We Tested the Official Trackers of 18 Shipping Lines

We opened each carrier’s official tracking page in a desktop browser with a real container number and recorded three things: which identifier types the form accepts (container number, bill of lading, booking reference), whether a URL carrying the number opens the result directly, and whether the carrier publishes a mobile app. Tracking pages were reached from the carrier’s own navigation, never from an aggregator.

A deep link counts as verified only when a URL containing the number returns the tracking result on load. Where a carrier accepts the parameter but ignores it — the page opens on an empty form — we record that as a tested negative, which is a different and more useful finding than silence.

Eleven carriers are marked "not tested". That is a statement about our coverage, not about their capabilities: we have not run the test, so we do not publish a verdict. Marking untested rows as "no" would inflate the headline finding and is exactly the kind of quiet error this study exists to avoid.

Fleet ranks are Alphaliner figures for 2026 and alliance membership is as of the same period. Prefix ownership comes from the BIC register, each entry confirmed against a second independent source — the method is described on how we verify data.

The headline is the deep-link column. Three carriers out of eighteen let a URL carry the container number, which means that for the other fifteen no tool — ours or anyone else’s — can hand a customer a working one-click tracking link. That single design decision explains why container tracking still feels like a chore in 2026.

The identifier columns are more uniform than expected: every tracker takes a container number and a B/L number, so the common advice to "use the B/L if the container number fails" is at least universally possible. Booking numbers are the split: 10 of 18 accept them, and the eight that do not are mostly the smaller regional lines.

The second half of this research is ownership. A container number begins with a four-letter BIC prefix identifying the owner, and the industry routinely gets specific prefixes wrong — including tracking aggregators that ought to know better. We hold 85 prefixes across 28 owners, each confirmed against two independent sources.

The misattributions worth naming: TCKU is Triton, a leasing company, though it is widely labelled as Maersk. TEMU and TGBU are Textainer, frequently credited to Triton — and TEMU predates the shopping app of the same name by decades. SEGU and GESU moved to Textainer when it absorbed the Seaco fleet in late 2025, so older tables still show them as Seaco. CBHU is registered to Florens, the COSCO group’s leasing arm, yet the boxes run in COSCO’s own fleet.

The distinction matters commercially, not just pedantically: 23 of our 85 prefixes belong to lessors, and a leased box is tracked with whichever line booked the freight, not with the company whose name is on the door. Send a customer to a leasing company’s website and they will find nothing at all.

Read the prefix before you open a website. If the four letters belong to a lessor, skip the owner entirely and go to the line named on your booking confirmation or bill of lading — the container tracking checker does this routing automatically.

Expect to type. With only three deep-link carriers, any workflow built on saved links will break for most shipments; a saved list of tracker URLs plus the number on your clipboard is the realistic setup.

Keep the B/L number to hand as the universal fallback: it is the one identifier every tested carrier accepts, and it survives the case where a container is re-stuffed or the number was misread off the door.

The full results are published as an open dataset — carriers.json and carriers.csv — free to reuse with attribution to contsize.com. Fields cover prefixes, role, tracker URL, accepted identifiers, deep-link status, app availability, alliance and fleet rank.

If you work at a carrier and our row is wrong, or you have tested one of the eleven untested trackers, tell us and we will re-test and update both the table and the dataset. Corrections are logged publicly on the about page.