Methodology
A strict, evidence based interpretation of the GHG Protocol. Activity first, LCA based, audit ready.
The method below is the standard we apply on every engagement. It is a strict interpretation of the GHG Protocol that is feasible today because activity data can be constructed at scale from financial evidence and documents. The outcome is an inventory that is auditable, comparable, and useful for decisions.
1. Purpose and scope
Purpose.
Produce an organisation level greenhouse gas inventory that is complete across upstream costs, auditable to source documents, and aligned to the GHG Protocol with clear disclosure of policies and boundaries.
Scope of this method.
Upstream emissions tied to what a company buys, materials, goods, transport, energy, and services. This covers the vast majority of spend and can be evidenced well.
CO2e only, default GWP100. A bridge table can be provided to another assessment.
Downstream categories can be added without changing the upstream ledger.
Two views are always maintained, a management view grouped by legal entity, cost centre, General Ledger Account ("GL Account"), and supplier, and a Protocol view tagged by scopes and the fifteen Scope 3 categories.
2. Alignment with the GHG Protocol principles
We operationalise the five accounting principles with specific controls.
GHG principle | Bardo control |
|---|---|
Relevance | Activity based by default, factors chosen by specificity to the real purchase and use context. |
Completeness | Coverage mapped from Accounts Payable ("AP") and GL to activity categories, no silent exclusions, documented gaps and plans. |
Consistency | Documented mapping rules and factor doctrine, versioned and carried across periods, base year policy and recalculation log. |
Transparency | Every value carries sources, scope notes, and change history, with access to evidence packs. |
Accuracy | Preference for supplier or product LCAs, quality scoring of factors, uncertainty quantified and reported. |
3. Method overview
There are two jobs behind any credible inventory. We standardise both.
Construct the activity that actually happened from financial documents, with units, supplier, model, route, or grade.
Select or construct the factor that best represents the activity, then calculate CO2e with documented boundaries and uncertainty.
The result is stored in a Carbon Footprint Inventory that records transaction, activity, factor, calculation, scope tags, and provenance.
4. Activity construction
Inputs.
AP and GL exports, invoices, purchase orders, supplier master, shipping and route documents, lease and tenancy information, asset lists.
Method.
Parse documents, normalise fields, and resolve entities, suppliers and products.
Link related documents to prevent double counting, for example an invoice and its attached freight line.
Find or reasoning to list physical units where required, kilograms, pieces, ton kilometres, kilowatt hours, and record units explicitly.
Assign activity types, for example material, transport, energy, service.
Map coverage to finance, accounts and cost centres, so completeness is visible and reconcilable.
Disclosure.
Activity construction policies and edge cases are recorded in method notes. Confidence and uncertainty signals travel with each activity.
5. Factor selection doctrine
A factor wins on two dimensions, specificity to the activity, and quality of evidence.
Order of preference.
Supplier and product specific LCA, verified or attested.
Evidence constructed LCA, built from transaction evidence and recognised life cycle data, fully documented and marked for upgrade when a supplier LCA becomes available.
Product or model specific LCA from public sources.
Process or material specific factor with a regional match.
Sector or regional average, kept only until a more specific factor is available.
Evidence carried by every factor.
Source, year, geography, system boundaries, functional unit, allocation rules where relevant, and reviewer notes. Each factor receives a quality assessment across technology, time, geography, completeness, and reliability. That assessment sets an uncertainty value that propagates to results.
Acceptance of supplier claims.
Supplier provided data is used when documentation meets the quality and boundary requirements. If it does not, we disclose the gap and use the best alternative.
6. Calculation and aggregation
Calculation.
Align units and functional units, multiply activity quantity by the chosen factor, carry uncertainty and scope notes.
Aggregation.
Protocol view, roll up into Scopes and Scope 3 categories, with clear disclosure of category policies and any exclusions.
Management view, roll up by finance dimensions, legal entity, cost centre, GL account, supplier, and time.
Both views are derived from the same activity and factor records, so numbers remain consistent.
7. Scope 3 upstream, where the complexity lives
Most reporting fails here because it relies on averages that inflate totals and erase context. Categories overlap, sources are vague, and no one can defend the results. Our method grounds every number in activity evidence, so it holds up under scrutiny.
7.1 Purchased goods and services (Scope 3.1)
Why today's numbers mislead
Most inventories reduce goods and services to spend averages. An invoice for "consulting," "office lease," or "steel delivery" becomes a sector factor per euro. The result: inflated totals, missing categories, and no connection to what actually happened.
Our approach
Activities, not accounts: Every purchase is resolved to the physical driver: kilograms of material, square meters of tenancy, hours of service, ton-kilometers of freight.
Services unpacked: A lease becomes energy + cleaning + maintenance + materials. SaaS becomes compute + storage + region. Consulting becomes hours + travel + materials.
Bundles split: Embedded freight, upstream energy, and pass-through costs are separated so Scope 3.1 stays distinct from 3.3, 3.4, or 3.6.
Evidence-first factors: Supplier or product LCAs whenever possible; evidence-constructed LCAs if none exist, documented and marked for upgrade.
The outcome
Scope 3.1 that reflects what was actually bought, not what the invoice called it. Comparable, defensible, and a basis for action, not a compliance guess.
What auditors see
Activity record linked to the invoice, factor record with sources and boundaries, change notes when the factor improves.
7.2 Capital goods, Category 3.2
The challenge
Large one time assets vary by technology, capacity, and configuration. Generic machine factors blur this. Invoices often bundle equipment, freight, installation, and service contracts. Results spike in one year, then become hard to compare.
Our method
Define the asset activity with model, capacity, configuration, manufacturer, commissioning year, and site.
Prefer a supplier or product LCA. If none exists, use an evidence constructed LCA from BOM and recognised process datasets, documented and marked for upgrade.
Unbundle the purchase, equipment in 3.2, inbound freight in 3.4, installation energy in 3.3, contractor services in 3.1, so categories stay clean.
Keep two views, Protocol view books the full cradle to gate footprint in the year of purchase, Management view allocates over useful life to match finance, both are stored with a clear reconciliation bridge.
Treat retrofits and major overhauls per financial treatment, capitalised items in 3.2, operational services in 3.1.
What auditors see
Asset records with factor sources and boundaries, the allocation policy and the reconciliation bridge, plus sampleable links to the invoice, PO, and spec sheet.
7.3 Fuel- and energy-related activities (not in Scopes 1 or 2), Category 3.3
The challenge
Upstream emissions from purchased electricity and heat — extraction, processing, and delivery — are often hidden or double counted. Transmission and distribution losses are ignored or blended into Scope 2, which makes results incomparable and non-auditable.
Our method
Model upstream generation and delivery losses for every unit of purchased energy.
Keep these upstream records linked to the same purchase that feeds Scope 2, so the full life cycle stays visible.
Tag consistently, with direct generation in Scope 2 and upstream emissions in 3.3, while management views present the combined total for clarity.
What auditors see
One purchase, two linked records: Scope 2 for the purchased energy, Scope 3.3 for upstream fuel and delivery, with scope notes that prevent overlap or omission.
7.4 Upstream transport and distribution, Category 3.4
The challenge
Freight emissions hinge on mode, route, and load factor. Generic kilometre averages ignore these drivers. On top of that, transport is often buried inside vendor invoices, misclassified as goods (3.1) or services.
Our method
Build ton-kilometre activities using internal data, or reconstructed route models.
For flights, calculate each leg using origin–destination, aircraft type, and cabin class. This ensures results are based on the specific route and load profile, not a broad kilometre average.
Select mode- and route-specific factors first; if missing, construct and document a traceable LCA.
Unbundle freight from goods invoices so 3.1 stays goods/services and 3.4 holds transport only.
What auditors see
Lane- and leg-level activities, factor records with route logic, and clear category tagging across 3.1 and 3.4.
7.5 Waste generated in operations, Category 3.5
The challenge
Treatment pathways vary by site and operator. Weights and materials are often missing.
Our method
Create waste stream activities with mass, material, and treatment.
Use treatment specific factors by region.
Where weights are missing, construct estimates from container data and vendor tickets, then mark for upgrade.
What auditors see
Stream level records with treatment logic and evidence from tickets or operators.
7.6 Business travel, Category 3.6
The challenge
Route and cabin class drive results. Global averages over or understate legs. Supplier claims vary in quality.
Our method
Build flight leg activities with route and class.
Use a route and class aware model as the primary factor set for flights.
Non flight travel follows mode and class factors.
Accept supplier claims only with documentation that meets quality checks.
What auditors see
Leg level activities, factor records with sources, and policy notes for claims.
7.7 Office tenancy and services, treated in 3.1 and 3.3 with Scope 2 links
The challenge
Operation differs by climate, grid, and building stock. Rent bundles energy and services, which hides categories.
Our method
Construct tenancy activities by address and area.
Build a location specific LCA for operation and services, not a global service average.
Split bundled energy to keep Scope 2 and 3.3 tagging correct, while management views show the complete tenancy footprint.
What auditors see
Tenancy records with local data, split lines for energy and services, and clear tagging.
7.8 Double counting prevention across categories
One place only mapping rules for frequent collisions, for example goods vs embedded freight, electricity generation vs delivery losses, energy embedded in materials.
Detectors flag conflicts. An exceptions log records the decision and the fix.
Results reconcile to AP and GL, which closes the loop on coverage.
9. Data quality and uncertainty
Quality assessment. For every factor we record quality indicators across technology, time, geography, completeness, and reliability.
Uncertainty. We translate the assessment into an uncertainty value at factor level and propagate uncertainty through calculations and category roll ups.
Disclosure. Category summaries include uncertainty bands and a plan to reduce them through specificity upgrades, supplier LCAs, better route data, and tenancy disclosures.
10. Double counting prevention and reconciliation
One place only mapping rules prevent overlaps, for example freight embedded inside a vendor invoice versus separate logistics lines, electricity generation versus delivery losses, energy embedded in materials.
An exceptions log records detected conflicts and their resolution.
Reconciliation routines tie totals back to AP and GL so coverage and movement can be explained in finance terms.
11. Boundaries, base year, and recalculation
Organisational boundary is recorded per entity, equity share or control, and applied consistently.
A base year is set. Recalculation triggers include structural change and method changes that materially affect totals.
When recalculation is required, we publish a reconciliation bridge so finance and auditors can explain movement.
12. Evidence package for assurance
Every reporting cycle includes the artefacts assurance teams need.
Method notes, boundary and category policies, energy purchase policy, data quality rules, base year status.
Sampling lists with links from any reported value to the exact transaction, the constructed activity, and the factor record.
Factor records with sources, boundaries, versions, reviewer notes, and change logs.
Reconciliation pack to AP and GL.
Factor catalogue summary, share of supplier or product LCAs, constructed LCAs, process factors, and averages, plus the improvement plan.
13. Limitations and disclosures
CO2e only, GWP100 by default.
Upstream focus by design. Downstream can be added, the upstream ledger does not change.
When no suitable factor exists, a constructed LCA is used with explicit documentation and marked for upgrade.
Offsets and avoided emissions are presented outside Scope totals.
The method reports best estimate values with uncertainty and provenance. We do not inflate numbers to appear conservative.
14. Summary of benefits for expert users
A strict, transparent interpretation of the Protocol that is practical at scale.
Activity based inventories that reconcile to finance and survive sampling.
Factor choices that favour specificity and evidence, not convenience.
Uncertainty that is explicit and falling as specificity improves.
A single ledger that supports filings and decisions without rework.
See how your data performs in a quick CSRD check
Book a demo, or open the ROI calculator to estimate time and cost.