Getting Carbon Data From Your Suppliers

Most supplier carbon programmes begin by sending a questionnaire to every supplier and end with a spreadsheet of partial responses nobody can use. The failure is in the sequence rather than in the effort.

Suppliers have no obligation to you under any of the instruments driving your request, and most do not hold product-level emissions data in any form. Starting from that reality produces a better plan than starting from the data you wish existed.

Step 1 — Screen before you ask

Build a first footprint entirely from secondary data. It will be wrong in detail and correct about concentration, and concentration is the only thing you need at this stage.

The usual finding is that a small minority of purchased inputs accounts for the large majority of upstream emissions. Those are the suppliers worth spending relationship capital on, and the rest can stay on secondary data indefinitely without materially affecting the result.

Step 2 — Ask through the right channel

Requests that arrive through procurement, alongside quality and delivery discussions, are answered. Requests that arrive from a sustainability address a supplier has never dealt with are not, or are answered late and thinly.

This is the single highest-return change available and it costs nothing except an internal agreement. It also means carbon data becomes a supplier performance topic rather than an annual survey, which changes how seriously it is treated.

Step 3 — Ask for the right thing

Requesting a single number is the common mistake, and it produces data that cannot be aggregated or defended. The additional fields cost the supplier little once they have done the calculation, and asking for them later means asking twice.

Step 4 — Expect a capability gap

A large share of suppliers will not hold the data, particularly smaller ones and those outside the EU. The first request is therefore often a request for a capability they do not have rather than for information they are withholding.

• Give a defined format and a worked example rather than an open question.
• Accept a partial figure with stated gaps over no figure at all.
• Point to a sector calculation method where one exists, rather than asking them to choose.
• Offer a realistic deadline; a four-week window on a first request is not realistic.

Larger buyers sometimes fund or provide calculation tooling for strategic suppliers. That is expensive and it works, and it is worth considering for the handful of suppliers who dominate a footprint and cannot answer.

Step 5 — Handle non-response deliberately

Some suppliers will not answer. Record what you asked, when, through which channel, and what came back. That record is required rather than merely tidy.

Under Directive (EU) 2022/2464 the transitional provisions on value chain data allow you to explain why information could not be obtained. Explaining requires having tried, and an organisation with no record of asking is in a materially weaker position than one that documented the attempt.

Contract terms are the durable fix

Everything above is a campaign. The durable version is a clause in purchase terms requiring product-level emissions data in a defined format, with notification when a formulation or production route changes.

Such a clause takes effect at the next order rather than retroactively, which is why adding it early matters more than the wording. A clause agreed this quarter reaches production roughly a year before one agreed in four quarters’ time.

What to do with what arrives

Store the figure with its metadata against the product record rather than in a reporting spreadsheet. A number separated from its boundary and asserting party degrades into an average the first time it is copied.

This is where passport infrastructure earns its place in a carbon programme. The passport does not calculate anything; it preserves the distinction between an evidenced supplier figure and a value someone typed in, which is precisely what assurance examines.

What incentives actually work?

Three, in descending order of effect. Making the data a condition of supply works and costs relationship goodwill. Making it a scored criterion in supplier evaluation works more slowly and is better tolerated. Asking politely works only where the supplier already holds the data.

A fourth that is underused: telling suppliers what you will do with the figure. Suppliers who understand that a good number wins them share behave differently from suppliers who assume the request is an administrative exercise with no consequence.

That only holds if it is true. If carbon data has no bearing on sourcing decisions, suppliers work that out quickly, and subsequent requests are answered with whatever is cheapest to produce.

How do you check what arrives?

Not by auditing every figure, which is not affordable. By sanity-checking against secondary data: a supplier figure far below the database average for that material is either a genuine achievement or a boundary difference, and it is worth knowing which.

The usual explanation is boundary rather than performance. A figure covering only manufacturing while the database entry covers cradle-to-gate will look dramatically better and is not comparable, which is exactly why the boundary field is requested alongside the number.

What about suppliers who send a corporate figure?

A common response is a company-level total, sometimes divided by revenue or output. It is not a product footprint and cannot be used as one, for the same reasons a corporate figure divided by units shipped is not.

The productive reply is specific rather than a rejection: acknowledge the corporate figure, explain that you need emissions attributable to the product you buy, and offer the format. Most suppliers sending a corporate total are being helpful with what they have rather than evading.

How long does this take?

A first cycle for a meaningful share of a footprint usually takes two to three quarters, and the constraint is supplier calculation capability rather than your own effort.

Second cycles are much faster because the format is established and the suppliers have done it once. Programmes that plan a single push and expect completeness are consistently disappointed; those that plan two cycles usually get there.