Electricity Carbon Footprint: Calculate a Defensible Result
A synthetic editorial image shows the back of a person holding a clipboard beside multiple meters, cabling and switches in an open panel. Through large windows are an office building and transmission towers. No legible customer name, meter serial, consumption quantity, emissions result or certificate appears. The retained scene supplies commercial metering context; it is not proof of an actual installation, a safe-work instruction, a measured site result or an approved contractual electricity claim.
Diagram text description
An electricity carbon footprint attributes emissions to a defined amount of electricity use. For a company's purchased electricity, the basic calculation is activity multiplied by a suitable emissions factor. The difficult part is establishing what the activity covers, whether the factor answers the right question and how much uncertainty remains in the records.
A defensible result connects four records: the site's electricity activity, the selected factor and method, the unit conversion, and the reported emissions. If a reviewer cannot follow those connections, a precise-looking number is not enough. A bill total can contain a duplicate; a meter can span the wrong dates; a factor can describe a different grid or an avoided-generation scenario.
This guide provides a reusable worksheet and a filled fictional site example. It reconciles overlapping bills, a meter replacement and a missing interval before calculating emissions. It also distinguishes a location-based result from a market-based view of the same load, without pretending that an unreviewed contract proves a lower footprint.
Begin with the electricity boundary
Start by identifying the reporting organisation, facility and electricity stream. Purchased electricity consumed within the applicable organisational boundary normally contributes indirect Scope 2 emissions. That is not the same as the site's direct fuel combustion, upstream fuel production or every other activity in its value chain.
Keep grid imports, onsite generation, self-consumption and exports in separate records. A grid-import meter already measures purchased electricity entering the site; subtracting rooftop solar generation from that reading can remove electricity that was never included. Exported electricity is not an automatic credit against the emissions attributed to purchased electricity either.
Record who owns or controls the generation equipment and how contractual attributes are treated. Those details can affect the wider inventory classification and the claims available. Do not use the simplified purchased-electricity worksheet to settle every leased-asset, generation or certificate question.
The result developed here is therefore deliberately narrow: purchased electricity attributed to a specified site and period under a stated method. Direct fuels, transmission and distribution losses where accounted for separately, other value-chain activities and a product's full life cycle remain outside this one calculation. They still need consideration in the relevant wider assessment.
Collect the right kWh data
Use one worksheet record for each unique meter interval, then reconcile it to bills and reporting dates. Store cumulative register readings separately from consumption: a register showing 105,000 kWh does not mean the facility used that much during the reporting period. Check multipliers, meter changes and whether readings are actual or estimated before calculating differences.
| Worksheet group | Fields to retain | Review decision |
|---|---|---|
| Identity and period | Site, meter serial, stream, interval start and end | Does this interval belong to this site and reporting year? |
| Reading evidence | Opening and closing registers, unit, multiplier, actual/estimated status | Can consumption be derived without crossing a reset or replacement? |
| Billing evidence | Invoice, original/correction reference, service dates, billed kWh or MWh | Is this new consumption, a duplicate or a replacement bill? |
| Coverage and adjustment | Gap dates, estimate basis, overlap removal, exclusions | What is measured, estimated, unresolved or routed elsewhere? |
| Generation and exports | Separate generation, self-use and export references | Were these already outside the purchased-import reading? |
| Factor and result | Method, region, data year, file/version, gases, factor unit, conversion, result | Can another reviewer reproduce the same attributed emissions? |
| Ownership | Preparer, reviewer, version and open decision | What evidence must be resolved before the figure is final? |
Table 1, shown one row at a time. Every field is retained.
Row 1
- Worksheet group
- Identity and period
- Fields to retain
- Site, meter serial, stream, interval start and end
- Review decision
- Does this interval belong to this site and reporting year?
Row 2
- Worksheet group
- Reading evidence
- Fields to retain
- Opening and closing registers, unit, multiplier, actual/estimated status
- Review decision
- Can consumption be derived without crossing a reset or replacement?
Row 3
- Worksheet group
- Billing evidence
- Fields to retain
- Invoice, original/correction reference, service dates, billed kWh or MWh
- Review decision
- Is this new consumption, a duplicate or a replacement bill?
Row 4
- Worksheet group
- Coverage and adjustment
- Fields to retain
- Gap dates, estimate basis, overlap removal, exclusions
- Review decision
- What is measured, estimated, unresolved or routed elsewhere?
Row 5
- Worksheet group
- Generation and exports
- Fields to retain
- Separate generation, self-use and export references
- Review decision
- Were these already outside the purchased-import reading?
Row 6
- Worksheet group
- Factor and result
- Fields to retain
- Method, region, data year, file/version, gases, factor unit, conversion, result
- Review decision
- Can another reviewer reproduce the same attributed emissions?
Row 7
- Worksheet group
- Ownership
- Fields to retain
- Preparer, reviewer, version and open decision
- Review decision
- What evidence must be resolved before the figure is final?
Copy those field groups into a workbook or controlled register. Preserve the original invoice and meter identifiers rather than replacing them with a total alone. The annual record should explain why each source was included once, excluded as a duplicate or retained as an estimate.
Fictional worksheet sample: record E-24 covers Kestrel Depot, site S-04, for calendar 2025. Import meter IMP-A opened at 100,000 kWh on 1 January and reached 105,000 on 1 July, giving 5,000 kWh. It reached 109,000 on 1 November, adding 4,000 kWh. Its final reading before replacement at the end of November is missing. New meter IMP-B opened at 5,000 on 1 December and closed at 7,200 on 1 January 2026, adding 2,200 kWh for December. All stated readings are invented actual-reading entries in the example, not measurements of a real facility.
For the invented calendar-2025 E-24 record, IMP-A supports 5,000 kWh for January to June and 4,000 for July to October. November uses an unresolved 800 kWh estimate. IMP-B supports 2,200 kWh for December using its own register difference. The drawing is a schematic period sequence, not a quantity-scaled chart. Never subtract the new serial’s opening reading from the old meter. The annual 12,000 kWh remains provisional.
Diagram text description
| E-24 reconciliation field | Filled worksheet value |
|---|---|
| Source bills | INV-01: 5,000 kWh; COPY-01: the same INV-01; INV-02: 4,000 kWh; INV-03: 2,200 kWh |
| Duplicate adjustment | Remove COPY-01's 5,000 kWh from the numerical sum; retain its identity and exclusion reason |
| Unique supported consumption | 5,000 + 4,000 + 2,200 = 11,200 kWh |
| November gap | 800 kWh provisional estimate, based only on the prior year's November record |
| Annual working quantity | 11,200 supported + 800 estimated = 12,000 kWh, or 12 MWh |
| Separate onsite stream | 1,000 kWh generated; 700 self-consumed; 300 exported; none subtracted from import-meter activity |
| Factor record | Fictional F-ILL-25: location-style generation CO2e, 0.40 kg per kWh; 400 kg per MWh |
| Reviewer and status | Rhea prepares; Omar reviews; version V1 is provisional because the missing interval is unresolved |
Table 2, shown one row at a time. Every field is retained.
Row 1
- E-24 reconciliation field
- Source bills
- Filled worksheet value
- INV-01: 5,000 kWh; COPY-01: the same INV-01; INV-02: 4,000 kWh; INV-03: 2,200 kWh
Row 2
- E-24 reconciliation field
- Duplicate adjustment
- Filled worksheet value
- Remove COPY-01's 5,000 kWh from the numerical sum; retain its identity and exclusion reason
Row 3
- E-24 reconciliation field
- Unique supported consumption
- Filled worksheet value
- 5,000 + 4,000 + 2,200 = 11,200 kWh
Row 4
- E-24 reconciliation field
- November gap
- Filled worksheet value
- 800 kWh provisional estimate, based only on the prior year's November record
Row 5
- E-24 reconciliation field
- Annual working quantity
- Filled worksheet value
- 11,200 supported + 800 estimated = 12,000 kWh, or 12 MWh
Row 6
- E-24 reconciliation field
- Separate onsite stream
- Filled worksheet value
- 1,000 kWh generated; 700 self-consumed; 300 exported; none subtracted from import-meter activity
Row 7
- E-24 reconciliation field
- Factor record
- Filled worksheet value
- Fictional F-ILL-25: location-style generation CO2e, 0.40 kg per kWh; 400 kg per MWh
Row 8
- E-24 reconciliation field
- Reviewer and status
- Filled worksheet value
- Rhea prepares; Omar reviews; version V1 is provisional because the missing interval is unresolved
The unadjusted bill sum is 16,200 kWh, not the annual purchased total. Removing the repeated 5,000 gives 11,200; the explicitly added 800 estimate gives the working 12,000. Its illustrative factor therefore produces 4,800 kg, or 4.8 metric tonnes CO2e, subject to the gap qualification explained below.
Open decision: obtain the missing IMP-A final reading or a supported replacement-period consumption record and test the November estimate against actual operating conditions. Consequence: keep E-24 provisional, disclose the estimated interval and revise both energy and emissions if better evidence changes it; do not label the annual quantity fully metered.
Choose a factor for the right place and year
Select a factor by geography, period, method and emissions boundary, not by whether its value looks favourable. A national average from one country is not a substitute for a site's appropriate regional grid factor somewhere else. Where a regional factor is suitable, retain how the site was assigned to that region.
Distinguish publication year from the year of the underlying electricity data. A file released this year can describe an earlier generation mix. Keep both dates and use the version appropriate to the reporting framework and analysis, explaining any unavoidable mismatch rather than quietly presenting it as an exact match.
Read the factor's unit and gas coverage. A CO2-only rate cannot simply be renamed CO2e; a generation factor does not automatically include upstream fuel production or transmission losses. Decide where those other emissions belong before combining datasets that use different boundaries.
The worksheet should retain the publisher, document or dataset version, exact factor row and original unit. If a preparer cannot establish the factor's geography or purpose, escalate the selection. An unresolved factor does not become valid because the spreadsheet accepts it. F-ILL-25 in the example is a teaching coefficient, not a recommendation for any real site's reporting year.
Check factor purpose before using it
An inventory factor and an avoided-generation factor answer different questions. An attributed electricity inventory asks what generation emissions are allocated to the electricity used during the reporting period. An avoided-generation analysis asks what generation might change under a defined intervention and counterfactual.
Average total-output rates are commonly used for the first task. Nonbaseload or other marginal-analysis rates can be relevant to particular avoided-generation questions, but their suitability depends on the project and method. A nonbaseload average should not be described as a universal model of every marginal generator at every hour.
Do not swap a project factor into the purchased-electricity inventory because an efficiency project is being discussed. First calculate the inventory on its stated basis. Assess a project's counterfactual impact separately, with its own boundary, assumptions and limitations.
That distinction prevents a spreadsheet from subtracting a modelled saving from an attributed result while still presenting the remainder as straightforward electricity-use emissions. Managers may want both analyses, but the worksheet must say which question each number answers.
Calculate with units that actually match
Write the units into the calculation before multiplying. With activity in kWh and a factor in kg CO2e per kWh, the energy units cancel and the result is kilograms CO2e. Dividing that mass by 1,000 gives metric tonnes CO2e. A kilowatt, by contrast, is a measure of power; a demand reading in kW cannot replace annual energy in kWh.
For E-24, the illustrative operation is 12,000 kWh × 0.40 kg CO2e per kWh = 4,800 kg CO2e. The same operation can be written as 12 MWh × 400 kg CO2e per MWh. It is not 12 MWh multiplied by 0.40 kg per kWh without a conversion.
For fictional E-24, 12,000 kWh times 0.40 kg CO2e per kWh equals 4,800 kg CO2e. Equivalently, 12 MWh times 400 kg CO2e per MWh gives the same 4,800 kg, or 4.8 metric tonnes. These are invented generation-only teaching coefficients, not recommended real factors. Activity includes an unresolved estimate, so the result remains provisional; upstream fuel and separate transmission-loss assessment are outside this example.
Diagram text description
Where a suitable source publishes pounds per MWh, first divide kWh by 1,000, multiply the resulting MWh by that rate, then convert pounds to kilograms using 0.45359237 kg per pound. Divide kilograms by 1,000 for metric tonnes. Keep the mass basis attached: metric tonnes are not US short tons.
Retain adequate precision during these steps and round the reported result consistently at the end. A correct conversion cannot rescue an unsuitable factor, but a visible dimensional calculation makes many otherwise hidden mistakes apparent.
A fictional electricity-footprint example
Fictional filled example: E-24 is Kestrel Depot's calendar 2025 purchased-electricity reconciliation, not a real inventory or an approved emissions factor. It uses import meters IMP-A and IMP-B with a documented replacement between them. Record E-24 closes its reporting period on 31 December 2025. The working annual quantity includes an estimated November interval and must remain qualified.
| Interval and meter | Evidence and reading difference | Purchased consumption | Treatment |
|---|---|---|---|
| January–June, IMP-A | 105,000 − 100,000 kWh; INV-01 agrees | 5,000 kWh | Include once; COPY-01 is excluded as the same invoice |
| July–October, IMP-A | 109,000 − 105,000 kWh; INV-02 agrees | 4,000 kWh | Include as a distinct non-overlapping interval |
| November, IMP-A | Final replacement reading unavailable | 800 kWh estimate | Provisional prior-year November proxy, not a meter difference |
| December, IMP-B | 7,200 − 5,000 kWh; INV-03 agrees | 2,200 kWh | Use the new meter's own register, not the old serial's closing value |
| Annual working total | 11,200 supported plus 800 estimated | 12,000 kWh | Reconciled working estimate, not fully measured annual use |
Table 3, shown one row at a time. Every field is retained.
Row 1
- Interval and meter
- January–June, IMP-A
- Evidence and reading difference
- 105,000 − 100,000 kWh; INV-01 agrees
- Purchased consumption
- 5,000 kWh
- Treatment
- Include once; COPY-01 is excluded as the same invoice
Row 2
- Interval and meter
- July–October, IMP-A
- Evidence and reading difference
- 109,000 − 105,000 kWh; INV-02 agrees
- Purchased consumption
- 4,000 kWh
- Treatment
- Include as a distinct non-overlapping interval
Row 3
- Interval and meter
- November, IMP-A
- Evidence and reading difference
- Final replacement reading unavailable
- Purchased consumption
- 800 kWh estimate
- Treatment
- Provisional prior-year November proxy, not a meter difference
Row 4
- Interval and meter
- December, IMP-B
- Evidence and reading difference
- 7,200 − 5,000 kWh; INV-03 agrees
- Purchased consumption
- 2,200 kWh
- Treatment
- Use the new meter's own register, not the old serial's closing value
Row 5
- Interval and meter
- Annual working total
- Evidence and reading difference
- 11,200 supported plus 800 estimated
- Purchased consumption
- 12,000 kWh
- Treatment
- Reconciled working estimate, not fully measured annual use
The opening and closing entries are cumulative registers. The first two IMP-A differences cover consecutive periods without counting the 1 July reading twice as consumption. IMP-B starts from its own 5,000 kWh register; subtracting that reading from IMP-A's last register would create a meaningless negative amount. Preserve the serial change and the dates in the source record.
Billing reconciliation is a separate cross-check. INV-01, COPY-01, INV-02 and INV-03 sum to 16,200 kWh, but COPY-01 represents the same first-half supply as INV-01. Deduct its 5,000 kWh contribution, not the underlying six months of electricity, to obtain 11,200 kWh supported by the unique intervals. Retain the duplicate document and the decision to exclude it from the calculation.
The fictional invoice sum is 16,200 kWh because COPY-01 duplicates INV-01’s 5,000 kWh. Excluding only that repeated contribution leaves 11,200 supported kWh. The separate unresolved November estimate adds 800, giving provisional 12,000 kWh. Keep COPY-01 and its exclusion reason in the evidence record. Onsite generation, self-use and exports are recorded separately and are not subtracted from imports.
Diagram text description
The 800 kWh November proxy comes from an invented prior-year record. No one has yet checked whether operating hours, occupancy or equipment use changed. Adding it covers the interval numerically; it does not transform the estimate into an actual reading. Onsite generation of 1,000 kWh, including 700 self-consumed and 300 exported, is recorded separately. The 12,000 kWh import figure is not reduced by either quantity.
Using fictional F-ILL-25, 12,000 × 0.40 = 4,800 kg CO2e, and 4,800 ÷ 1,000 = 4.8 metric tonnes CO2e. Equivalently, 12 MWh × 400 kg CO2e per MWh gives the same mass. The coefficient assumes a location-style generation basis covering CO2e; it excludes upstream fuel production and a separate transmission-loss assessment. It is not a real grid value or an observed customer result.
The supported 11,200 kWh accounts for 4,480 kg at that illustrative rate; the estimated 800 kWh accounts for another 320 kg. This split tells a reviewer where the uncertainty enters. If later evidence establishes 1,000 kWh for November, the annual quantity would become 12,200 kWh and the same coefficient would yield 4,880 kg. That 80 kg difference would correct missing input evidence, not show a new physical increase after the year ended.
Unresolved review: Rhea must retrieve the old meter's final record or another defensible account of November consumption, and Omar must decide whether the proxy remains suitable. Reporting consequence: describe 4.8 tonnes as a provisional attributed estimate for this defined electricity stream, not a fully measured or assured figure; retain V1 and issue a traceable revision if the gap is resolved differently.
Where a market-based figure changes the answer
A location-based and a market-based result can describe the same electricity load through different allocation methods. The first uses an appropriate grid-average basis. The second follows the applicable hierarchy of contractual and other eligible information. Where dual reporting applies, keep the two results separately labelled; adding them would count the same electricity twice.
For fictional E-24, the location-style calculation uses the same provisional 12,000 kWh that appears in the worksheet. A separate market-based result has not been established. The example contains no demonstrated instrument coverage, retirement or cancellation evidence, eligible supplier-specific rate or reviewed residual-mix selection. Its market view is therefore unresolved, not zero.
Lack of a special renewable contract is not permission to omit a required market-based total indefinitely. Determine the applicable factor hierarchy, including residual-mix data and permitted fallbacks, and document the selection. A missing certificate and a missing eligible calculation basis are different problems.
One top node shows fictional E-24 provisional purchased electricity of 12,000 kWh. Two outgoing connectors lead to parallel accounting views: location, 4.8 metric tonnes CO2e using an illustrative 0.40 kg per kWh coefficient; and market, unresolved because the applicable eligible information has not been reviewed. The load contains 11,200 kWh supported by unique source intervals plus an 800 kWh missing-month estimate. These arrows do not divide the physical electricity or represent successive process stages. Where both methods apply they describe the same consumption and are reported separately, never summed. Unresolved is not a zero factor or permission to omit a required result; the reviewer must establish applicable instrument, residual-mix or permitted fallback evidence.
Diagram text description
Contractual electricity attributes are also different from carbon offsets. Do not replace an instrument-quality review with a purchase receipt for credits. This worksheet supports the evidence handoff; it does not certify a contract, establish physical electricity provenance or resolve every jurisdiction's reporting rules.
How to interpret a changing result
When the footprint changes, inspect the inputs before describing progress. Lower consumed kWh, a different grid factor, a newly included site and corrected source coverage can all move the result. A market-based change may reflect qualified contractual evidence rather than lower electricity demand.
Keep the previous dataset and factor version, then explain the comparable difference. For example, resolving E-24's missing interval would revise the historical estimate. The revised number should not be presented as electricity consumed during a later period. Similarly, changing the factor while holding the kWh constant changes attributed emissions without demonstrating that this site used less power.
For a management comparison, distinguish the reported inventory from any explicitly labelled constant-factor analysis used to understand consumption effects. Do not quietly substitute an old factor for a required reporting-year method. Document the order and assumptions of any numerical bridge.
Use the outcome to choose an action: investigate a load increase, improve a weak source, review procurement evidence or explain grid changes. A lower headline result alone cannot tell a manager which of those happened.
Common claims that a number cannot support
State the claim at the same boundary as the calculation. A qualified purchased-electricity Scope 2 result does not establish a complete company footprint, a product life-cycle footprint or zero environmental impact. The emissions from equipment, fuels, materials and other activities do not disappear because they are outside this worksheet.
A contractual allocation also does not prove that a facility physically received only renewable electrons or that a specific generator supplied it in every hour. A lower attributed rate is not evidence that demand vanished. Explain the method rather than allowing a familiar word such as renewable to carry a wider claim than the records support.
For E-24, the narrow statement is a provisional location-style estimate based partly on an assumed missing interval and a fictional factor. It must not be reused as a real reporting result. In a real inventory, replace the teaching inputs with reviewed records and suitable coefficients, then qualify the remaining uncertainty before making external claims.
Keep a calculation a reviewer can reproduce
Keep a calculation packet that another person can reproduce without relying on the preparer's memory. The reported tonnes are the final output of linked evidence, not a substitute for the evidence itself. Retain original versions and corrections, including documents excluded from the numerical total.
| Packet component | What the reviewer should be able to trace |
|---|---|
| Site and scope | Facility boundary, purchased stream, reporting dates and separately routed generation or exports |
| Activity | Meter serials, actual/estimated readings, multipliers, invoice dates and interval quantities |
| Reconciliation | Duplicate identities, replacement bills, meter changes, gap assumptions and revised totals |
| Factor | Method, location, data vintage, publisher/version, original unit and emissions coverage |
| Arithmetic | Unit conversion, intermediate mass, rounding and final labelled result |
| Additional method view | Instrument eligibility, volume coverage and applicable residual or fallback data |
| Review | Preparer, reviewer, approval state, unresolved decisions and retained calculation versions |
Table 4, shown one row at a time. Every field is retained.
Row 1
- Packet component
- Site and scope
- What the reviewer should be able to trace
- Facility boundary, purchased stream, reporting dates and separately routed generation or exports
Row 2
- Packet component
- Activity
- What the reviewer should be able to trace
- Meter serials, actual/estimated readings, multipliers, invoice dates and interval quantities
Row 3
- Packet component
- Reconciliation
- What the reviewer should be able to trace
- Duplicate identities, replacement bills, meter changes, gap assumptions and revised totals
Row 4
- Packet component
- Factor
- What the reviewer should be able to trace
- Method, location, data vintage, publisher/version, original unit and emissions coverage
Row 5
- Packet component
- Arithmetic
- What the reviewer should be able to trace
- Unit conversion, intermediate mass, rounding and final labelled result
Row 6
- Packet component
- Additional method view
- What the reviewer should be able to trace
- Instrument eligibility, volume coverage and applicable residual or fallback data
Row 7
- Packet component
- Review
- What the reviewer should be able to trace
- Preparer, reviewer, approval state, unresolved decisions and retained calculation versions
The reviewer should independently reproduce at least the material arithmetic and inspect the coverage decisions, rather than merely comparing the final number with the spreadsheet's displayed total. Give open issues a responsible owner. If the result changes, preserve which input, factor or decision caused the revision so a later report can explain it accurately.
Choose a system that preserves source distinctions
An evidence-organisation system is useful only if it preserves these distinctions. Ask whether it can retain source identifiers, meter intervals, factor versions, correction decisions and the difference between an estimate and an actual reading. A neat dashboard without those relationships can make review harder rather than easier.
Test a small handoff before depending on a tool: another reviewer should recognise COPY-01, follow the IMP-A to IMP-B change and find the unresolved November assumption. They should be able to reproduce the result from permitted records and understand what remains undecided.
These are requirements for the chosen process, not assertions about any named provider's live calculator, automatic integration or certificate-assurance capability. Confirm working functions and export rights with evidence. The record must remain understandable when the preparer, reviewer or software changes.
What to do before reporting the figure
Before reporting, confirm the boundary, reconcile the activity, review the factor and reproduce the units. Check that each interval appears once, estimates are visible and onsite or exported electricity has not been subtracted from an import quantity without a valid reason. Then establish which method views are required and whether their evidence is sufficient.
For E-24, the next action is not to remove the provisional label. It is to resolve the November replacement record, test the proxy and complete the outstanding method review. A small worksheet can be useful precisely because it exposes a decision that an annual total would hide.
Keep the practical distinction in view: this article's task is to assemble and calculate an electricity-attributed result. A fuller method comparison or procurement assessment answers additional questions. The final handoff should give the next reviewer enough information to continue without guessing, while preventing a limited calculation from becoming an unsupported wider claim.
Scope 2: purchased energy
