Getting carbon data out of accounting, HR and purchasing

Twelve PDFs in the facilities inbox: a year of electricity. A spreadsheet the fleet manager exports on the second of every month: a year of fuel. Four clicks in the accounting system: a year of purchases.

None of it was produced for you, and that is exactly why you will get it quickly. The rule that saves weeks on a carbon footprint fits in one sentence: ask for the documents a department already makes in the normal course of its work. The day you ask someone to create one, you go behind their real job, and their real job never stops.

The map fits on one page

Before sending a single message, write down, for each emission source, the department that holds the data and the document that carries it. For a mid-sized company the list is short:

Writing the map down turns a vague campaign into a finite list. A finite list can be managed; a campaign cannot.

Energy and fuel are already counted

Electricity and gas are the easy win, because the invoice states the consumption itself, in kilowatt-hours and cubic meters. For a multi-site company the only real work is consolidation: one line per site, one column per energy, twelve months with no gaps. Resist converting from euros. Prices move for reasons that have nothing to do with consumption, and a footprint built on spend will show a reduction the year your supplier contract improves.

Fleet fuel behaves the same way: card statements give liters per vehicle, which is exactly the activity data the calculation wants. Refrigerants hide one level down, in maintenance reports. The quantity recharged into an air-conditioning or cold-storage circuit over the year is the quantity that leaked, and for a company with a lot of cold, that line surprises people.

Commuting is the only real questionnaire

Nothing in your systems knows how people get to work, which makes this the one source where asking employees directly is the right instrument rather than a last resort. Limit the questionnaire to what the calculation consumes: usual mode of transport, one-way distance, days on site per week. Three questions, five at the very most, answerable in a minute.

Response rates fall with every question added, and the arithmetic favors brevity: 30% of employees answering three questions, extrapolated transparently, beats 5% answering a beautiful survey.

Good to know: remote work belongs in this questionnaire rather than in a separate one. A day at home removes a commute and adds home energy, and "days on site per week" captures both sides in a single number.

Purchases come through accounting

Purchased goods and services are usually the largest scope 3 line and always the least obedient, because the underlying data belongs to hundreds of suppliers who mostly do not have it either. The workable first-year method starts inside your own building: an export of spend by expense category, mapped to monetary emission factors.

Ask for the full year, at the finest category level the chart of accounts supports, with intercompany flows excluded. One file, one department, covering what may be more than half your footprint. Those factors are averages, which is a real limitation and the subject of the physical-versus-monetary rule; for collection purposes it changes nothing, because you need the export either way.

Freight sits between two worlds. If you buy transport as a service, ask logistics or your forwarder for tonnages, lanes or distances, and modes. Forwarders field this request routinely now, and many will return emissions figures directly.

Collection is a dependency list

Once the map is drawn and the requests are out, what remains is follow-through, and follow-through is a table: source, owner, document requested, date promised, date received, gaps noted. Review it weekly. Escalate through your sponsor when a source is two weeks late, not two months, because a source that is two months late has usually not been refused, it has been forgotten.

Expect gaps, and write down how each was filled: a missing quarter extrapolated from the other three, a small site estimated from a comparable one. The footnote matters more than the patch. Next year's footprint will be built by someone who needs to know what was measured and what was assumed, and that someone may well be you.

So finish the table before you touch a single emission factor. When every row has a document behind it, the hard part of the footprint is over, and the multiplication that everyone thinks of as the work turns out to be the fast part.

FAQ

What data do you need for a carbon footprint?

Activity data per emission source: energy consumption from invoices (kWh, m³), fuel from fleet card statements, refrigerant recharges from maintenance reports, travel from booking exports, commuting from a short employee questionnaire, purchases from an accounting export by expense category, and freight volumes by mode from logistics or the forwarder.

Why is purchase data the hardest part of a carbon footprint?

Because the underlying emissions data belongs to hundreds of suppliers, most of whom do not have it. A first footprint therefore starts from the company's own accounting export, mapped to monetary emission factors, and improves specific categories with physical or supplier data in later years.

How do you collect commuting data?

With a short employee questionnaire, since commuting exists in no company system. Ask usual transport mode, one-way distance, and days on site per week, and keep it under five questions. A modest response rate extrapolated transparently is enough for a first footprint.