Carbon Reporting

Spend-Based vs Activity-Based Carbon Accounting: Which Is Right?

January 13, 2026· 23 min read· by AI Sustainable Future Team

The spend-based versus activity-based question is one of the first method choices in any carbon inventory. The usual answer, "activity-based is more accurate, spend-based is cheaper, use activity data when you can", is too blunt to act on. The GHG Protocol gives more specific guidance: prioritise the categories that matter most, then choose data for each one based on your goals, the category's size and the data you can actually get. This guide walks through how to make that call category by category, and what the numerical difference looks like in an illustrative example.

Short definitions first. Activity-based accounting multiplies a physical quantity (kilowatt-hours, gallons of fuel, kilograms of refrigerant, passenger-miles flown) by a per-unit emission factor. In the US, a common source is the EPA GHG Emission Factors Hub, whose 2025 edition also covers freight, business travel, commuting and waste; the UK government conversion factors are a common alternative. Spend-based accounting multiplies the amount paid to a vendor by a sector-average factor per dollar. For US purchases, a widely used source is EPA's Supply Chain Greenhouse Gas Emission Factors v1.3 (July 2024, 1,016 commodities, 2022 dollars), built from the USEEIO model; the Cornerstone Sustainability Data Initiative published a v1.4 update in October 2025. Both methods appear in the GHG Protocol's Technical Guidance for Calculating Scope 3 Emissions. Neither is right for every category.

The operational question you are actually answering is: for this Scope 3 category, in this reporting period, given the data I have and the category's share of my footprint, which method can I defend to a reviewer, and is the data reasonable to collect? Answer it once for each of the 15 categories and you have your methodology.

What the GHG Protocol says about choosing data

The Corporate Value Chain (Scope 3) Standard sorts data into primary data, from specific activities in your value chain (a supplier's own emissions, your own fuel records), and secondary data: industry averages, published databases, financial data and proxies. Section 7.1 asks you to focus data collection on the activities with the largest expected emissions, the biggest reduction opportunities and the most relevance to your goals. Section 7.3 then sets out how to choose:

Any combination is allowed. You may mix primary and secondary data within the inventory.
Primary data where it matters. Collect high-quality primary data for high-priority activities, and use supplier data for activities where you want to track reductions.
Secondary data elsewhere. It fits activities you did not prioritise, activities where primary data is not available, and cases where the secondary data is of higher quality.
Disclose what you used. You must report the types and sources of data, including emission factors and GWP values, and the percentage of emissions calculated from supplier data.

For Category 1 (purchased goods and services), the Technical Guidance describes four methods, from most to least specific to the supplier:

Supplier-specific method: the supplier gives you a cradle-to-gate figure for the product it sold you. Example: your print vendor reports the footprint of your job.
Hybrid method: the supplier gives you its own Scope 1 and 2 emissions allocated to your purchase, and you add its upstream emissions from its activity data or from secondary factors.
Average-data method: you know the physical quantity bought (kilograms, units) and apply an industry-average cradle-to-gate factor per unit.
Spend-based method: you multiply the amount spent by an EEIO factor per dollar. The guidance says to use it when the other three are not feasible, and it is the usual screening method.

Two caveats. Box 1.1 warns that a more specific method does not guarantee a more accurate result: supplier data can be less accurate than an industry average, depending on its sources and how emissions are allocated to your purchase. And you can use different methods for different purchases within one category. So the guidance sets a direction: use more specific data where the category matters and the data can be obtained, and record why. A reviewer who sees spend-based factors where you held activity data in your own records (fleet fuel, for example) will ask why; spend-based factors on a small category, with a documented reason, raise little.

Which method fits, category by category

The 15 Scope 3 categories fall into three rough groups: those where activity data is usually at hand, those where spend-based is the realistic default for a small company, and hybrids.

Categories where activity data is usually available

Category 3, Fuel- and Energy-Related Activities: upstream emissions of the fuel you burn and the electricity you buy, plus transmission and distribution losses. You already hold the fuel volumes and kWh from Scope 1 and 2, so apply upstream and loss factors to them. The EPA Hub publishes grid gross loss percentages by eGRID subregion.
Category 4, Upstream Transportation and Distribution: third-party freight you paid for. If freight bills show weight and distance, use the distance-based method (tonne-miles or tonne-kilometres times a factor per mode). If they do not, the Technical Guidance allows the spend-based method.
Category 6, Business Travel: a travel-booking export (Concur, Navan or similar) holds flight segments and distances, so apply distance-based factors by flight length. The Technical Guidance treats spend-based as the fallback when distance or fuel data is not available.
Category 7, Employee Commuting: the Technical Guidance lists fuel-based, distance-based and average-data methods, usually fed by an employee survey. Your ledger only shows commuting the company pays for, such as transit passes or mileage reimbursements, so a spend-based figure covers part of the category at most.
Category 11, Use of Sold Products: for companies that sell energy-using products, units sold × lifetime energy use × grid factor. For those companies it can be the largest category in the inventory.

Categories where spend-based is the realistic default

Category 2, Capital Goods: the year's additions to property, plant and equipment. The Technical Guidance offers the same methods as Category 1, but getting cradle-to-gate data from the maker of every asset is rarely practical for a small company, so spend-based with EEIO factors is the usual route.
Category 5, Waste Generated in Operations: the Technical Guidance's methods all start from tonnes of waste by type and treatment. If your hauler reports tonnage, use it; with only invoices, a spend-based estimate is a rough screen. The EPA v1.3 factor for waste management services is high, 0.988 kg CO₂e per dollar, so a modest waste bill can still matter.
Category 8, Upstream Leased Assets: leased assets not already in your Scope 1 and 2; usually small for office-based companies. When the landlord bills a flat rent that includes utilities, the Technical Guidance's average-data method estimates emissions from floor space by building type, and the EIA's Commercial Buildings Energy Consumption Survey publishes US energy use per square foot by building type.

Hybrid categories (supplier data for top vendors, spend-based for the tail)

Category 1, Purchased Goods and Services: often the largest category for service firms, and the one most worth getting right. Year 1: spend-based for every vendor is a reasonable screen. Year 2 onward: collect supplier data from the vendors that carry the most estimated emissions, and keep spend-based for the long tail. This is where the progression described below matters.
Category 9, Downstream Transportation and Distribution: if carriers give you weight and distance, use the distance-based method for the main carriers. Otherwise use spend-based.
Categories 10, 12, 13, 14 and 15: many small service companies find these not relevant. The Scope 3 Standard requires you to account for all Scope 3 emissions and to disclose and justify any exclusion, so write down the reason for each.

Illustrative example (fictional company): Category 1 calculated two ways

A fictional 85-person architecture firm has $3.8M of non-utility vendor spend in one fiscal year, across 2,847 AP line items. Below, its Category 1 is calculated spend-based throughout, then with supplier data for the top 10 vendors. The factors are EPA v1.3 values with margins, in kg CO₂e per 2022 dollar, applied to spend as stated to keep the arithmetic readable. In a real inventory you first restate spend in 2022 dollars, which lowers the result somewhat for later years.

Path A: spend-based for everything

Each line is classified to a NAICS commodity and multiplied by its factor.

Three architecture subcontractors ($1.2M): Architectural Services (NAICS 541310), 0.103 kg CO₂e per dollar. Subtotal: 123.6 tCO₂e.
Cloud hosting ($200K): Data Processing, Hosting, and Related Services (518210), 0.093. Subtotal: 18.6 tCO₂e.
Commercial printer ($150K): Commercial Printing (323111), 0.236. Subtotal: 35.4 tCO₂e.
Five other top vendors ($950K): IT services, $300K at 0.089 (26.7 t); software subscriptions, $150K at 0.080 (12.0 t); legal, $200K at 0.041 (8.2 t); engineering consultants, $200K at 0.103 (20.6 t); janitorial services, $100K at 0.214 (21.4 t). Subtotal: 88.9 tCO₂e.
Long tail (a few hundred vendors, $1.3M): about 80 commodity codes, averaging roughly 0.14 kg CO₂e per dollar. Subtotal: 182.0 tCO₂e.
Spend-based Category 1 total: 448.5 tCO₂e.

Path B: supplier data for the top vendors, spend-based for the rest

The top 10 vendors account for $2.5M of the $3.8M (about two thirds). The firm sends each a short data request. In this illustrative example, the responses are:

Architecture subcontractors ($1.2M): two of the three ($800K) send their Scope 1 and 2 totals, their own Scope 3 Category 1 (what they buy), and revenue for the same year. Allocated by the firm's share of their revenue, that is 0.07 kg CO₂e per dollar: $800K × 0.07 = 56.0 t. The third ($400K) does not respond and stays at the EEIO factor: $400K × 0.103 = 41.2 t. Subtotal: 97.2 tCO₂e (against 123.6 t spend-based).
Cloud hosting ($200K): the provider sends an account-level emissions report for the firm's usage: 10.0 tCO₂e (against 18.6 t). Check which scopes a provider's figure includes before you swap it in.
Commercial printer ($150K): the printer reports the paper used on the firm's jobs (30 tonnes) and its press energy. A cradle-to-gate paper factor from a cited database, plus the printer's allocated Scope 1 and 2, gives 40.0 tCO₂e (against 35.4 t). This supplier sits above its sector average.
Five other top vendors ($950K): no primary data; they stay spend-based. Subtotal: 88.9 tCO₂e.
Long tail ($1.3M): spend-based as before. Subtotal: 182.0 tCO₂e.
Hybrid Category 1 total: 418.1 tCO₂e, about 7% lower than the spend-based 448.5 tCO₂e.

What the comparison shows

The total moved by 30.4 tCO₂e. That is inside the uncertainty of a spend-based screen: our engineering estimate for spend-based screening is roughly ±25-40% at the category level (the GHG Protocol publishes no figure).

The direction differed by supplier. The two responding architects and the hosting provider came in below their sector averages, and the printer came in above. A sector-average factor can be too high or too low for any given supplier.

The larger gain is information the spend-based run cannot give: which suppliers sit above or below their sector average, and a supplier-level baseline to track reductions against. Table 7.5 of the Scope 3 Standard lists these as the advantages of primary data. For disclosure, record the share of the total calculated from supplier data (here about 25%), which the Standard asks you to report.

When spend-based is the right answer, and when it is a liability

Spend-based is appropriate when:

It is your first year of reporting. You do not yet have supplier relationships or activity data, and you need a screening estimate to find out which categories matter.
The category is a small share of the total. Spending 40 hours to move a $10K waste bill from spend-based to tonnage data may shift a few tonnes in an inventory of several hundred. Leave it spend-based, document the reason and move on.
Primary data is not reasonably available. The vendor does not respond, the activity data does not exist in usable form, or collecting it would cost more than the category's size justifies.
You are screening for relevance. The Scope 3 Standard (Box 7.1) describes EEIO data as particularly useful for the initial screen that decides which categories get more effort.

Spend-based becomes a liability when:

The category is a large share of your footprint and you have made no attempt at supplier data. If Category 1 is 70% of your total and is still 100% spend-based in year two, a reviewer will read that as a data program that has not started.
Suppliers in the sector differ widely. An EEIO factor is a sector average and cannot distinguish between suppliers in the same sector (Scope 3 Standard, Table 7.5). Two data-centre operators on different electricity grids, for example, get the same factor per dollar.
You are seeking third-party assurance. The assurance provider will look at whether the method suits each material category, and a material category that stays fully spend-based after year one invites questions.
The number has to drive a decision. If Category 1 feeds a reduction target or a procurement choice, a sector average cannot tell you which supplier to prioritise. A spend-based figure also moves with what you pay: a price rise looks like an emissions increase, and a supplier's real reductions never show. Section 7.3 of the Scope 3 Standard recommends supplier data for this case.

A three-year data maturity progression

A carbon program does not go from zero to full supplier data in one year. Here is a sequence that suits many small companies.

Year 1, screening

Scope 1 and 2 calculated from activity data (fuel volumes, kWh, refrigerant) with current factors, such as the EPA Hub and eGRID in the US. Scope 3 screened with spend-based factors where you have spend, and simple estimates where you do not, to find the relevant categories; for a services firm these are often Categories 1, 6 and 7. Each excluded category gets a one-line reason. Category 1 is usually 100% spend-based in year one, which is a normal starting point.

SB 253 note: California's SB 253 applies to entities with more than $1bn in revenue doing business in California. The first reports, Scope 1 and 2 only, are due 10 November 2026 (CARB). Scope 3 reporting starts in 2027, and the statute allows secondary data, including spend-based and industry-average data, for Scope 3. The draft rules CARB previewed at its July 2026 workshop require limited assurance for Scope 1 and 2 only. If you supply a company in scope, Scope 3 data requests may follow.

Year 2, top-vendor engagement

Categories 6 and 7 refined with booking data and a better commuting survey. Category 1 starts its hybrid transition: pick the vendors with the largest estimated emissions, send each a data request, accept whatever comes back, and keep spend-based factors for non-responders. In the example above, that moved about 30% of Category 1 spend onto supplier data. The total can move in either direction.

Year 3, supplier data built into procurement

Data requests become part of vendor onboarding: new vendors above a spend threshold are asked for emissions data at setup. Supplier-data coverage of Category 1 rises, other relevant categories are built out, and the relevance screen is refreshed. The inventory now has a documented method per category, a known share of supplier data and a base year.

One caution: when a category moves from spend-based to supplier data, the year-over-year change mixes real emission changes with the change of method. The Scope 3 Standard (Chapter 9) requires you to recalculate base-year emissions when method changes or better data have a significant effect. If your figure falls because supplier data came in below the sector average, brief the CFO before the report goes out: the drop reflects better data, and actual emissions may not have changed.

Vendor engagement playbook for getting primary data

The choke point for moving Category 1 onto supplier data is vendor response. Response rates to first-time requests are often low, so plan from the start to default non-responders to a cited spend-based factor and to record that choice.

Who to ask

Start with the vendors that carry the most estimated emissions, which is not always the list with the most spend; the Scope 3 Standard (section 7.1) cautions that spend may not correlate well with emissions. In the example, the printer ($150K) carries 35.4 t and the law firm ($200K) 8.2 t: less spend, more than four times the emissions. Rank vendors by spend × factor, take the top ten or so, and do not spend hours chasing vendor #47.

What to ask for (in this order of preference)

A cradle-to-gate footprint for the products or services they sold you, if they have one (the supplier-specific method).
Their Scope 1 and 2 emissions, their upstream Scope 3 for purchased goods and services, and their revenue, for the same period. Allocate by your share of their revenue: (their emissions ÷ their revenue) × what you paid them. If they give only Scope 1 and 2, add their upstream emissions from secondary data (the GHG Protocol's hybrid method), since Scope 1 and 2 alone understate a cradle-to-gate figure.
Their published sustainability report or CDP response, from which you can take the same figures.
A commitment to provide data next year if they cannot this year. That sets up your year-two baseline.

A short request email

Subject: Carbon data request, [your company name] FY[year] emissions inventory. Three short paragraphs: (1) we are building our carbon inventory, and your supply to us is one input; (2) the data points we need (listed) and the format; (3) the deadline (30 days out), and a note that without their figures we will use a sector-average factor, which may not reflect their actual performance. That last point gives suppliers below their sector average a reason to reply.

What to do when they do not respond

Default to the spend-based factor, document it, and add the vendor to next year's outreach list with a note. Do not hold the whole report for one vendor. Partial supplier-data coverage with a documented rationale is far more useful than no report.

Where Carbon Draft fits in this progression

Carbon Draft is built for the year-one screening step. You upload one accounting or bank export (CSV or Excel from QuickBooks, Xero, NetSuite, Sage, SAP or Tally, or a DATEV, SIE or XAF file) and, optionally, your kWh, fuel volumes, gas and refrigerant. The engine classifies the ledger lines and applies cited factors, including EPA v1.3 for US purchases (adjusted to each purchase's year with the BEA GDP deflator), eGRID 2023 for US electricity and the EPA Emission Factors Hub 2025 for US fuels. Scope 3 covers up to 8 of the 15 categories: 1, 2, 4, 5, 6 and 7 where your ledger shows that spend, 3 derived from your electricity, and 12 when you enter end-of-life tonnages. The report states which categories it does not cover, so you can screen the rest.

Basic, $20: Scope 1 and 2.
Complete, $49: adds Scope 3.
Disclosure-Aligned Report, $299: adds the full priced ledger (every transaction with its factor), per-row confidence, an Excel calculation workbook, and a framework section for GHG Protocol, CDP, CSRD (ESRS E1) and California SB 253 that sets out what each requires, what the report contains and the gap. Scope 3 coverage is the same as in Complete.

Every report is a screening-grade inventory marked not assured: a year-one baseline a reviewer or assurance provider can start from, which you replace with supplier data in years two and three. See pricing and the methodology for factor sources.

If you are starting from zero this year, upload your accounting export at aisustainablefuture.com/carbon-draft and get a cited, screening-grade carbon report in minutes. Scope 1 and 2 from $20, Scope 3 from $49.

If you are a CPA firm building this as a client service, the workflow is in our CPA's Guide to GHG Protocol Scope 1, 2 and 3, and partner report packs are on the for accountants page.

If you are a small supplier facing a customer request, our guide to what climate disclosure rules mean for small business suppliers covers where to start.