Does the Greenhouse Gas (GHG) Protocol say a year-on-year change of over 10 percent in activity data needs checking?

Only as an example. The GHG Protocol Corporate Standard says to compare current-year activity data with historical trends and investigate patterns that are not consistent, adding that changes of over 10 percent from year to year may warrant further investigation. It is a prompt to look, not a limit.

The wording

It sits in the standard's list of measures for the quality of activity data: compare current year data with historical trends. If data do not exhibit relatively consistent changes from year to year, the causes should be investigated, and changes of over 10 percent may warrant further investigation. The word is "may", and the 10 percent is given as an example.

Other checks in the same list

• Compare activity data from several reference sources, for example government survey data or trade association data, with your own, and compare facilities with each other.
• Check that base year recalculation procedures and boundary decisions were applied consistently.
• Investigate whether biases are known, for example operations at smaller sites left out unintentionally.
• Convert fuel consumption to energy units before applying carbon content factors.
• Set up collection so the same data can be gathered efficiently in future years.

Events the standard says to describe

When facilities report to head office, the standard asks them to describe events and changes that affect reported data: acquisitions, divestitures, closures, technology upgrades, and changes of reporting boundary or calculation method. These are the first places to look when a figure moves.

Compare quantities first, then emissions

My reading: test the activity data, such as litres and kilowatt hours, before the emissions. A new grid factor or fuel factor moves emissions with no change in activity, and a change in activity can be hidden by a change in factor. The example below shows the second.

My list of usual causes, not the standard's: an unrecorded bill or a duplicate, an estimated bill trued up later, a unit slip such as litres against kilolitres, a production change, and a month that falls into the wrong year.

Two figures, one over 10 percent and one under

• Diesel: 48,000 litres in financial year (FY) 2024-25 and 53,760 in FY 2025-26. Change = 5,760 / 48,000 = 12.0%, over 10%, so look at it
• Emissions at 2.645 kilograms (kg) of carbon dioxide (CO2) per litre: 126.96 tonnes and 142.20 tonnes, up 15.24 tonnes
• Electricity: 480,000 kilowatt hours (kWh) and 523,200. Change = 43,200 / 480,000 = 9.0%, under 10%, so this test does not flag it
• Emissions with each year's grid factor: 480,000 x 0.710 = 340.80 tonnes, and 523,200 x 0.675 = 353.16 tonnes, up 3.6%
• Electricity quantity rose 9.0% but emissions only 3.6%, because the grid factor fell from 0.710 to 0.675

Quantities are example numbers. Ten percent is the standard's example. Set your own trigger, write it down and apply it the same way every year.

What Senseible does here

Senseible reads fuel invoices and electricity bills into monthly records, so the months are on file to set against last year's before a verifier or buyer does.

• Verifier asks why emissions changed
• Verifier asks why a month looks unusual
• Internal review before external verification
• Past years when the grid factor changes
• Estimated bill trued up next year

Sources: GHG Protocol Corporate Accounting and Reporting Standard (revised edition), chapters 6 and 7 (activity data quality measures, reporting from facilities), GHG Protocol Corporate Standard (ghgprotocol.org), CEA CO2 Baseline Database, grid factors for FY 2024-25 and FY 2025-26 (Central Electricity Authority)

Last reviewed 2026-11-05. The 10 percent is the standard's example, not a rule. Use the grid factor for the year of each bill.