Scope 3 emissions: How to build a credible program.

Scope 3 emissions represent the vast majority of most companies' carbon footprints. A consumer goods manufacturer may find that 85% of emissions come from purchased materials and packaging. A financial institution's financed emissions typically dwarf those of its direct operations; CDP has reported that portfolio (financed) emissions average roughly 700 times a financial institution's operational emissions. Professional services firms typically have far higher Scope 3 (business travel, purchased services, commuting) than Scope 1 and 2.

Ignoring Scope 3 means ignoring the majority of your climate impact. It also means missing strategic risks: supply chain carbon costs (CBAM), customer decarbonization requirements, and investor scrutiny of value chain exposure.

Regulators are increasingly mandating Scope 3. California SB 253 requires it for all covered entities. CSRD requires it for material categories. Under SBTi's V1.3.1 rules, companies must set Scope 3 targets when Scope 3 exceeds 40% of total emissions (nearly always); SBTi's V2.0 standard (June 2026) instead requires targets for categories representing 5% or more of Scope 3 (see our SBTi guide). CDP's climate questionnaire asks for detailed Scope 3 data. The direction is clear: Scope 3 measurement is becoming table stakes.

The GHG Protocol divides Scope 3 into 15 categories across upstream and downstream activities.

(1) Purchased goods and services: emissions from production of everything the company buys, often the largest category.

(2) Capital goods: emissions from production of purchased capital equipment.

(3) Fuel- and energy-related activities not in Scope 1 or 2: upstream emissions of purchased fuels and electricity (extraction, refining, transmission losses).

(4) Upstream transportation and distribution: third-party logistics for inbound materials.

(5) Waste generated in operations: disposal and treatment of waste from the company's facilities.

(6) Business travel: flights, hotels, rental cars.

(7) Employee commuting: emissions from employees traveling to and from work.

(8) Upstream leased assets: emissions from leased assets not in Scope 1 or 2.

(9) Downstream transportation and distribution: third-party logistics for products sold.

(10) Processing of sold products: emissions from further processing by downstream companies.

(11) Use of sold products: emissions from customers using the product (for example, energy consumption during product life).

(12) End-of-life treatment of sold products: disposal, recycling, incineration.

(13) Downstream leased assets: emissions from assets leased to others.

(14) Franchises: emissions from franchise operations.

(15) Investments: emissions from equity and debt investments (critical for financial institutions, measured using PCAF).

Three primary approaches exist for measuring Scope 3, each trading accuracy for feasibility.

Spend-based estimation uses Environmentally Extended Input-Output (EEIO) models (like USEEIO or EXIOBASE) to estimate emissions from procurement spend. Modern platforms like Gravity apply BEA industry-specific inflation adjustment and foreign exchange conversion automatically so that spend data and emission factors are compared in consistent dollar years. AI-powered factor inference can map spend descriptions to EEIO factors in bulk, dramatically accelerating Category 1 measurement. Advantages: easy to implement with existing procurement data, provides full coverage quickly. Limitations: assumes industry-average emission intensity, sensitive to price fluctuations, cannot distinguish between high- and low-carbon suppliers.

Average-data methods use industry-average emission factors applied to physical activity data (for example, kg CO₂e per tonne of steel, per kg of aluminum). More accurate than spend-based but requires knowledge of the specific materials and quantities purchased, not just the dollar amounts.

Supplier-specific primary data is the gold standard: actual emissions data collected directly from suppliers through engagement programs. Most accurate and most actionable, but requires active engagement with hundreds or thousands of suppliers and depends on their measurement capabilities.

Collecting primary data from suppliers requires a structured engagement program. Best practices include:

Prioritize by materiality. Typically 20% of suppliers represent 80% of supply chain emissions. Focus primary data collection on the highest-emitting categories and suppliers first.

Make it easy – reduce surveys. Start with AI Enrichment: scanning the web to pull in what they’ve already disclosed publicly. Next, provide clear, standardized data request templates only where required. Offer multiple response formats (online forms, spreadsheet uploads, integration with supplier sustainability platforms). Reduce friction at every step.

Provide incentives. Preferred supplier status, longer contract terms, and procurement scoring that weights sustainability data provision can motivate supplier participation.

Build capability. Many suppliers, especially smaller ones, may not have their own carbon accounting programs. Providing guidance materials, calculation tools, or connections to affordable measurement services builds the ecosystem.

Track data quality over time. Measure the percentage of Scope 3 emissions covered by primary data versus estimates. Set improvement targets (for example, 50% primary data coverage within 2 years).

Gravity's supply chain module automates data request workflows, processes supplier responses using AI, and tracks data quality progression from spend-based to primary data over time.

Under SBTi's V1.3.1 criteria, companies must set Scope 3 targets when Scope 3 emissions exceed 40% of total emissions, which applies to nearly all companies, and those targets must cover at least two-thirds of total Scope 3 emissions. SBTi's V2.0 standard (published June 2026) changes this approach, requiring targets for categories that each represent 5% or more of Scope 3 emissions. Both rulesets are currently valid — V1.3.1 remains available for submissions through January 31, 2028 — so confirm which applies to your submission. See our SBTi guide for the full V1-to-V2.0 transition.

The target-setting approaches include absolute contraction (reducing total Scope 3 emissions by a defined percentage) and supplier or customer engagement targets (requiring a percentage of suppliers/customers to set their own science-based targets within a defined timeframe).

SBTi's FLAG (Forest, Land, and Agriculture) Guidance adds sector-specific requirements for companies with significant land-use emissions. The Financial Institutions framework covers financed emissions under Category 15.

Meeting Scope 3 targets requires continuous measurement, not one-time estimation. Organizations need reliable base-year inventories that can be recalculated when structural changes occur, annual inventory updates showing progress, and the ability to attribute reductions to specific interventions (supplier engagement, material substitution, logistics optimization).

Data availability is the number one challenge. Many suppliers do not track their own emissions, and spend-based estimates have wide uncertainty ranges. Solution: Start with what you have. Spend-based estimates provide directional guidance. Improve iteratively.

Category boundaries can be confusing. Some emissions could be classified under multiple categories (for example, upstream logistics could be Category 1 or Category 4 depending on contractual arrangements). Solution: Document boundary decisions transparently and apply them consistently.

Double counting between categories (for example, fuel used in upstream logistics counted in both Category 1 and Category 4) requires careful boundary management. The GHG Protocol provides guidance on avoiding double counting.

Scope 3 increases apparent total emissions, which can be uncomfortable to disclose. However, transparency about the full footprint is increasingly expected and rewarded by stakeholders. SBTi validation provides a credible framework for showing that high Scope 3 numbers are being addressed.

Year-over-year changes can be driven by methodology improvements rather than actual emission changes, making trend analysis difficult. Solution: Separate methodology-driven changes from real changes in reporting, and recalculate the base year when methodologies change.