Scope 3 Emissions Are 80% of Your Footprint and the Hardest to Measure

Every sustainability director I have talked to in the past two years eventually says the same thing: "We sorted out Scope 1 and 2. Then we looked at Scope 3 and realized we had barely scratched the surface." It is a humbling moment, and it happens to companies that have already invested heavily in their emissions programs.

The numbers back this up. Across most industries, Scope 3 represents between 70% and 90% of a company's total GHG footprint. For consumer goods companies, the figure often exceeds 95%. For financial institutions, it can be even higher once financed emissions (Category 15) are included. You can switch to 100% renewable electricity, electrify your entire fleet, and optimize every facility you own — and still leave the vast majority of your climate impact unaddressed.

Why Scope 3 Is Structurally Different

Scope 1 and 2 emissions are yours to control. You burn the fuel. You buy the electricity. You have utility bills, fleet records, and fuel receipts. Measurement is a data management problem, not an estimation problem.

Scope 3 is different because the emissions happen outside your walls, across a supply chain you do not own and often do not fully see. The GHG Protocol identifies 15 categories within Scope 3, ranging from purchased goods and services (Category 1) to investments (Category 15). Each category requires a different data collection approach, different emission factors, and different assumptions when primary data is unavailable.

Category 1 alone — purchased goods and services — is the largest source for most manufacturers and retailers, and it requires either supplier-specific data (which almost no one has at scale), spend-based estimation using average economic input-output factors, or a hybrid approach. None of these is precise. All of them require judgment calls that need to be disclosed and defensible.

The Three Methods and Their Honest Limitations

The GHG Protocol Corporate Value Chain Standard describes three main calculation approaches for upstream Scope 3 categories:

Spend-based: Multiply your supplier payments by an industry-average emission intensity factor (kg CO2e per dollar spent). Fast to implement, easy to explain to boards. The problem is that two suppliers in the same industry category can have carbon intensities that differ by a factor of five or more. You are averaging over that variation, which produces a number that may be directionally correct but will not survive supplier-level scrutiny.

Supplier-specific (primary data): Get actual emissions data from each supplier, ideally verified by a third party. This is the gold standard. It is also impractical for most companies at scale — the typical mid-market manufacturer has 200 to 500 direct suppliers, and asking each of them for audited GHG data requires a supplier engagement program that takes years to build.

Hybrid: Use primary data where you can get it (usually top suppliers by spend), and fall back to spend-based or average-data methods for the long tail. This is what most credible reporters are actually doing, and it is the approach the CDP and CSRD frameworks accommodate when primary data is not available.

Where Companies Get Stuck

The most common failure mode is analysis paralysis. Companies spend 18 months trying to decide which method to use before collecting any data. Meanwhile, their CSRD disclosure deadline is approaching, investors are asking for Scope 3 estimates, and the internal team is arguing about whether a spend-based number is "misleading."

The better approach: start with spend-based for a complete baseline, even if imperfect. A published, documented estimate with clear methodological limitations is more useful than a gap in your disclosure. It also gives you a reference point for prioritizing which suppliers to engage first. Concentrate your supplier engagement budget on the 20 to 30 suppliers that account for 70 to 80% of your Category 1 spend. That is where data quality improvements will have the most impact on your reported total.

Category 11 (use of sold products) is the other category where companies frequently get stuck. For any company selling physical products, the emissions generated during product use often exceed everything upstream. An appliance manufacturer that ignores Category 11 is presenting a fundamentally incomplete picture of its climate impact. Calculating it requires product usage data — energy consumption during operation, expected product lifetime, regional electricity grid factors — which demands coordination between sustainability, product engineering, and sales teams who rarely work together.

What Good Looks Like in Practice

A specialty chemicals company we worked with had Scope 1 and 2 emissions totaling approximately 48,000 tonnes CO2e. Their initial Scope 3 estimate, using spend-based methods for Category 1 and activity data for Categories 3, 4, and 9, came in at 340,000 tonnes — roughly seven times larger. The exercise forced a complete reorientation of their decarbonization strategy. Switching to green energy for manufacturing, which had been the headline initiative, would reduce total footprint by less than 5%. The actual levers were supplier selection and product reformulation.

That reorientation is uncomfortable. It means engaging suppliers who have no idea what a Scope 3 questionnaire is, renegotiating contracts, and in some cases replacing established vendors. But it is also the only path to material emissions reductions for most companies.

Where Reporting Requirements Are Going

Both the SEC's climate disclosure rule and the CSRD require Scope 3 disclosure when material. The SEC rule, following its partial stay by the courts, still requires large accelerated filers to begin Scope 3 reporting on a phased timeline. CSRD requires Scope 3 for all large companies with EU operations starting with financial year 2024, reported in 2025. The direction is clear: Scope 3 is becoming a compliance requirement, not just a voluntary best practice.

Companies that have been deferring Scope 3 work because "the methodology is too uncertain" are running out of runway. Imperfect data with transparent methodology is accepted. No data is not.