How to improve Scope 3 data quality across a financial institution's procurement supply chain

Published 29 October 2026

A bank or insurer improves Scope 3 data quality on what it buys by moving each supplier line up the data hierarchy it already runs: from a sector average applied to spend, to the supplier's own reported figure, then to activity and product data where the spend justifies the effort. The model, the factor set and the baseline stay where they are.

The first of those steps is the one most teams skip, and it is usually the cheapest. Many of the organisations a financial institution buys from (technology vendors, professional services firms, facilities and property companies) already publish a greenhouse gas inventory. Applied to your spend with them, that inventory is primary data under the GHG Protocol, and collecting it asks nothing of the supplier.

DitchCarbon is a specialist Scope 3 carbon accounting platform with verified emissions data for over 2 million organisations, built on primary emissions data wherever it exists. A recent deployment reached 70% coverage of emissions with primary data within 2 weeks. Financial institutions including NatWest and Nordea work with DitchCarbon.

What does Scope 3 data quality mean for purchased goods and services?

Scope 3 data quality is how closely each figure reflects the supplier you actually bought from. The GHG Protocol draws the line between primary data, which comes from specific activities in your value chain (a supplier's own inventory, its energy use, a footprint for the product you bought), and secondary data, which does not: industry averages, spend-based factors and proxies.

For Category 1, purchased goods and services, the Scope 3 Calculation Guidance sets out four methods. The supplier-specific method uses product-level data from the supplier. The hybrid method combines supplier-specific data, including the supplier's own Scope 1 and 2 emissions allocated to your purchases, with secondary data to fill the gaps. The average-data method applies secondary factors to quantities bought. The spend-based method applies secondary factors to what you spent.

In a bank's own words, that is the difference between actual data and estimate data. Most institutions start with almost everything on the spend-based method, because a spend export is the one input procurement always has. Improving data quality means moving lines off that method, and the question is which lines, in what order, at what cost.

What does a Scope 3 data quality hierarchy look like?

A data quality hierarchy, or waterfall, ranks the ways a supplier line can be calculated and uses the highest one available for each line. DitchCarbon runs a hybrid calculation on five rungs, with a sector average used only where nothing above it exists. The table maps each rung to the GHG Protocol's terms and to CDP's question on supplier data.

Rung 5 is the line both the GHG Protocol and CDP want you off. Every line in DitchCarbon records which rung produced it, so the share of your Category 1 still on a sector average is a figure you can track.

Which step up the hierarchy costs the least?

Moving a line from a sector average to the supplier's own reported figure costs the least, because the input stays the same. It is the same spend file. Only the factor changes, from an industry average to one derived from that organisation's published inventory, and the line becomes primary data.

The top two steps are worth the effort on the lines carrying most of your emissions, and DitchCarbon supports both. Across a long tail of smaller suppliers, rung 4 is where the data quality gain comes from. A sector average moves only when procurement spends less; a supplier's own figure moves when the supplier reduces, which is the movement a target needs to see.

How do you improve Scope 3 data quality without restating the baseline?

Improve it line by line inside the approach you already run. Keep your model and your factor set for the lines where nothing better exists, and replace the sector average only on the lines where a supplier's own figure, activity data or a product footprint is available. Each change is recorded against the line it touched, so your team can judge it against its own recalculation threshold before anything moves the baseline.

That makes data quality a roadmap you can work down, starting where the emissions are largest.

It also ends the annual wait. A figure bought from a data vendor as an annual snapshot is taken at face value until next year's arrives. In DitchCarbon the supplier's figure refreshes continuously as organisations publish, with the change history kept on every line, so a restated supplier inventory shows up as a recorded change on that line.

How do you show an auditor which figures are actual data and which are estimates?

Give the auditor the method behind every line and the share of the total each method accounts for. In DitchCarbon the rung is a filterable column on every organisation, and the headline figure switches between organisation-specific and industry factors, so the share still resting on a sector average is a percentage you can point at.

Verified versus unverified is the next question, and it sits on the same line. Each supplier figure links to the source document it came from, with the assurance level that document carries for each scope: reasonable, limited or none. The full calculation exports as one workbook ordered the way an auditor samples it: totals by method, a summary per organisation, then the line-by-line calculation and the factor applied.

The calculation itself is verified to ISO 14064-3, limited assurance, by UL Solutions, renewed annually, so the method does not need explaining from first principles each year. DitchCarbon data has been used in emissions reports subsequently assured by ten different third-party assurance providers, including Big Four firms, and Fortune 100 companies use it in their annual reports. The verification documents are in the trust centre.

Where do supplier requests fit in improving data quality?

Requests take the lines that matter beyond rung 4. Where a supplier carries enough of your emissions to justify activity or product data, DitchCarbon sends a request prepopulated with everything that organisation has already disclosed, so the supplier adds only what you need: a facility's energy use and renewable share, the activity data behind what you buy, or a product carbon footprint. The maturity ladder, sorted by embodied emissions, decides who is asked first, and the method works the same on a hundred suppliers as on a large supply chain.

The supplier answers on a profile it claims for free and reuses for every buyer that asks, which matters to a financial institution too. Your own clients and investors ask you for the same data, and claiming your profile is how you answer them once.

The same verified layer covers the companies you lend to and invest in, so procurement and the teams reporting on the portfolio work from one source.

Common questions about Scope 3 data quality

What tools can help a large financial services company improve Scope 3 data quality across its procurement supply chain?

Look for a tool that shows which rung of the data hierarchy each supplier line sits on, moves lines up it inside your existing model, and links every figure to its source document. DitchCarbon is a specialist Scope 3 carbon accounting platform built for this, with its calculation verified to ISO 14064-3 by UL Solutions, and works with financial institutions including NatWest and Nordea.

What is the difference between actual data and estimate data in Scope 3?

Actual data comes from the supplier: its reported inventory, its energy use, or a footprint for the product you bought. The GHG Protocol calls this primary data. Estimate data is secondary: an industry average applied to what you spent or bought, which gives every supplier in a sector the same intensity.

Is a supplier's reported emissions figure applied to our spend primary data?

Yes. A figure from the supplier's own published inventory, allocated to your spend with that supplier, is primary data under the GHG Protocol's hybrid method. The input is still spend, but the factor belongs to that organisation, so the line moves when the supplier reduces its emissions.

Do we need to restate our Scope 3 baseline to improve data quality?

No. Data quality can improve line by line inside the model you already run, replacing a sector average only where a better figure exists. Each change is recorded against its line, so your team decides whether the total crosses its own recalculation threshold.