How to calculate Scope 3 emissions from spend data, step by step

To calculate Scope 3 emissions from spend data, you multiply how much you spent with each supplier, grouped by category, by an emission factor expressed in kilograms of CO2-equivalent per dollar. Pull your accounts payable export, map every line to one of the fifteen Scope 3 categories, apply the matching spend-based factor, and sum the result. The output is a screening estimate: good enough to tell you where your footprint actually sits, and where to spend effort collecting better data next.

That is the whole method in one paragraph. The reason it takes a page is that each of those steps has a judgment call in it, and the difference between a defensible number and a useless one is in how you handle them. This is the standard first move for any company facing Scope 3 reporting, because spend is usually the only complete record you have of your entire value chain.

Why start from spend at all?

Scope 3 is every indirect emission in your value chain: your suppliers' factories, your freight, your business travel, your products in use. For most companies it is 70 to 90 percent of the total footprint. The problem is that you do not have activity data for most of it. You do not know the kilowatt-hours your contract manufacturer burned or the tonne-kilometers your freight traveled. What you do have, for every supplier, is what you paid them. Spend is the one dataset that covers the whole value chain, which is why the GHG Protocol explicitly allows the spend-based method and why every screening inventory starts here.

Step 1: Pull and clean the accounts payable export

Export a full year of AP lines: supplier name, general ledger account, description and amount. A year, not a quarter, because you want seasonality and annual purchases in the picture. Then clean it enough to be classifiable. Merge duplicate supplier spellings, strip out anything that is not a real purchase of goods or services (payroll, taxes, intercompany transfers, loan repayments), and make sure amounts are in one currency. If your expense and vendor data is already categorized cleanly, this step is short; if it is a raw dump, budget real time for it, because everything downstream inherits its mess.

Step 2: Map each line to a Scope 3 category

Every purchase belongs to one of the fifteen Scope 3 categories. In practice a handful of them carry almost everything. For most companies the concentration looks like this:

• Category 1, purchased goods and services. The biggest bucket for nearly everyone: raw materials, components, cloud, software, professional services, packaging.
• Category 2, capital goods. Machinery, vehicles, IT hardware, buildings. Accounted in the year you buy them.
• Category 4, upstream transport. Inbound and outbound freight you pay for.
• Category 6, business travel. Flights, hotels, rail, rideshare on company expense.

The classification key is what was bought, not who sold it. A payment to a "logistics" vendor might be freight (category 4) or a leased warehouse (category 8). This is the step where a spreadsheet quietly breaks down: with ten thousand lines, nobody can later explain why row 4,812 was called category 1, and an assurance reviewer will ask. The discipline is to record the reason for each mapping, not just the mapping.

Step 3: Apply the spend-based emission factors

A spend-based emission factor tells you how many kilograms of CO2e are associated with a dollar of spend in a given category. The most-used public source in the US is the EPA's environmentally-extended input-output (EEIO) factor set, which gives a kgCO2e-per-dollar figure for each industry sector. You match each spend category to the closest EEIO sector, multiply, and you have an estimate.

A worked line: you spent $180,000 with a steel supplier, mapped to category 1. If the relevant EEIO factor is around 1.1 kgCO2e per dollar, that line is roughly 198 tonnes of CO2e. Do that for every line, sum by category, and the totals show you where your footprint concentrates. Note the honest caveat: because the factor is an industry average, two suppliers in the same sector get the same intensity even if one runs on renewables and one on coal. Spend-based numbers tell you the shape of the problem, not the precise size.

Step 4: Know when spend is not good enough

The spend-based method is a screening tool, and its whole purpose is to work out where to stop using it. Once you know that, say, category 1 is 60 percent of your footprint and driven by three suppliers, you graduate those to activity data: ask the suppliers for their actual product-level emissions, or for the physical quantities you can pair with better factors. You leave the long tail of small, immaterial categories on the spend estimate, because refining them would cost more than it changes the number.

This is also what regulators expect. California's SB 253 and Europe's CSRD both accept spend-based estimation in early years and expect you to improve the material categories over time. What they do not accept is a number nobody can reproduce. The full menu of US obligations that this inventory feeds is on climate disclosure software.

Doing it without the spreadsheet

The math here is trivial. The work is the classification: ten thousand lines, each needing a category, a factor and a recorded reason, kept linked to the source invoice so the number survives a review. That is exactly the pass our carbon accounting software automates: it reads the AP export, drafts a category and factor for every line with a confidence score, and routes only the low-confidence tail to a human. Every posting keeps its link back to the invoice, so when an auditor samples a figure, the answer is a click. You can run that classification on your own lines in the demo, and read the plain-language Scope 3 definition if you want the boundaries first.

Emission factors and figures here are illustrative of the spend-based method; use current EPA EEIO or equivalent factors for your own calculation, and treat spend-based results as a screening estimate rather than an audit-grade inventory. Regulatory references describe rules as published in July 2026.