Scope 3 Category 1: A Practical Guide

A procurement analyst opens an ERP export containing 4,000 purchase-order lines and asks a simple question: which of these belong in Scope 3 Category 1? The spreadsheet contains raw materials, software renewals, packaging, freight, contract manufacturing, office supplies, and services. Some lines have a usable SKU. Others contain free-text descriptions that change every time a buyer raises an order.

That's the operational reality of Scope 3 Category 1, purchased goods and services. The accounting definition is established, but producing a credible inventory requires more than assigning an emission factor to total procurement spend. You need a durable product identity, a defensible calculation method, supplier evidence, clear boundaries, and records that can be refreshed without rebuilding the inventory from scratch.

• What Scope 3 Category 1 Actually Covers
• The boundary follows the purchase and physical flow
• Choosing the Right Calculation Method
• Mapping Purchases to Product Identifiers
• Resolve the unit before applying the factor
• Three Worked Examples for the Same Product
• Common Mistakes That Distort Category 1 Numbers
• Shortcuts that undercount
• Building a Repeatable Supplier Data Workflow
• Evidence Governance and Audit Readiness
• A 90 Day Plan to Operationalize Category 1
• Days 1 to 30 focus on scope and identity
• Days 31 to 60 compare the methods
• Days 61 to 90 lock the workflow

The GHG Protocol defines Category 1 as all upstream, cradle-to-gate emissions from products and services purchased or acquired during the reporting year, covering both tangible goods and intangible services. The category includes emissions from extraction, production, and upstream processing before the reporting company takes control of the input. The GHG Protocol Category 1 guidance also makes clear that the category excludes the other upstream Scope 3 categories.

That boundary resolves much of the analyst's spreadsheet problem. A purchased fabric, component, carton, finished product, cloud service, or professional service can belong in Category 1. A machine bought for the factory belongs in Category 2, capital goods. Purchased fuel and electricity-related activities belong in Category 3, while inbound transport that the reporting company treats as a separate logistics activity belongs in Category 4. Leased space falls under the relevant leased-asset category, not Category 1.

!A diagram explaining Scope 3 Category 1 emissions covering purchased goods and services using ERP and PO data.

Category 1 commonly contains:

• Product inputs: raw materials, ingredients, components, sub-assemblies, and co-manufactured goods.
• Product presentation: packaging, labels, inserts, and other materials purchased for sale or delivery.
• Indirect purchased inputs: dyes, detergents, softeners, maintenance consumables, and similar materials that may not appear in a bill of materials.
• Services: professional services, cloud computing, software licences, and other procured services.

Services still count even when the supplier delivers no physical item. A consulting engagement, hosted software subscription, or cloud contract is a purchased service, so its upstream emissions are assessed within Category 1 unless another Scope 3 category provides the appropriate boundary.

The practical boundary rule is straightforward: goods and services bought and consumed in the fiscal year belong in the inventory, whether procurement records them as direct spend or as an operating expense. Cost classification helps finance teams, but it doesn't determine the Scope 3 category.

Practical rule: Classify each purchase by what was acquired and how its upstream emissions flow into the business, not by the account code alone.

The calculation method determines how much of your Category 1 result reflects actual supplier performance and how much reflects a broad proxy. The GHG Protocol's Scope 3 calculation guidance recognises a hierarchy: supplier-specific or product-level data is strongest, activity or average-data methods are intermediate, and spend-based EEIO methods are weakest.

Supplier-specific data can include a product carbon footprint, a verified LCA, or a supplier's facility and product information that supports a calculated result. It takes effort to review, but it gives procurement teams something they can act on. If a supplier changes material inputs, energy sources, yield, or production location, the record can reflect that change rather than leaving the business tied to a sector average.

Average-data methods use an emission factor for a material, process, product type, or industry sector. They're useful when a supplier can identify what it sells but can't provide a product-level footprint. The result is usually more representative than a pure spend estimate, especially when you have physical quantities such as kilograms, litres, metres, or units.

Spend-based EEIO is a practical fallback. It converts monetary spend into emissions using broad economic-sector averages. That makes it fast and useful for fragmented tail spend, marketing, general professional services, and early-stage inventories. It also means that price changes, inflation, currency treatment, and supplier margin can move the emissions result even when the physical product hasn't changed.

A sensible rule is to pursue supplier-specific data for the largest and most material product groups where product-level disclosure exists. Use average-data factors when the supplier can describe the product but can't support a product footprint. Reserve spend-based methods for gaps rather than using them as a permanent substitute for supplier engagement.

Mixing methods is normal. One product record may use a supplier factor, another may use an average material factor, and a third may use spend-based EEIO. The important control is to record the method, factor source, reporting year, unit, boundary, and assumptions for every line or product record. A hybrid inventory is defensible when the hierarchy is explicit. A spreadsheet that mixes factors without source metadata isn't.

A purchase line can look complete in an ERP while remaining unusable for product-level accounting. “Black tee,” a supplier part number, and an internal finished-good code may all describe the same item. Resolve that relationship before selecting an emission factor, or the same product can receive different treatment across sites, reporting years, or sourcing cycles.

Build a persistent product record, not a temporary spreadsheet key. Clean the ERP description and supplier reference, then connect the line to the internal item master. Reconcile manufacturer part numbers with the bill of materials, add a GTIN or other GS1 identifier where available, and retain the supplier's identifier as a secondary reference. Keep prior identifiers when a sourcing event or ERP migration creates a new code.

The quantity and the factor must describe the same basis. A supplier may report fabric in kilograms while procurement buys rolls. A packaging supplier may invoice cases while the item master stores units. A chemical supplier may use litres while the factor is expressed per kilogram. Convert each quantity to a documented unit before multiplying it by the factor, and retain the conversion basis with the record.

Classification needs the same discipline. Map spend categories to a recognised economic or product taxonomy, such as UN CPC or GICS where appropriate. Do not let free-text interpretation decide which factor applies. For a product-level factor, the identifier must resolve to the product, material, or component covered by that factor.

!A four-step diagram illustrating the process of mapping ERP purchase lines to final product identifiers.

System changes and acquisitions expose weak mappings quickly:

• Duplicate records: Different internal codes may refer to the same supplier product.
• Free-text references: “Black tee,” “T-shirt black,” and a supplier part number may identify one underlying item.
• Cycling internal codes: A code that changes with each sourcing event prevents consistent year-over-year tracking.
• Uncontrolled packaging records: Packaging bought separately may never be linked to the finished product.

A usable record retains identifiers, factor versions, approved evidence, and component relationships. It should remain traceable after an ERP migration. The Shopstar guide to SKU number basics for makers and creators can help teams standardise SKU logic before connecting procurement records to emissions data.

The same identifier set can support the inventory, product disclosure, and a Digital Product Passport definition. That continuity prevents teams from re-keying the product whenever reporting requirements change. It also makes coverage cutoffs and evidence status easier to reproduce in later reporting years.

Consider one cotton t-shirt SKU purchased at $7.50 per unit, with 250,000 units purchased each year. The annual procurement value is therefore $1,875,000, calculated from the stated unit cost and quantity. The examples below are illustrative calculations using the assumptions provided, not a claim that one factor applies to every cotton garment.

The spend-based approach applies an EEIO factor of roughly 0.4 kg CO2e per dollar, producing approximately 750 tCO2e annually. That result is quick to generate, but it reflects a broad economic average. It doesn't identify the actual cotton content, dyeing process, electricity mix, supplier yield, or manufacturing location.

The average-data approach applies a cradle-to-gate cotton garment factor of roughly 5.5 kg CO2e per unit, producing approximately 1,375 tCO2e. This method uses the physical product quantity, so it avoids the direct price sensitivity of spend-based accounting. It still represents an average product and may not capture the supplier's actual process.

A supplier-specific example uses a Tier 1 knit-and-dye supplier's reported 4.8 kg CO2e per unit, combined with an upstream cotton lint factor, producing approximately 1,250 tCO2e annually. The result is not automatically superior because a supplier supplied it. Reviewers still need to check system boundaries, allocation, data vintage, exclusions, and whether the upstream cotton input has been counted consistently.

The spread is wide enough to change the Category 1 total materially. For a first inventory, spend-based data can be a defensible starting point when better evidence isn't available, provided the methodology file labels it as an estimate. The next move should be to replace it with physical or supplier-specific data for products where the result could influence procurement decisions.

A spreadsheet can look exact while its controls are weak. A spend-based result shown to two decimal places may still rely on a broad sector factor that converts price into kilograms of CO2e. The GHG Protocol calculation guidance places spend-based methods lower in the data-quality hierarchy, particularly when price does not reflect physical quantity or service intensity.

A product identifier rarely captures every purchased input. Indirect materials, packaging for freight, dyes, detergents, softeners, labels, and co-manufactured sub-assemblies may sit in separate procurement records. The UNFCCC material on measuring Scope 3 purchased goods and services identifies inputs such as dyes, detergents, and softeners that may not appear in a bill of materials.

Inbound transport requires a boundary decision. If the supplier's price includes upstream transport and that movement is also assigned to Category 4, the inventory can count it twice. Do not remove transport by default. Record the boundary, then apply the same treatment across mapped products and reporting years.

Check these failure points before accepting the total:

• Unreviewed contract manufacturing: Supplier services may contain electricity, process materials, and production inputs that the purchase description does not show.
• Stale factors: A fixed base-year factor can misrepresent the procurement mix after a sourcing change.
• Net-only spend: Returns, rebates, credits, and cancellations can change purchased quantity and spend without a documented reconciliation.
• Sector averages on major SKUs: One broad factor can conceal supplier differences and weaken later reduction claims.

Keep a passport-style record for each product or input, including its identifier, method, factor, boundary, evidence, and coverage status. That record makes assumptions visible when data improves and prevents a new spreadsheet shortcut from replacing an old one.

Uncertainty is acceptable when it is labelled, bounded, and linked to a replacement plan. Presenting an estimate as a measured product footprint is not.

A supplier may return a polished footprint that cannot be matched to the product your company purchased. Prevent that failure at the request stage. Start with a controlled product register containing the mapped SKU, supplier part number, purchased quantity, reporting period, required unit, production site, and requested boundary. A specific request gives procurement and suppliers a shared reference instead of asking for “all emissions data.”

Build the workflow around the product record, not the document received. Ask for an ISO 14067 product carbon footprint, verified LCA report, an EPD whose relevant life-cycle stage matches the boundary, a CDP Climate Change response with its scope checked, or a third-party product carbon footprint aligned with relevant PACT or GHG Protocol technical guidance. A document title does not establish comparability. Review what the figure includes, which product and site it covers, its reporting period, allocation approach, and unit.

!A diagram outlining the four-step process for building a repeatable supplier data workflow for business management.

Use four controlled actions, with an owner and status for each:

1. Request: Send a product-specific questionnaire with deadlines, units, boundaries, and evidence requirements.
2. Review: Match the evidence to the product identifier, production site, purchased quantity, reporting period, factor, system boundary, and allocation method.
3. Approve: Record the reviewer, decision status, assumptions, limitations, and selected calculation method.
4. Refresh: At the next cycle, request changes to the quantity, site, factor, or supporting document rather than rebuilding the record.

Approved evidence stays attached to the persistent product identifier. A structured supplier product data collection workflow can support this pattern, although a controlled database with documented ownership can also work. The control points matter more than the software: versioned evidence, named approval, change history, and a defined fallback when a supplier does not respond.

Set the fallback hierarchy before collection begins. Use the latest approved supplier record while its boundary remains valid, then an average-data factor, then a spend-based estimate when physical or product data is unavailable. Record the reason, affected products, and replacement task. The next reporting cycle should show whether the estimate was replaced, not carry the shortcut forward.

A verifier won't only inspect the final Category 1 total. They'll want to trace a sample from the reported result back to the purchase record, product identifier, quantity, emission factor, source evidence, calculation, reviewer, and boundary decision. If those links live across email, personal drives, and an unversioned spreadsheet, the result becomes difficult to defend even when the arithmetic is correct.

The minimum evidence package should include:

• Calculation method: Supplier-specific, average-data, spend-based, or hybrid.
• Factor source: Dataset name, version, unit, geography, and applicable product or sector.
• Supplier support: Attestation, LCA, PCF, EPD, or other referenced evidence.
• Assumption log: Allocation, cut-offs, conversions, exclusions, and treatment of transport.
• Review record: Named approver, review date, status, and unresolved limitations.

Version history matters because factors and supplier records change. Immutable timestamps and access controls show which evidence supported a published result at the time of approval. That control becomes more important when the same product data feeds a Digital Product Passport or product-level claims under frameworks such as the ESPR or EU Battery regulation.

The chain-of-custody resource is relevant when product identity, material provenance, and evidence continuity need to remain connected across transactions. Category 1 governance follows the same principle: retain the relationship between the product, its upstream evidence, and the approved calculation.

A governed record turns the annual inventory into a repeatable disclosure input. Each reporting period becomes a refresh and exception process, not a fresh attempt to reconstruct what procurement bought.

A workable rollout starts small enough to control and broad enough to expose the data problems. Select 20 pilot SKUs across the top three spend categories and the top three material types. Include at least one product with good supplier data, one with partial information, and one that currently relies on spend-based estimation.

Extract supplier, purchase order, quantity, mass, unit, currency, product description, internal SKU, supplier SKU, and accounting category from the ERP. Reconcile duplicates, returns, credits, and unit conversions. Then create the governed product record that will hold the factor, evidence, calculation, and approval status.

Don't wait for perfect supplier data before fixing identity. A supplier can't reliably answer a product request if the buyer and sustainability team disagree about which SKU, site, or component the request concerns.

!A 90-day plan infographic illustrating a three-phase approach to operationalize Category 1 with clear milestones and goals.

Run spend-based, average-data, and supplier-specific calculations in parallel wherever the inputs allow it. The purpose isn't to choose the largest or smallest result. It's to expose where method choice drives uncertainty and where supplier evidence would change the procurement decision.

Use the comparison to prioritise outreach. A product with a large quantity, weak data, and a material supplier relationship deserves more attention than a small service purchase that can remain on a documented spend-based fallback.

Set the coverage rule, escalation path, evidence requirements, reviewer roles, and fallback hierarchy. The available guidance points toward tighter expectations for coverage and auditability. A 2026 guide describing a 95% Scope 3 coverage floor with 5% allowed as justified gaps illustrates the direction of travel, while the same industry discussion emphasises transparency about primary-data coverage, factor databases, and improvement plans.

At hand-off, publish the methodology file, attach approved evidence to the product records, and assign owners for the next refresh. The same structured dataset can then support product-level disclosures and CSRD-aligned reporting without another manual mapping exercise.

DPP Grid provides persistent product identities, supplier data requests, evidence-linked fields, approval states, and versioned product records that can connect Category 1 calculations with Digital Product Passport workflows. Visit DPP Grid to assess how a governed product record could support your next supplier-data cycle and reporting hand-off.