# A buyer wants product-level emissions from us but we only have factory bills. Which level of data does the Greenhouse Gas (GHG) Protocol Scope 3 Standard ask a supplier for?

Canonical: https://senseible.earth/answers/supplier-data-level-product-facility-or-corporate

The Scope 3 Standard ranks five levels, most specific first: product, activity or production line, facility, business unit, corporate. It asks for the most product-specific data available, and says data above product level may need allocation. A diversified supplier should give more granular data than a uniform one.

## The five levels, most specific first

1. Product level: cradle-to-gate emissions for the product of interest.
2. Activity, process or production line level: emissions or activity data for the activities, processes or lines that make the product.
3. Facility level: emissions or activity data for the facilities or operations that make the product.
4. Business unit level: for the business units that make the product.
5. Corporate level: for the entire corporation.

## The rule for choosing

When collecting primary data from partners, companies should obtain the most product-specific data available. Product-level data is more precise because it relates to the good or service purchased and avoids the need for allocation. If it is not available, suppliers should try to provide activity, process or production line data. If that is not available, facility data, and so on down the list.

Data at activity, production line, facility, business unit or corporate level may require allocation. The standard allows either the buyer or the supplier to do the allocating.

## When corporate-level data is acceptable

The standard's Box 7.4 compares two suppliers. The first only produces professional services across its units, with similar emissions throughout. The buyer decides corporate-level data will reflect the service it buys. The second has one unit in professional services and another in manufacturing, with widely different intensities. The buyer decides not to use corporate-level data and asks for facility or business unit data. The buyer makes a qualitative call from the nature of the supplier's activities.

## Why allocation is the weak spot

- The standard says allocation adds uncertainty and may be especially inaccurate when a facility makes a wide variety of products that differ a lot in their emissions. It should be used only when more accurate data is not available.
- Ways to avoid it: product-level data from partners, separate sub-metering of energy use, for example per production line, or engineering models for each product.
- No allocation is needed if the facility produces only one output, or if the emissions of each output are quantified separately.

## What a plant can do now

1. One product: your facility figure divided by units made gives plant emissions per unit with no allocation. The standard's product level is cradle-to-gate, which also counts what your inputs carried in, so tell the buyer what your figure includes.
2. Several products: send the facility total with production by product, and ask which allocation the buyer uses. The standard lists the allocation method used, or the information the buyer needs to allocate, among the items a request may ask for. This is my reading of how to use it.
3. Keep the bills behind each facility figure. They are the evidence at any level.

## A one-product plant at facility level

- Electricity: 150,000 kilowatt hours (kWh) x 0.675 kilograms (kg) of carbon dioxide (CO2) per kWh (financial year (FY) 2025-26 grid factor) = 101,250 kg
- Diesel: 20,000 litres x 2.645 kg CO2 per litre = 52,900 kg
- Fuel and power total: 101,250 + 52,900 = 154,150 kg, or 154.15 tonnes
- One product, 1,000 tonnes made: 154.15 / 1,000 = 0.154 tonnes of CO2 per tonne for fuel and power at the plant, with no allocation needed

Quantities are example numbers. If the same plant made two products, the 154.15 tonnes would need sub-metering or an allocation key, and the standard's decision tree tries sub-metering and engineering models before physical keys.

## What Senseible does here

Senseible reads fuel invoices and electricity bills into monthly records, which gives the facility-level figure and the bills behind it. Splitting a facility across products is not something those records do on their own.

- [One set of figures for several buyers](https://senseible.earth/answers/one-carbon-dataset-for-several-buyers)
- [Buyer sends its own questionnaire format](https://senseible.earth/answers/buyer-sends-own-carbon-questionnaire-format)
- [Buyer wants one company number for factory and trading arm](https://senseible.earth/answers/buyer-wants-one-company-number-factory-and-trading-arm)
- [Free calculators](https://senseible.earth/calculators)

Sources: [GHG Protocol Corporate Value Chain (Scope 3) Accounting and Reporting Standard, chapters 7 and 8 (Table 7.7, Box 7.4, sections 8.1 to 8.3)](https://ghgprotocol.org/sites/default/files/standards/Corporate-Value-Chain-Accounting-Reporing-Standard_041613_2.pdf), [CEA CO2 Baseline Database, FY 2025-26 grid factor (Central Electricity Authority)](https://cea.nic.in/cdm-co2-baseline-view?mappingId=234&lang=en)

Last reviewed 2026-11-05. Levels and wording are from the 2011 Scope 3 Standard. Which level a buyer accepts is the buyer's call, so ask before you spend time splitting a facility figure.