Methodology Overview

Activity-Based, Evidence-First

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.

Key Principles

"Activity-First" Data Modeling

Every purchase is translated into a physical quantity and activity. For a given transaction, Bardo asks "What did we actually buy or do?" – e.g. kilograms of steel, kilometers of freight transport, kilowatt-hours of electricity, hours of consulting, and uses that as the basis for emissions calculation. This contrasts with a spend-based mindset that asks "How much money did we spend in this category?". By using actual usage data where available, the approach adheres to the GHG Protocol's guidance to seek higher-quality (more specific) data for significant emissions sources. The trade-off is more data collection effort, but it dramatically improves accuracy per activity.

Supplier/Product Specific Emission Factors Preferred

When possible, Bardo uses supplier-specific emission factors, meaning if Supplier A has published an emission per unit for their product (or provided primary data), that is utilized. If no direct data exists, Bardo will construct a factor from similar products or known industry data and documenting assumptions (this is sometimes called a "proxy LCA" or evidence-based factor). Only as a last resort does Bardo use generic industry averages. This mirrors the data quality hierarchy many standards endorse: use primary data for key suppliers, and secondary data (averages) for less material items. By prioritizing specificity, the inventory becomes more reflective of the company's actual supply chain rather than a generic economy. Over time, as better data becomes available (say a supplier issues a new LCA), Bardo upgrade the factors, ensuring continuous improvement of accuracy.

Services Are Accounted via Their Drivers

Services often involve indirect emissions (e.g. an IT service's emissions come from data centers, an office lease's emissions come from building energy use, cleaning chemicals, etc.). Rather than accept that "services have no data" and assign an arbitrary factor per dollar, Bardo breaks down services into constituent parts. For example, for cloud services, it identifies usage metrics (compute hours, GB-month of storage) and region, aligning with how cloud providers measure carbon footprints. For an office lease (tenancy), it estimates the portion of building energy, water, waste services, etc., attributable to that space. This level of modeling ensures that outsourced or intangible activities are not a blind spot, they're measured with the same rigor as physical goods. It's widely acknowledged that broad-brush approaches to services can mislead; by modeling them, Bardo provides numbers that are more actionable (e.g. informing a decision to consolidate office space or choose cloud regions with cleaner energy). In essence, no category is left as "too hard to model", because those often hide significant emissions which, if simply averaged, could distort the footprint or inflate it without offering insight to reduce it.

Bundled Emissions Are Separated ("Unbundled")

Many purchased products or services have embedded emissions from other categories: for instance, the production of a machine (capital good) includes emissions from manufacturing (Category 2) and often shipping that machine to the buyer (Category 4), plus perhaps installation services (Category 1) and even upstream energy used in its production (Category 3). Bardo's approach is to separate these components for clarity. In practice, the system might record multiple line items for one supplier invoice: e.g., 2000 kg CO₂ for manufacturing of equipment (Cat 2), 100 kg CO₂ for freight of that equipment (Cat 4), 50 kg for installation services (Cat 1). Each is tagged to its proper category. This yields clean category totals that make sense (manufacturing vs transport vs services are distinct) and avoids double counting. It also makes year-over-year changes easier to explain, e.g., if you buy the same machine but shipping method changes, you'd see Category 4 change but Category 2 remain comparable.

Scope 2 Emissions Treated with Proper Upstream Accounting

Bardo ensures Scope 2 (purchased electricity, heat, etc.) is reported per protocol (market- and location-based as required), and additionally accounts for the upstream lifecycle of energy in Scope 3. By linking these records, a company can see the full impact of its energy procurement. For example, buying 1 MWh of electricity has a direct Scope 2 impact (say 0.3 tCO₂ if grid emissions intensity is 0.3 kg/kWh), but also upstream emissions from extracting fuel, generating and transmitting that electricity (perhaps an additional 0.05 tCO₂ in Category 3). Bardo would record both, tied together. Many simplistic tools might ignore the 0.05 or accidentally merge it into Scope 2. Bardo keeps them separate for compliance (Scope 2 vs 3) yet linked for internal analysis, giving a more complete picture of energy emissions.

Reconciled, Versioned, and Auditable

Each reporting cycle (e.g. annual footprint) is handled as a controlled snapshot. Bardo reconciles the activity coverage to financial totals as mentioned, and locks the period's data for audit. If factors or methodologies update in future cycles, changes are tracked with versioning, so if an emission factor is refined, one can see the difference it would make (this helps explain changes year-over-year that are due to methodology, not actual activity change). The inventory thus has a memory: it can be reproduced exactly as reported for a given year, even if improvements are made later. Every data point has provenance (source of factor, who/when it was reviewed or modified, etc.). This approach treats carbon data with the same rigor as financial data, which is increasingly expected by regulators and auditors. In contrast, spreadsheet-based approaches often have trouble with version control (e.g. a formula change might alter prior results without a clear record).

Human-in-the-Loop for Edge Cases

While much of Bardo's process is automated or systematic, it isn't a black box software that simply outputs numbers unreviewed. Analysts review low-confidence items or unusual transactions. For example, if an outlier transaction is flagged (perhaps a very large spend in an unrecognized category or an ambiguous description), a human expert intervenes to determine the correct activity and factor. These resolutions are then encoded into the rules going forward. This managed aspect ensures that the methodology can handle real-world data quality issues (like miscoded spend or novel purchases) that purely automated tools might mishandle. Traditional consulting naturally has humans in the loop (they're doing everything), and pure software has none; Bardo aims for a middle ground where automation covers the bulk, and experts address the tricky parts, improving the overall quality continuously.

The Hybrid Approach

By adhering to these principles, Bardo's method aligns with what the GHG Protocol calls the "hybrid approach", using an initial spend-based mapping for completeness, then overlaying activity data for accuracy. This hybrid method is recommended because it ensures all economic activities are accounted for while maximizing precision where data allows. In Bardo's case, the "spend mapping" is essentially the process of building the ledger from financial records (ensuring completeness), and the "activity overlay" is populating each line with the best possible activity data and factor. The outcome is a carbon footprint that is both comprehensive and precise, closing what is widely known as the "accuracy gap" in corporate carbon footprints.