — have no unit price. They stay in every value total and are left out of every price figure. We say so under each affected table.

Filings that don't add up are handled, not ignored. A weight error that implies an impossible price, or a $1 "sample" declaration, stays in the value totals, is left out of the price figures, and is named in the limits section so you can see exactly which records we treated differently.

Buyers and sellers

The exporter is the company shipping from origin; the buyer is the company receiving abroad, as named on the declaration. We tidy up names: a company spelled two ways is one company, and a group that appears on both sides of its own shipment is identified as such, not counted twice. Shipments between related companies are separated from real sales before we draw any conclusion about demand or price.

For each market we report what share of filings name the buyer (rather than "to order"), so you know where the buyer list is complete enough to act on.

Gaps are findings too

Coverage varies by country and by month. If a country that should dominate a product is barely in the record, the report says so and does not draw conclusions from its absence. One month can't show seasonality, trends or how stable a supplier is, and we don't pretend it can; where a longer series exists, we show it.

Every report ends with a limits section, an appendix listing every destination market with its counts, and a list of definitions, so nothing in the report rests on an undefined word.

What we don't do

  • No scraped or unlicensed data. Our records are licensed customs and port data, and we never reproduce raw database records in a report.
  • No surveys, no estimates dressed up as facts, no averages across codes that mix different products.
  • No guarantees of results. Our recommendations are informed opinions on the evidence available on the date of issue.
  • No substitute for a customs broker. Confirm tariff classification and regulatory requirements with a licensed broker before you commit to anything.

Turnaround, confidentiality and corrections

Standard reports are delivered within seven working days. Professional and Enterprise plans get priority and a three-working-day turnaround. Each report is prepared for the client named on its cover and is not for passing on. Your questions, codes and reports are seen only by you and your analyst.

If a delivered report contains a factual error, we issue a corrected edition free of charge. If we can't answer a question because the records don't exist for the period you asked about, you can choose a replacement question or a pro-rata refund of that report.

Methodology — SeaLines