I buy from 100 suppliers. Can I collect real data from the biggest few and estimate the rest, the way the Greenhouse Gas Protocol suggests?

The Greenhouse Gas (GHG) Protocol's coffee example does this. 10 of 100 suppliers account for 85 percent of the beans, so the company may use supplier data for those 10 and may extrapolate to 100 percent based on the 85 percent. Smaller buys such as sugar can use average-data or spend-based factors.

What the guide's coffee example does

A coffee company bought beans from 100 suppliers. Ten of them account for 85 percent of the quantity. The guide says the company may calculate emissions for those ten with the supplier-specific or hybrid method, and may then extrapolate to 100 percent based on the 85 percent covered.

It spent 20 million United States dollars (USD) on beans and USD 1 million on sugar. Because sugar is a small share of spend, the guide says the company may choose not to engage the sugar suppliers and use secondary factors, through the average-data or spend-based method.

The practical approaches the guide lists

• More accurate data and calculations for large contributors.
• Less accurate data and calculations for small contributors.
• Grouping or combining similar activity data, such as goods and services.
• Taking data from a representative sample and extrapolating to the whole.
• Using proxy techniques.

Steps for a buyer in India

1. Total the tonnes or rupees bought from each supplier over the year, from purchase invoices.
2. Rank suppliers and add up their cumulative share. The guide gives no cut-off; 85 percent is only its example.
3. Ask the top suppliers for product or hybrid data, and keep the ones that reply as measured.
4. Scale the measured emissions to the whole category, and say in your method note that the rest is extrapolated.

The caution that comes with scaling

Scaling assumes the suppliers you did not measure look like the ones you did (our reading). If the big suppliers are cleaner or dirtier than the long tail, the scaled figure is off by that difference. Say which share of the category is measured and which is extrapolated.

Scaling the measured share, with example numbers

• Bought in the year: 4,000 tonnes. Top 10 suppliers: 85 percent, so 4,000 x 0.85 = 3,400 tonnes
• Their measured emissions, from supplier data: 1,700 tonnes of carbon dioxide equivalent (CO2e), which is 1,700 / 3,400 = 0.5 tonnes per tonne bought
• Scaled to 100 percent: 1,700 / 0.85 = 2,000 tonnes CO2e, the same as adding 600 tonnes x 0.5 = 300 tonnes for the unmeasured part
• Spend check from the guide's figures: sugar USD 1 million of USD 21 million total, about 4.8 percent

The 4,000 tonnes and 1,700 tonnes CO2e are our example numbers. The 85 percent, USD 20 million and USD 1 million come from the guide's example.

What Senseible does here

Senseible reads fuel invoices and electricity bills into monthly records. It does not rank your suppliers or scale supplier figures; do that from your purchase ledger as above.

• Filling gaps in Scope 3 data
• Sampling methods for Scope 3
• Category 1 method order
• Scope 3 supplier evidence

Sources: Greenhouse Gas Protocol Technical Guidance for Calculating Scope 3 Emissions, Introduction ("Using a combination of calculation methods")

Last reviewed 2026-11-05. The 85 percent and the spend figures are from the guide's example. Our example numbers are labelled.