Scope 3: a practical guide to the GHG Protocol's 15 categories

Why Scope 3 is the hardest, and the most relevant

The GHG Protocol divides corporate emissions into three scopes. Scope 1 covers direct emissions: combustion in boilers, owned fleet, industrial processes. Scope 2 covers purchased electricity. Both are relatively simple to calculate because the company controls the data.

Scope 3 is different. It covers every indirect emission across the value chain, upstream (suppliers, inbound logistics, corporate travel) and downstream (use of sold products, waste treatment, investments). For most companies, Scope 3 is between 70% and 90% of the total carbon footprint.

That weight creates a paradox: the most significant scope is also the hardest to measure. The data sits across hundreds of suppliers, carriers, logistics partners and end customers. Quality varies enormously. And the organisational boundaries are not always clear.

Even so, Scope 3 has become unavoidable. CDP requires Scope 3 reporting for A and A- scores. SBTi requires Scope 3 targets when it accounts for more than 40% of total emissions. The European CSRD requires detailed value chain disclosure. And CVM 193 in Brazil points the same way.

The good news: you do not have to calculate everything with absolute precision from the first cycle. The GHG Protocol Scope 3 Standard allows a progressive approach: start with sector estimates, refine with primary data over time, and focus on the hotspots that actually move the needle.

The 15 Scope 3 categories in one table

The GHG Protocol divides Scope 3 into 15 categories. The first 8 are upstream (before the product leaves the company). The next 7 are downstream (after the sale).

The typical method for each category, and the factor databases, are in the tools table further down.

Upstream (categories 1-8), one by one

The upstream categories cover everything that happens before the product reaches the company gate.

Category 1, purchased goods and services. The production emissions of every input the company buys: raw materials, components, packaging, outsourced services. It is usually the largest Scope 3 category for manufacturing and retail. The calculation can use a spend-based approach, physical quantity, or primary supplier data. Spend-based is the simplest, multiplying the amount spent by economic emission factors (from EEIO models, such as the EPA's USEEIO or Exiobase), but it is also the least accurate.

Category 2, capital goods. Production emissions of capital assets acquired during the year: machinery, equipment, buildings, vehicles. Unlike financial accounting, the GHG Protocol recommends accounting for all emissions in the year of acquisition, with no depreciation. Companies making large infrastructure investments will see spikes here.

Category 3, fuel- and energy-related activities not included in Scopes 1 and 2. The emissions of extracting, producing and transporting the fuels and electricity the company consumes but which are not counted in Scopes 1 and 2. It includes electricity transmission and distribution losses. In Brazil, with a predominantly hydro power mix, this tends to be smaller than in countries with fossil-heavy grids, but it is not negligible.

Category 4, upstream transportation and distribution. Emissions from transporting purchased inputs from direct suppliers to the company, in vehicles it does not control. It also includes every freight service the company contracts and pays for, including outbound freight of sold products and transport between the company's own sites. The care needed is not to count the same freight in Category 1 as well, when the input's factor already includes transport. In Brazil, where road transport carries more than 60% of freight, this category is usually significant.

Category 5, waste generated in operations. Emissions from treating and disposing of the solid waste and effluent the company generates. It covers landfill, incineration, recycling and effluent treatment. The calculation method depends on the waste type and final destination. Companies with industrial processes generating large volumes of organic waste will see substantial figures.

Category 6, business travel. Emissions from air and ground travel and accommodation for employees. Relatively simple to calculate when the company has a centralised travel agency. Tools such as the ICAO calculator estimate flight emissions from origin, destination and class. The absolute volume is usually small, but it is a highly visible category internally.

Category 7, employee commuting. Emissions from employees' home-to-work travel. It requires an internal survey covering transport modes, distances and frequency. In Brazilian cities with heavy traffic and limited public transport, this category can surprise. Companies with home-working programmes achieve measurable reductions here.

Category 8, upstream leased assets. Emissions from operating assets the company leases from third parties and which were not included in Scopes 1 and 2. Examples: rented premises, vehicles under operating leases, leased equipment. If those assets were already accounted for in Scopes 1 and 2 under the operational control approach, this category is zero.

Downstream (categories 9-15), one by one

The downstream categories cover what happens after the product leaves the company gate.

Category 9, downstream transportation and distribution. Emissions from transporting and storing the sold product on its way to the end consumer, when the company does not pay for that service. Outbound freight that the company contracts and pays for stays in Category 4. It covers distribution centres, refrigerated storage, retail and last-mile delivery carried out by third parties. For e-commerce and retail, this category grows with logistics reach.

Category 10, processing of sold products. Emissions from processing intermediate products sold to industrial customers. Relevant for producers of steel, resins, paper and other inputs that get transformed before final use. It requires knowledge of the customer's processes, which makes data collection especially challenging.

Category 11, use of sold products. Emissions generated during the useful life of the products the company sells. For vehicle manufacturers, fuel combustion is accounted for here. For appliance makers, energy use. For software or service companies, this category may be zero or irrelevant. In sectors with energy-intensive use, it is frequently the largest Scope 3 category.

Category 12, end-of-life treatment of sold products. Emissions from disposing of and treating products after consumer use. It covers landfill, incineration and recycling. It requires assumptions about product lifetime and the disposal mix in the market served. Producers of single-use packaging have significant exposure here.

Category 13, downstream leased assets. Emissions from operating assets the company leases to third parties. Relevant for vehicle, property and equipment rental businesses. If those assets were already included in Scopes 1 and 2, this category is zero.

Category 14, franchises. Emissions from franchisee operations. It applies only to franchisors using the operational control approach in Scopes 1 and 2 (and which therefore excluded franchises from those scopes). Fast-food chains, filling station networks and retail chains find substantial figures here.

Category 15, investments. Emissions associated with the company's financial holdings: equity, debt, project finance. Particularly relevant for banks, insurers, asset managers and holding companies with diversified stakes. PCAF is the reference standard for this category in financial services.

The 7 most common Scope 3 mistakes (and how to avoid each one)

Knowing the 15 categories is the first step. The second is not repeating the mistakes that show up in almost every first cycle. Because Scope 3 depends on third-party information, with variable quality and slow collection, the first attempt usually reaches independent verification with structural gaps. These are the seven mistakes that come up most, and how to avoid each one.

1. Skipping categories without justification

The GHG Protocol requires every omitted category to carry a documented justification: materiality below the defined threshold, or data unavailable with an action plan for the next cycle. Silent omission is a non-conformity.

How to avoid it: record the screening of all 15 categories in a relevance matrix, with the reason for each exclusion and the cycle in which it will be revisited.

2. Using secondary data when primary data exists

Spend-based or activity-based with a default factor is only acceptable when primary data is unavailable. When it exists, the verifier expects it to be used. Ignoring it lowers the quality of the GHG Inventory and can lead to a qualification in independent verification.

How to avoid it: before applying an average factor, check whether the supplier, the carrier or an internal system already holds the actual activity data.

3. Confusing "outsourced" with Scope 3

An outsourced fleet for your own use (a third-party driver operating a vehicle under the company's operational control) is Scope 1. A carrier contracted and paid for by the company is Scope 3, category 4, even for outbound freight. Category 9 is reserved for transport of the sold product that the company does not pay for. A rental car for occasional use falls under category 6. It is the most recurring confusion in logistics.

How to avoid it: apply the operational control test case by case and document the decision. The right question is who operates the vehicle and who is responsible for maintenance, not who owns it.

4. Ignoring business travel (category 6)

It is one of the fastest categories to calculate, because the travel agency or the HR system already consolidates the data. Even so, it is often forgotten. At professional services firms it can reach 15-20% of total emissions.

How to avoid it: ask the travel agency for the annual report of trips, hotel stays and travel classes at the start of the cycle.

5. Underestimating purchased goods and services (category 1)

Usually the largest source, and it needs spend-based or activity-based done properly. The classic mistake is applying one generic factor across a base of 5,000 different suppliers.

How to avoid it: start with aggregate factors per product or service class and refine year by year, prioritising the classes with the highest spend and carbon intensity.

6. Forgetting use of sold products (category 11)

For equipment, vehicle and electronics manufacturers it is frequently the largest category. In some software models, server consumption at the client also weighs in. A company in these sectors that ignores category 11 may be omitting more than 50% of its emissions.

How to avoid it: include category 11 in the screening from the first cycle and model consumption over the product's useful life. The calculation method is in the FAQ at the end of this guide.

7. Not engaging suppliers early

Without primary data, precision drops and the verifier asks for an improvement plan. Leaving supplier contact to the end of the cycle also squeezes deadlines and lowers response rates.

How to avoid it: start with the 20-30 suppliers that make up 70-80% of spend. It is the path with the best return, and the step-by-step is in the engagement section below.

Primary vs secondary data: when to use each

The GHG Protocol distinguishes two types of data for Scope 3.

Primary data is information specific to the actual activity: the exact quantity of steel bought from a specific supplier multiplied by the emission factor of the mill that produced it. It is the most accurate, but it requires direct engagement with suppliers and can be hard to obtain at scale.

Secondary data is an estimate based on sector averages, economic factors or proxies. Examples: emission factors per unit of spend in a given sector (spend-based), average factors per tonne of commodity, average transport distances by mode. Easier to obtain, less accurate.

The practical recommendation is a hybrid approach. In the first inventory cycle, use secondary data across all 15 categories and identify the hotspots, the 3 to 5 categories that concentrate most of the emissions. In the second cycle, start collecting primary data for those hotspots. In later cycles, expand primary data coverage progressively.

In practice, this becomes a methodological hierarchy. For each category, climb one step per cycle:

1. Primary supplier data (best): their verified actual emissions, allocated to your consumption.
2. Activity-based with a local factor: physical unit × verified regional factor.
3. Activity-based with a default factor: physical unit × IPCC or DEFRA factor.
4. Spend-based: monetary value × monetary factor (EEIO models, such as USEEIO or Exiobase).
5. Sector default value: a single figure, with no collection. Last resort.

SBTi asks for near-term Scope 3 targets to cover at least 67% of Scope 3 emissions. That does not mean you need primary data for 67%, it means you need to include enough categories to cover that share.

Supplier engagement: how to collect the data

Collecting Scope 3 data is, in essence, an exercise in value chain engagement. Your most relevant suppliers need to understand why you are asking for data, exactly what is needed, and how it will be used.

Step 1, prioritise. Identify the suppliers accounting for the largest share of your spend or volume and cross-reference with the most carbon-intensive sectors. Typically 20% of suppliers account for 80% of upstream Scope 3 emissions.

Step 2, communicate. Send a formal communication explaining the initiative, the deadlines and the expected response format. Include a standardised form. Many companies use platforms such as CDP Supply Chain or their own questionnaires.

Step 3, technical support. Many suppliers, especially SMEs, have no emissions inventory of their own. Offer support: webinars, guides, simplified calculators. Mangue Tech provides a supply chain engagement module that sends questionnaires, collects responses and calculates emissions automatically.

Step 4, validate. Supplier-declared data needs cross-validation. Compare with sector benchmarks, check consistency over time and identify outliers. Emission factors far above or below the sector average deserve investigation.

Step 5, iterate. Supplier engagement is cyclical. Each year, data quality and coverage are expected to improve. SBTi accepts supplier engagement targets, in which a share of the chain, measured by emissions, adopts its own reduction targets.

Suggested operational roadmap

Putting the data hierarchy and supplier engagement together, a realistic roadmap for the first cycles looks like this:

Tools and emission factors for each category

Each Scope 3 category has its own emission factor sources and calculation tools.

The Mangue Tech platform maintains more than 60,000 up-to-date emission factors mapped to each category, drawing on MCTI, IPCC AR6, DEFRA and Ecoinvent. The system automatically suggests the most appropriate factor for each type of data collected.

How Mangue Tech automates Scope 3

The traditional approach of calculating Scope 3 in Excel spreadsheets has serious limits: error risk, no traceability, difficulty updating and no efficient way to audit.

The Mangue Tech platform solves that with a structured workflow. The Scope 3 module lets you configure the relevant categories and import purchasing data (ERP), logistics (TMS) and travel (travel agency) via integrations or bulk upload. The system applies the latest emission factors automatically, calculates emissions by category and produces reports ready for CDP, SBTi and the GHG Protocol.

The supply chain engagement module completes the workflow: it sends tailored questionnaires to suppliers, collects structured responses, validates data against benchmarks and feeds the results straight into the Scope 3 calculation.

Crab.AI, the Mangue Tech AI, maps invoices to emission factors across the 15 categories in seconds, with an audit trail. The consulting team validates materiality, the collection plan and the narrative. Typical result for a new client: a first complete Scope 3 in 8 to 12 weeks, covering more than 90% of spend, at a quality approved in independent verification under limited assurance.

The result: a complete, auditable, up-to-date Scope 3 inventory, without the weeks of manual work the traditional process demands.