Spend-Based vs Activity-Based Emissions: A Decision Framework

Every mid-market carbon accounting conversation eventually hits the same question. Do we use spend-based factors, or do we chase activity data? The wrong answer, in both directions, costs quarters of engineering and finance time. The right answer is almost always both, in specific ratios, with a documented transition path.

Here is the framework for choosing.

What is the difference between spend-based and activity-based emissions?

Spend-based emissions multiply a dollar amount by an emission factor expressed as kg CO2e per dollar. Activity-based emissions multiply a physical quantity (kWh, miles flown, kg of steel) by a factor expressed as kg CO2e per unit of that activity.

Both are GHG Protocol Scope 3 methods. Both are used in CDP-verified reports. Both pass limited assurance when documented correctly. The difference is precision, not legitimacy.

• Spend-based answers the question "roughly, what does our spending in this category emit?" using sector averages.
• Activity-based answers the question "specifically, what did our actions in this category emit?" using primary or product-specific data.

Neither is a downgrade. They serve different purposes at different maturity stages.

When should you use spend-based emissions?

Spend-based works when at least one of these three conditions holds.

• The category is immaterial. Under 5% of your total footprint. Spending weeks refining a category that will not move the total is theater, not science.
• Activity data does not exist. Your suppliers have not published product carbon footprints, your commuting patterns are hybrid and unstable, or your capital goods are consumed once and never traced.
• You are building a first baseline. A spend-based inventory across all 15 Scope 3 categories in two weeks is more valuable than a perfect activity-based inventory of three categories in six months.

For most mid-market companies with their first inventory, 60 to 80% of the footprint starts spend-based. That is normal.

When should you use activity-based emissions?

Upgrade to activity-based when three conditions hold together.

1. The category is material. Above 10% of your total footprint, or singled out by a customer or regulator.
2. Activity data exists in a system you already run. Cloud provider consoles, travel platform exports, utility bill kWh, supplier questionnaires you can reasonably request.
3. The upgrade reduces your reporting risk more than it costs. Enterprise customers asking pointed questions, CSRD assurance in the next 12 months, or a public reduction target that needs a defensible baseline.

Cloud compute is the clearest example. If you spend $2M a year on AWS, spend-based estimates put that at roughly 200 to 400 tCO2e depending on the factor. AWS's own footprint tool reports region-specific numbers that are often 50 to 80% lower. The activity-based number is defensible, cheaper, and closer to reality. Upgrade it on day one.

How do you decide category by category?

Use a two-axis rubric. Materiality on one axis, data availability on the other.

The rubric prevents two failure modes. First, chasing activity data for immaterial categories. Second, defaulting to spend-based when the activity data is right there in a system you already pay for.

What does a hybrid Scope 3 inventory look like in practice?

A typical mid-market SaaS company inventory ends up structured like this.

• Category 1: purchased goods and services. Top 10 suppliers by spend, primary supplier data. Next 50 by spend, spend-based with sector factors. Long tail, blended factor.
• Category 2: capital goods. Spend-based unless you buy hardware directly.
• Category 3: fuel and energy activities. Activity-based, derived from Scope 1 and 2 data.
• Category 6: business travel. Activity-based from travel platform data.
• Category 7: employee commuting. Activity-based from annual survey.
• Cloud compute (inside Cat 1). Activity-based from cloud provider dashboards.
• Categories 4, 5, 8, 9, 10, 12, 13, 14, 15. Documented as immaterial or not applicable with calculation.

The report shows the method used per category. This is not a hack; this is how the standard is meant to be applied.

How do you transition without restating everything?

Restatement is the friction most teams fear. Here is how to handle it cleanly.

• Restate the baseline year. If you have a public reduction target tied to a baseline (e.g., "50% reduction by 2030 from 2024 baseline"), restate the baseline every time you upgrade a material category. Otherwise your progress claims are unfalsifiable.
• Do not restate intermediate years unless material. If the change is under 5% of the total footprint, note the change and move on.
• Document every method change. In the methodology appendix: category, old method, new method, reason, size of impact. Assurance reviewers care about the audit trail, not the direction of the change.
• Never silently switch mid-year. If you upgrade a category in Q3, use the new method for the full year, not partial.

The mistake is treating restatement as a failure. Restatement done well is a signal of maturity. Silent revisions are the signal that gets flagged.

What emission factor datasets should you standardize on?

Four datasets cover most mid-market needs.

• EPA USEEIO. Free, US-focused, sector-average. Good default for spend-based US emissions.
• DEFRA / BEIS. UK government factors, extensive coverage of fuels, travel, and materials. Global reach.
• Exiobase. Global multi-regional input-output model. Good for global purchased goods.
• Ecoinvent. Higher-resolution product-level factors. Paid. Use for activity-based estimates on physical goods.

Pick one primary source per category. Document the version. Only switch when your assurance provider requests it or when a new version corrects a factor you rely on materially.

The mistake to avoid

The trap is treating spend-based versus activity-based as a purity contest. Companies that hold out for "real data" spend two years building inventory infrastructure while their customers move on to competitors with a defensible number today. Companies that go all spend-based and never upgrade wonder why their reduction claims never move the number. Both are wrong. The correct posture is a hybrid inventory that starts spend-based on day one, upgrades material categories with data that already exists, and documents every choice so the auditor never has to ask.