Transport emissions reporting fails on data, not on methodology. The calculation itself is settled: distance, weight and a factor. What breaks is that the shipment records feeding it were never designed to be audited, and by the time an auditor asks where a number came from, the answer is a spreadsheet nobody can reproduce.
This guide follows the chain from one shipment to a disclosure that survives review: what to capture, which standard applies, where modelled data is acceptable and where it is not, and what an audit trail actually has to contain.
The data model: what one shipment has to carry
Every shipment that will end up in a disclosure needs six fields. Not more, and rarely fewer:
- Origin and destination. Postcode level at minimum. City level introduces error that compounds across a year.
- Distance actually travelled. Not the straight line, and not the planned route if the vehicle deviated.
- Chargeable weight. The basis for allocating a vehicle’s emissions across the loads it carries.
- Vehicle profile. Category and, where known, Euro class and fuel type.
- Load factor. The share of vehicle capacity your freight occupied.
- Empty running. The positioning leg attributable to your shipment.
The last two are where most reporting quietly understates. A shipment reported on its loaded leg alone ignores the kilometres run to reach it, which on some lanes exceeds the loaded distance. The cost side of the same problem is covered in how to reduce empty miles.
Which standard applies, and why it matters for the number
Two references govern European road freight reporting. ISO 14083 defines how to quantify greenhouse gas emissions of transport chains, including how to handle hubs and allocation between consignments. The GLEC Framework operationalises it for logistics and is what most European shippers and carriers align to in practice.
The practical consequence is allocation. Two companies can move identical freight and publish different figures, legitimately, because one allocated by weight and the other by tonne-kilometre, or because one included the empty positioning leg and the other did not. If your disclosure does not state its allocation basis, it is not comparable and an auditor will say so.
Primary data, modelled data and the line between them
Primary data means measured: actual distance from telematics, actual fuel from the carrier, actual weight from the consignment note. Modelled data means derived from factors and averages.
Modelled data is acceptable, and for most shippers it is unavoidable in year one. What is not acceptable is failing to label which is which. A defensible disclosure states the share of tonne-kilometres backed by primary data and shows that share rising year on year. Auditors treat a static 100% modelled position across three reporting cycles as a governance finding, not a data finding.
The practical route to primary data is the carrier’s telematics, which is the same connection that produces freight visibility. That is why emissions reporting and visibility tend to become one project: both depend on connecting the same vehicles.
What CSRD actually requires from transport
Under CSRD and the ESRS E1 climate standard, transport sits in Scope 3 and is material for most manufacturers, retailers and distributors. Materiality is the trigger: if upstream and downstream transport is material to your value chain, it must be reported, with methodology disclosed and year-on-year comparison possible.
What that means operationally is less about the total and more about traceability. You must be able to explain how a figure was produced, which data was measured and which modelled, and why the boundary was drawn where it was. Our guide to CSRD and freight covers the reporting obligations in detail, and the CSRD Scope 3 freight emissions calculator lets you produce a first estimate from data you already have.
The audit trail, in practice
An audit trail is not a report. It is the ability to take any published figure, drill to the shipments behind it, and show for each one where distance, weight and factor came from. Three requirements follow:
- Shipment-level retention. Aggregates cannot be re-derived. Keep the rows.
- Factor versioning. Emission factors are revised. A recalculated prior year must record which factor version produced which figure.
- Change history. When a carrier restates a distance, the restatement and its reason must be visible.
Software that produces a headline number without exposing the shipments underneath will pass a board review and fail an audit. This is the distinction between a dashboard and a reporting system, and it is worth establishing before the first disclosure rather than after.
A sequence that works
Start by fixing shipment records, not by choosing a platform. Capture the six fields on every movement for one quarter and you will discover which lanes have no usable data, which is where the project’s real work is. Then connect the carriers covering the largest share of your tonne-kilometres, because primary data on 30% of volume is worth more than modelled data on all of it. Only then automate.
Continuous tracking of the resulting figures is what carbon visibility software does, and the operational levers that actually move them, rather than the accounting ones, are collected in green logistics strategies that reduce emissions.
