All insights & guides

Guide · Carbon Accounting

Scope 3 Categories: All 15 Explained with Examples

March 2026 · Updated May 2026 · 14 min read

Scope 3 emissions are 70–90% of most companies' carbon footprint and the hardest part of any inventory. The GHG Protocol Corporate Value Chain Standard splits them into 15 categories. This guide takes you through every one — what's in, what's out, how to calculate it, a real worked example, the trap most companies fall into, and which industries should care most.

TL;DR

  • 15 categories: 1–8 upstream (your suppliers, inputs), 9–15 downstream (sold products, investments).
  • 3–5 categories usually dominate. Do a materiality screen first — don't try to perfect all 15.
  • Method ladder: supplier-specific > hybrid > activity-based > spend-based. Climb it where it matters.
  • The classic mistake: double-counting Category 1 with Categories 3–4. Read the GHG Protocol Technical Guidance.

What are Scope 3 emissions?

Scope 3 emissions are all indirect greenhouse gas emissions across a company's value chain — both the production of inputs you buy (upstream) and the use of products you sell (downstream). They sit alongside Scope 1 (direct emissions from owned operations) and Scope 2 (purchased electricity, steam, heat and cooling).

Under the GHG Protocol Corporate Value Chain (Scope 3) Standard, Scope 3 is divided into 15 categories. CSRD/ESRS E1, ISSB IFRS S2, CDP and SBTi all reference this same structure, so getting it right once carries you through every major reporting framework.

All 15 categories at a glance

Click any category to jump to its detailed breakdown.

Upstream (your value chain ↑)

Downstream (your products in use ↓)

Which categories matter most for your industry

Don't try to perfect all 15. Most companies have three or four dominant categories. Here's the rough map — use it as a starting point, then run your own materiality screen.

IndustryDominant Scope 3 categoriesTypical share of Scope 3
ManufacturingCat 1 (purchased goods) · Cat 11 (use of products)60–80%
AutomotiveCat 11 (use of vehicles) · Cat 1 (steel, aluminium, batteries)80–95%
Software / TechCat 1 (cloud, hardware) · Cat 2 (data centres) · Cat 7 (commute)60–80%
Retail / Consumer goodsCat 1 (product sourcing) · Cat 11 (product use) · Cat 12 (packaging EoL)85–95%
Financial services / BanksCat 15 (financed emissions)>95%
Asset managers / PECat 15 (investments)>98%
Real estateCat 13 (leased assets) · Cat 1 (construction materials, embodied carbon)70–90%
Logistics / FreightCat 4 (upstream) · Cat 9 (downstream) · Cat 1 (vehicles, fuel)80–95%
Food & beverageCat 1 (agricultural inputs) · Cat 4 (cold chain) · Cat 12 (packaging)80–95%
Oil & GasCat 11 (combustion of sold fuels)>90%
Energy & UtilitiesCat 3 (T&D losses) · Cat 11 (use of sold energy)60–80%
Professional servicesCat 1 (purchased services) · Cat 6 (travel) · Cat 7 (commute)60–80%

Upstream categories (1–8)

Everything that happens before a product reaches you. Suppliers, inputs, energy upstream of your meter, transport into your sites, employee commuting and travel.

CATEGORY 01 · UPSTREAM

Purchased goods and services

Emissions from the production of products and services your company buys — everything from raw materials and components to professional services and SaaS subscriptions.

✓ In scopeAll cradle-to-gate emissions of every purchased good or service, including extraction, processing, manufacturing, packaging and transportation up to the supplier's gate.
× Not hereCapital goods (Cat 2), purchased fuel and energy upstream (Cat 3), transport from supplier to you (Cat 4).

How to calculate: Supplier-specific (best) → hybrid (mix supplier & secondary data) → activity-based (kg of steel × emission factor) → spend-based (£ spent × industry factor). Walk up the ladder where it matters most.

Worked example: A mid-market manufacturer spends £12M/year on steel. Spend-based: £12M × 2.3 kg CO2e/£ (UK steel industry average from BEIS) = 27,600 tCO2e. Activity-based: 5,000 tonnes × 1.9 tCO2e/tonne (EU verified steel) = 9,500 tCO2e. Supplier-specific data drops it further if the supplier uses electric arc / scrap steel. Three answers, same activity.

Pitfall: Often the largest Scope 3 category, so worth getting right. Spend-based factors hide real reduction opportunities — once you've done one, climb to activity-based for your top 20 spend lines.

CATEGORY 02 · UPSTREAM

Capital goods

Emissions from producing capital assets you've purchased — buildings, machinery, vehicles, IT hardware, equipment. The embodied carbon of the things you own.

✓ In scopeCradle-to-gate emissions of newly purchased capital assets in the reporting year. The full amount is reported in the year of purchase (no amortisation).
× Not hereOperational emissions of running those assets (Scope 1 / 2). Capital expenditure recorded for accounting but not delivered yet.

How to calculate: Same ladder as Category 1. For buildings, use embodied carbon factors (kg CO2e per m² by building type and structure). For IT hardware, use Product Carbon Footprints (PCFs) from manufacturers.

Worked example: A retailer fits out a new 10,000 m² warehouse. Embodied carbon factor for tilt-up concrete warehouse: ~700 kg CO2e/m². Total: 7,000 tCO2e — all reported in the construction year.

Pitfall: Easy to forget completely. Then a single year's capital investment can spike Scope 3 dramatically. Don't be surprised; flag it in narrative.

CATEGORY 03 · UPSTREAM

Fuel- and energy-related activities

Emissions from producing the fuels and electricity you buy — the upstream piece that doesn't fall into Scope 1 or 2. Includes 'well-to-tank' fuel emissions, electricity T&D losses, and trading of energy.

✓ In scopeWell-to-tank (WTT) emissions of fuels burned in Scope 1. Upstream emissions of electricity in Scope 2. Transmission & distribution losses. Purchased electricity not consumed (e.g. resale).
× Not hereCombustion at your facility (Scope 1). The electricity itself once it crosses your meter (Scope 2).

How to calculate: Apply WTT and T&D loss factors to Scope 1 / Scope 2 consumption volumes. UK BEIS publishes annual factors; US EPA & DEFRA do the same.

Worked example: A UK office uses 500,000 kWh grid electricity (Scope 2 = ~107 tCO2e at 2025 UK grid factor). Add T&D losses of ~8% and WTT of the generation fuels: roughly 25–30 tCO2e additional in Cat 3.

Pitfall: The double-count trap. If you used spend-based for Category 1 fuel purchases, you've already implicitly included WTT — adding Cat 3 on top double-counts. Activity-based Cat 1 and a separate Cat 3 is fine; spend-based Cat 1 and a separate Cat 3 is not.

CATEGORY 04 · UPSTREAM

Upstream transportation and distribution

Emissions from moving purchased goods from suppliers to your sites, plus any third-party logistics services you paid for.

✓ In scopeInbound freight (road, sea, air, rail). Warehousing and distribution paid for by you. Logistics services purchased.
× Not hereTransport in your own / leased vehicles (Scope 1). Transport between supplier production sites (already in the supplier's Cat 4 or upstream of Cat 1).

How to calculate: Tonne-kilometres × mode-specific emissions factor (kg CO2e per tonne-km for road / sea / air / rail). Activity-based factors are widely available from GLEC / SmartWay / BEIS.

Worked example: 2,000 tonnes of components shipped 8,000 km by sea = 16,000,000 tonne-km × 0.012 kg CO2e/tonne-km = 192 tCO2e. Same volume shipped by air = ~10,000+ tCO2e. Mode matters enormously.

Pitfall: Same double-count risk as Cat 3 if Cat 1 was spend-based. Also: lots of teams forget the warehousing component — pure mileage isn't the full story.

CATEGORY 05 · UPSTREAM

Waste generated in operations

Emissions from disposal and treatment of waste produced by your own facilities. Landfill methane, incineration CO2, wastewater treatment.

✓ In scopeAll solid and liquid waste streams generated at your facilities: general waste, recycling, hazardous, food waste, wastewater. Includes transport to disposal site.
× Not hereCustomer or downstream waste (Cat 12). Waste from leased assets where the lessor pays.

How to calculate: Mass of waste by stream × treatment-specific factor. Recycling typically has a low net factor (sometimes negative on credit basis); landfill has the highest.

Worked example: 500 tonnes mixed commercial waste to landfill × 467 kg CO2e/tonne (DEFRA 2025) = 234 tCO2e. Divert to recycling: ~21 kg CO2e/tonne = 11 tCO2e. Tenfold reduction for free if your waste contractor changes.

Pitfall: Smaller than people expect for most companies (often <1% of Scope 3). Don't over-invest in measurement here unless you're in heavy industry or chemicals.

CATEGORY 06 · UPSTREAM

Business travel

Emissions from employees travelling for business in vehicles your company doesn't own: flights, rail, hire cars, taxis. Hotel nights typically included.

✓ In scopeFlights, train tickets, hire cars, taxis, hotel stays. RFI (radiative forcing index) typically applied to flights for total climate impact.
× Not hereTravel in company-owned/leased vehicles (Scope 1). Commuting (Cat 7).

How to calculate: Flights: passenger-km × class-and-haul factor (short / long, economy / business / first). Hotels: nights × country-specific factor (DEFRA publishes). Hire car: actual mileage.

Worked example: One London-New York return in economy: ~2,500 kg CO2e per passenger (with RFI). The same flight in business: ~4,500 kg. Five executives travelling business class: ~22.5 tCO2e for one trip.

Pitfall: Travel data lives in TMC systems, expense reports and credit card feeds. Get the integration right once and this category becomes easy. The RFI factor (whether to apply, typically 1.7–2.0×) is a methodology choice you should declare.

CATEGORY 07 · UPSTREAM

Employee commuting

Emissions from employees travelling between home and their work location. Increasingly includes home-working energy use since 2020.

✓ In scopeAll commuting modes: car, bus, train, walking/cycling (zero). Increasingly includes WFH energy — GHG Protocol Technical Guidance addendum covers this since 2022.
× Not hereBusiness travel (Cat 6). Travel in company cars (Scope 1 / Cat 1 depending on ownership).

How to calculate: Survey employees on commute pattern, mode, distance — usually quarterly or annually. Then headcount × average commute × mode factor. WFH: hours WFH × country-grid kWh per hour assumption.

Worked example: 500 employees, average 20 km commute round-trip, 220 working days, 60% car / 40% public transport. Car: 500 × 0.6 × 20 × 220 × 0.17 kg/km = 224 tCO2e. Public transport: 500 × 0.4 × 20 × 220 × 0.04 = 35 tCO2e. Total: ~260 tCO2e/year.

Pitfall: Survey fatigue. Sample carefully and document assumptions. WFH is a relatively new add — check current GHG Protocol guidance on Homeworking before including.

CATEGORY 08 · UPSTREAM

Upstream leased assets

Emissions from operating assets your company leases (you're the lessee) that aren't already in your Scope 1 or 2.

✓ In scopeLeased space where the landlord pays utilities (so they fall outside your Scope 1/2). Leased vehicles where you don't have operational control. Leased equipment.
× Not hereLeased assets already in your Scope 1/2 under operational control approach. Capital lease arrangements you treat as owned.

How to calculate: Floor area × building energy intensity factor; or vehicle miles × emissions factor. The judgement call is the consolidation approach (operational control vs equity vs financial control) — declare it.

Worked example: A consultancy occupies 5 leased floors where the landlord runs HVAC and lighting from a single building meter. Estimate the proportional share: 5 floors of 40 × 5,000 m² building × 120 kWh/m²/year × UK grid factor = roughly 32 tCO2e/year.

Pitfall: Easy to either double-count with Scope 2 or miss entirely. Establish your consolidation approach first, then assign every leased asset to either Scope 1/2 or Cat 8.

Downstream categories (9–15)

Everything that happens to your products after they leave you. Distribution, processing, customer use, end-of-life, plus assets you lease to others and your investments.

CATEGORY 09 · DOWNSTREAM

Downstream transportation and distribution

Emissions from transporting sold products to customers and through distribution channels, where the customer (not you) is paying for the logistics.

✓ In scopeLast-mile delivery, retail distribution, customer-paid freight. Storage at distributors / retailers.
× Not hereOutbound transport you paid for (already in Cat 4 or your Scope 1). Customer travel to a retail store (not a Scope 3 category).

How to calculate: Tonne-km × mode factor, same as Cat 4. Often estimated based on average product weight, average customer location, and standard delivery routes.

Worked example: An e-commerce retailer ships 1M parcels/year, average 2 kg, average 200 km last-mile by van. 1M × 0.002 t × 200 km × 0.17 kg/tkm = 68 tCO2e/year.

Pitfall: Boundary with Cat 4 is the most-asked question in Scope 3. Rule: who pays for the freight? Pays you / paid by you = Cat 4. Pays customer / handled by reseller = Cat 9.

CATEGORY 10 · DOWNSTREAM

Processing of sold products

Emissions from further processing of intermediate products you sell — e.g. you sell steel coil, your customer rolls and stamps it into car panels. Their processing emissions are yours under Cat 10.

✓ In scopeAll third-party manufacturing operations on intermediate products you sell, before they become a final consumer good.
× Not hereUse phase of final products (Cat 11). Companies selling only finished goods don't have a Cat 10.

How to calculate: Volume of intermediate sold × typical downstream processing energy × grid factor. Often heavily assumption-driven for upstream suppliers.

Worked example: A chemicals manufacturer sells 100,000 tonnes of intermediate polymer to plastics converters. Average converter energy: ~1.5 GJ/tonne × 100,000 = 150,000 GJ × grid factor ≈ 8,300 tCO2e.

Pitfall: Almost no real data; estimate heavily and document. Sometimes excluded if your products are mostly finished goods.

CATEGORY 11 · DOWNSTREAM

Use of sold products

Emissions from customers using your products over their lifetime. Massive for energy-using goods (cars, white goods, software) and combusted products (fuels, lubricants).

✓ In scopeDirect use-phase emissions (electricity to run the device, fuel burned in a vehicle, methane leaks from sold equipment). Indirect (i.e. expected use) use phase per GHG Protocol guidance.
× Not hereEnd-of-life disposal (Cat 12). Service / maintenance you don't sell separately.

How to calculate: Units sold × lifetime use × per-unit use emissions. For cars: km driven over life × gCO2/km. For appliances: kWh/year × lifetime years × grid factor. For fuels: tonne sold × combustion factor.

Worked example: A car maker sells 50,000 vehicles averaging 110 gCO2/km. Average lifetime 200,000 km. Each vehicle: 22 tCO2e. Fleet: 1.1M tCO2e for one year's sales. Bigger than the entire rest of the company's footprint combined.

Pitfall: Often the dominant category but completely ignored by companies that don't sell hardware. If you sell electrified products, this is the category to obsess over.

CATEGORY 12 · DOWNSTREAM

End-of-life treatment of sold products

Emissions from the eventual disposal of the products you sell. Landfill, incineration, recycling, composting — all attributed to the original seller.

✓ In scopeAll end-of-life pathways for sold products, weighted by typical real-world disposal mix in the regions you sell into.
× Not hereOperational waste (Cat 5). Service / refurbishment emissions (Cat 11).

How to calculate: Mass of product sold × expected disposal mix × treatment-specific factor. Packaging is a major sub-component — track separately if material.

Worked example: A consumer goods company sells 10,000 tonnes/year of plastic-packaged product (8% packaging by weight). Packaging EoL: 800 t plastic × 40% landfill × 1.4 tCO2e/t = ~450 tCO2e (packaging EoL alone).

Pitfall: Often confused with Cat 5 (your own operational waste). Cat 5 is what you throw out from your sites. Cat 12 is what customers throw out at the end of using your products.

CATEGORY 13 · DOWNSTREAM

Downstream leased assets

Emissions from operating assets your company owns and leases to others. The flip side of Cat 8.

✓ In scopeOffice buildings or warehouses you own and lease to tenants. Vehicles or equipment you own and lease out. Solar panels owned by you and operated by a customer.
× Not hereAssets in Scope 1 / 2 (you have operational control). Sold products (Cat 11).

How to calculate: Floor area × tenant energy intensity, or asset operating hours × emissions factor. For commercial real estate, GRESB asset-level energy data is the standard source.

Worked example: A REIT owns 50,000 m² of leased office space at 150 kWh/m²/year energy intensity. UK grid factor: 7.5M kWh × 0.214 = 1,605 tCO2e/year.

Pitfall: Real estate's dominant Scope 3 category. Don't confuse with embodied carbon (Cat 1 / Cat 2).

CATEGORY 14 · DOWNSTREAM

Franchises

Emissions from operating franchise locations not consolidated into Scope 1 / 2 under your chosen consolidation approach.

✓ In scopeOperational emissions of franchisees (energy, refrigerants, vehicles) where you are the franchisor and they're independently operated.
× Not hereWholly-owned outlets (Scope 1 / 2). Franchisees' own purchases (their Cat 1).

How to calculate: Per-outlet energy estimate × number of franchises × grid factor. McDonald's, Subway, KFC and similar franchisors lead reporting practice here.

Worked example: A coffee chain franchises 500 outlets, each ~40,000 kWh/year. Total: 20M kWh × UK grid factor = ~4,280 tCO2e/year.

Pitfall: Only applies if you franchise. Many companies skip it correctly. If you do franchise, the data has to come from franchisees — build the supplier engagement workflow.

CATEGORY 15 · DOWNSTREAM

Investments

Emissions associated with your company's investments — equity, debt, project finance, managed assets. The 'financed emissions' of banks, asset managers and private equity firms.

✓ In scopeEquity investments, debt investments, project finance, managed assets, client services. Allocated by share of ownership / financing.
× Not hereWholly-owned subsidiaries (consolidated into Scope 1/2). Pension fund assets owned by your employees rather than your company.

How to calculate: Follow PCAF (Partnership for Carbon Accounting Financials) methodology. Investee company emissions × (your investment ÷ investee enterprise value). PCAF defines six asset-class methodologies (listed equity, business loans, project finance, commercial real estate, mortgages, motor vehicle loans).

Worked example: A bank has £100M loan to an industrial company with £1B enterprise value emitting 500,000 tCO2e. Attribution: (100M/1,000M) × 500,000 = 50,000 tCO2e financed by the bank, in Cat 15.

Pitfall: Cat 15 is essentially the entire Scope 3 footprint of financial institutions. PCAF data quality scores (1=best, 5=worst) should be disclosed alongside the numbers — transparency about methodology beats false precision.

The calculation methods ladder

Four methods. Climb the ladder where the category is material; use the bottom rung for categories that aren't.

RUNG 4
Spend-based

£ spent × industry-average kg CO2e per £.

Use when: first-year inventory, immaterial categories, no other data available.

RUNG 3
Activity-based (average)

Physical units (kg, kWh, tonne-km) × product-class average factor.

Use when: you know physical volumes but not supplier-specific data.

RUNG 2
Hybrid

Supplier-specific for top-spend lines, average data for the rest.

Use when: you've engaged your biggest suppliers but can't get them all.

RUNG 1 · BEST
Supplier-specific

Actual product- and supplier-level emissions data from each vendor.

Use when: the category is material and assurance / SBTi rigour is needed.

Six common Scope 3 mistakes

  1. Double-counting Categories 1, 3 and 4. Spend-based Cat 1 already includes upstream fuel/energy and transport. Adding Cat 3 and Cat 4 separately overstates the footprint.
  2. Treating Cat 5 (operational waste) as Cat 12 (sold-product EoL). Cat 5 is what your sites throw away. Cat 12 is what customers throw away after using your products. Different denominators, different factors.
  3. Ignoring Category 11 because you "don't sell hardware". Software, services and SaaS still have a use phase if customers run them on infrastructure you sold or specified.
  4. Spend-based forever. First inventory year, fine. Year three, you're being lazy. SBTi and CSRD assurance increasingly demand activity-based for material categories.
  5. RFI for flights left unstated. Multiplying short-haul air emissions by 1.7–2.0× (radiative forcing) is a methodology choice that materially changes Cat 6. Declare it.
  6. Annualising capital goods. Cat 2 is reported in the year of purchase, full amount, no amortisation. Many teams instinctively spread it over the asset's life — that's not GHG Protocol-compliant.

Materiality screening: where to focus

Before climbing the methods ladder, screen which categories actually matter. A one-week screening exercise saves months of effort. Three steps:

  1. Apply spend-based factors to all 15 categories. Don't overthink — even imprecise spend-based numbers reveal proportions.
  2. Rank by absolute tCO2e. The top 3–5 categories typically cover 80%+ of total Scope 3.
  3. Invest measurement effort proportionally. Activity-based for top categories. Hybrid for next tier. Stay spend-based for the long tail.

SBTi rule of thumb: If Scope 3 is >40% of your total footprint, SBTi requires a Scope 3 reduction target. Most companies cross that threshold easily once they include Cat 1 and Cat 11.

FAQs

How many Scope 3 categories are there?

Fifteen, defined by the GHG Protocol Corporate Value Chain (Scope 3) Standard. Categories 1–8 are upstream (suppliers and inputs); categories 9–15 are downstream (sold products and investments).

Which Scope 3 categories are typically largest?

It depends on the industry. Category 1 (Purchased goods and services) is largest for most companies. Category 11 (Use of sold products) dominates for energy-using product makers — cars, appliances, software. Category 15 (Investments) is largest for banks and asset managers. Category 4 (Upstream transport) is largest for logistics and distribution.

Do I have to report all 15 Scope 3 categories?

Under CSRD/ESRS E1 and ISSB IFRS S2, you must report every material category and explain any you exclude. Under the GHG Protocol Standard itself, reporting is voluntary but should be complete or justified. In practice, do a materiality screen first — usually 3–5 categories dominate.

What's the difference between spend-based and activity-based Scope 3?

Spend-based multiplies pounds/dollars spent by an industry-average emissions factor. Quick to do, low accuracy. Activity-based uses physical units — kg, kWh, tonnes-km — multiplied by a unit-level emissions factor. More accurate. Supplier-specific is the gold standard.

What's the most common Scope 3 mistake?

Double-counting between Category 1 and Categories 3–4. Most companies who use spend-based factors for Category 1 are already implicitly including transport and energy. Adding Categories 3 and 4 on top is double-count.

How long does a Scope 3 inventory take to build?

A first-cut spend-based inventory across all 15 categories takes 2–4 weeks. Moving to activity-based and supplier-specific data is the multi-year journey. Most CSRD-ready companies are in year 2 or 3 of that journey.

Does ESG:ONE calculate all 15 Scope 3 categories?

Yes. We support all 15 GHG Protocol categories with the appropriate methods per category. The library includes 140,000+ emissions factors including DEFRA, EPA, Ecoinvent, EXIOBASE and BEIS.

Stop chasing Scope 3 in spreadsheets.

ESG:ONE automates collection across all 15 categories — supplier portals, AI extraction from utility bills and invoices, 140,000+ emissions factors, full audit trail. Live in 3 weeks.

Related guides

CSRD 2026 requirements →

Post-Omnibus scope, deadlines and ESRS checklist.

Measuring ESG financial impact →

Methods for P&L, cost of capital and valuation effects.

12 ESG platforms compared →

Honest comparison of the leading ESG reporting platforms.

We use cookies to enhance your experience. By continuing to visit this site you agree to our use of cookies. Learn more