Primary vs secondary emissions data for Scope 3: what counts, and where to get it

Primary emissions data is data that comes from a specific activity inside your own value chain: a supplier's reported inventory, a product footprint they published, the fuel a carrier actually burned. Secondary data is everything that does not: a sector average, a factor per pound of spend, a regional grid mix. Most Scope 3 teams believe the first kind has to be collected. Most of it already exists. This page sets out what the GHG Protocol counts as primary, where each kind sits in a Scope 3 calculation, what each costs to obtain, and where to find primary emissions data before you send a single request.

Definitions from the GHG Protocol Corporate Value Chain (Scope 3) Standard. Page last reviewed 22 September 2026.

What is primary emissions data?

The GHG Protocol Scope 3 Standard defines primary data as data from specific activities within a company's value chain, and secondary data as data that is not from specific activities within that value chain. The test is provenance, not format. A figure is primary if it was produced by, or measured at, the organisation or activity you are accounting for. It is secondary if it describes a category of organisation or activity in general.

Three kinds of data pass the test for a purchased good or service:

• The supplier's own reported greenhouse gas inventory. A company that publishes its Scope 1 and 2 emissions and its revenue has produced primary data about itself. Allocated to what you bought from it, that is primary data about your Category 1 line.
• A product carbon footprint the supplier published. A cradle-to-gate or cradle-to-grave figure per unit for the thing you bought, produced to a named standard.
• Activity data from the value chain. Litres of fuel a carrier burned on your freight, kilowatt hours a contract manufacturer consumed on your run, tonnes of a material a supplier shipped to you.

One kind fails it: an industry average. A factor per pound of spend or per tonne of output, drawn from an environmentally extended input-output model or a lifecycle database, describes the sector rather than the supplier. It is secondary by definition, however large and well maintained the library it came from.

Why does the distinction matter for Scope 3?

Because the kind of data decides what the number can do.

A Category 1 figure built from a sector average on spend moves when procurement spends less. It cannot move when a supplier cuts its emissions, because the factor does not know the supplier exists. Two suppliers in the same sector get the same figure per pound whatever either of them has done. A team can report that number. It cannot act on it, engage with it, or show an auditor where it came from beyond the name of a database.

The same line calculated from the supplier's own inventory moves when the supplier reports a reduction, separates that supplier from its peers, and traces back to a document the supplier published. That is what reporting standards are now asking about: CSRD and CDP both ask how much of a Scope 3 total rests on supplier-specific data rather than averages, and PCAF scores a financed emissions figure from 1 to 5 on the same axis.

The distinction also decides the work. If primary data has to be collected, Scope 3 is a survey programme with a response rate as its ceiling. If most of it already exists, Scope 3 is a matching problem, and collection is for the gaps.

Is my supplier's published inventory primary data?

Yes, and this is the point most teams miss.

Readers usually hear "primary" as activity or product data that the supplier hands over for your specific purchase. That is primary data, and it is the hardest kind to get. But the supplier's own corporate inventory is also data from a specific activity within your value chain: the activity is the supplier's operation, and the supplier measured it. The GHG Protocol's definition turns on where the data came from, not on whether it was produced for you.

The step that makes it usable is allocation. A supplier's total emissions divided by its revenue gives an emission factor that belongs to that supplier alone. Multiply it by what you spent with them and you have a Category 1 figure for that line. The data is primary, reported by the organisation itself, and the allocation is yours. That is the sentence an auditor wants: primary emissions data, reported by the organisation, allocated to your spend.

This is not the spend-based method. The GHG Protocol defines its spend-based method as spend multiplied by secondary emission factors, meaning industry averages. Once the factor comes from the supplier's own published inventory it is no longer secondary data, so it is no longer the spend-based method, even though the input is still a line of spend. Spend is the input. The factor decides whether the result is primary or secondary. The market collapses those two constantly, and the collapse is what makes primary data sound like a project rather than a substitution.

The four methods, and where each sits

The Scope 3 Standard gives four calculation methods for purchased goods and services. Three produce primary data, one produces secondary data, and they do not cost the same.

The first row is the one most teams do not know is available. It converts a line from something nobody can act on into something they can, with no additional burden on the team and none on a supplier that has already published. The product and activity rows are better still, and they cost real time from you and often work from the supplier, so they belong on the lines carrying the emissions rather than across a whole tail. A programme that claims every improvement is effortless is describing a brochure.

The bottom row is not wrong. It is the floor. Where a supplier has published nothing and nothing comes back, a sector average closes the inventory, and the honest thing is to show how much of the total still rests on it rather than to hide the share inside a single figure.

Where does primary emissions data already exist?

Before anyone is asked, three kinds of primary data exist for a large share of a typical Category 1 list or portfolio.

Published inventories. Listed companies, large private companies in jurisdictions with mandatory reporting, and anyone who has answered CDP has put their Scope 1 and 2 figures, and often their revenue, into the public record. Matching those disclosures to the legal entities on your supplier list is the work, and it is a matching problem rather than a collection one: DUNS, LEI and ISIN identifiers, parent and child hierarchies, name changes and acquisitions.

Published product carbon footprints. Manufacturers in materials, chemicals, packaging, electronics and building products increasingly publish PCFs and EPDs per product, to PACT, ISO 14067 or EN 15804. Each one is a primary figure per unit, with a standard, a boundary and often a third-party verifier attached.

Data suppliers have already provided for other buyers. A supplier that has answered one customer's request for activity or product data has done the work once. Where it agrees to share that data more widely, a new buyer can use it from day one. Sharing is always the supplier's decision, and a disclosure made privately for one client is never shared with anyone else. What the supplier chooses to make general is what widens the pool.

This is the layer DitchCarbon is built on: verified emissions data for over 2 million organisations, built on primary emissions data wherever it exists, matched to your suppliers or holdings by entity resolution against DUNS, LEI and ISIN identifiers. Each organisation records which of the four methods produced its figure, so the share of your baseline still resting on secondary data is a filterable number rather than a claim, and every figure carries its source and change history. In a recent deployment that reached about 60% of a large supplier base within 2 weeks, before a single request went out.

How do you get primary emissions data from suppliers who have not published?

By asking only where it changes the total, and by asking for less.

Most of a supplier list is a long tail whose emissions barely move the number. Requests belong with the organisations where a primary figure would change your Category 1 total, which is a short list once the published layer is in place. Those requests should arrive prepopulated with what the organisation has already disclosed, so the supplier reviews and corrects rather than compiles from scratch. Prepopulated requests get a higher response rate than a cold survey.

The sequence matters. First, coverage that exists before anyone is asked, so often no request is needed. Second, where a figure does need improving, a targeted request that goes out prepopulated. Third, where nothing has been disclosed and nothing comes back, the sector average closes the inventory and the gap is shown rather than hidden. Any model that has to ask every supplier before it knows anything adds to the queue of forms that supplier's one-person sustainability team already receives, whoever is running it.

How do you show the primary share to an auditor?

Record the method per line, and make it visible.

For each organisation in the baseline, record which method produced its figure: its own reported inventory, a published product footprint, activity data, or an industry average. Then the questions an auditor and a reporting standard both ask fall out as filters. What proportion of Category 1 spend is on primary data, actual data rather than an estimate? Which suppliers moved from secondary to primary this year? Which figures changed since last quarter, and what document changed them? Every figure carrying its source and change history is what turns "audit-ready" from a phrase any vendor can print into something a third party can follow back. On the DitchCarbon side, the calculator that produces those figures is verified to ISO 14064-3, limited assurance, by UL Solutions, renewed annually; the reports are in the trust centre.

If you are the supplier being asked

If you landed here because a customer or an investor asked you for primary emissions data, the fastest answer is the data you have already published. A company that reports its inventory has primary data; the buyer's job is to find and allocate it. Claim your organisation's profile, check what is already held against your name, add the product footprints and activity data you are willing to share, and choose who can read them. Answer once and every buyer using the same layer reads the same answer, instead of the fifteenth version of the same form.

The short version

Primary emissions data is data from a specific activity in your value chain; secondary data is a category average. A supplier's own published inventory, allocated to your spend, is primary data, and it costs nothing extra to use. Product footprints and activity data are primary too, and they cost real effort, so spend it where the emissions are. A sector average is the floor, not the method. DitchCarbon starts every baseline from the primary data that already exists for over 2 million organisations, shows which method stands behind each figure, and sends requests only where a primary figure would change the answer.