Generic unbranded metal bottle, material samples and measuring equipment at a product testing bench
Unit for Carbon Footprint: CO2e, Product Denominators and Comparisons
Diagram text description

Original synthetic editorial photograph. Generic unbranded metal bottle and raw material specimens at a product-testing bench; not a certified customer product or a quantitative result. Spatial grouping, branches, proportions or directional paths carry the claim with concise labels.

The unit for carbon footprint reporting is usually a mass of carbon-dioxide equivalent, written g CO2e, kg CO2e or metric tonnes CO2e. The “e” says that the result can include several greenhouse gases expressed on a common warming-impact scale. A complete footprint label also states what the mass refers to: one product, a defined amount of material, a service delivered over a stated lifetime, or an organization's emissions during a reporting period.

Thus “4 kg CO2e” is not yet an interpretable product result. Four kilograms per bottle, per kilogram of material and per thousand uses answer different questions. The production stages counted matter as well. A cradle-to-gate value stops at the factory gate; a cradle-to-grave value reaches use and end of life. Equal numbers with different system boundaries do not mean equal product impacts.

The quickest useful check is to read numerator, denominator and boundary together. Ask: which greenhouse gases are included, how much service or product is represented, and which life-cycle stages were counted? This article explains those three parts so a footprint is not reduced to a bare number.

Why the e in CO2e matters

Carbon dioxide is one greenhouse gas. Methane, nitrous oxide and other gases can also contribute to warming, but a kilogram of each does not have the same warming effect. Carbon-dioxide equivalent, or CO2e, expresses a defined amount of each gas as the mass of CO2 with the same estimated effect over the method's selected time horizon. The calculation uses a global warming potential, usually abbreviated GWP. ISO describes CO2e as a comparison unit, not a claim that every gas physically becomes CO2.

A label reading “kg CO2” may therefore refer only to carbon dioxide. A label reading “kg CO2e” can cover the gases included in its method. The difference is important for a material supply chain with methane or nitrous-oxide sources. Do not silently change CO2 into CO2e by adding an “e” to an old figure. The other gas amounts and applicable GWP values must be accounted for.

Gas masses require weighting before aggregation.
Gas masses require weighting before aggregation.
Diagram text description

GWP expressed as kg CO2e per kg of the particular gas. Use one stated method and horizon; never sum raw CH4 and N2O mass as though already CO2e. No numeric GWP supplied.

The chosen method and version matter when comparing studies. A changed GWP set can alter the CO2e result even if the physical activity is unchanged. A reader needs the gas basket and method reference before concluding that one product has improved or that two suppliers measured the same thing.

Grams, kilograms and metric tonnes

Mass prefixes change the scale of the reported figure, not the underlying subject. One metric tonne equals 1,000 kilograms; one kilogram equals 1,000 grams. A product component may have a footprint shown in grams CO2e, a finished good in kilograms CO2e and a company's annual inventory in metric tonnes CO2e. Converting the mass prefix is straightforward, but converting the denominator is a separate problem.

For example, 2.5 kg CO2e is 2,500 g CO2e and 0.0025 metric tonnes CO2e. That arithmetic does not tell you whether the number is per unit, per kilogram of output or for an entire batch. It also does not turn a kg CO2 result into kg CO2e. Only change the prefix after confirming that the gas coverage and boundary are otherwise identical.

A prefix changes scale, not the footprint boundary.
A prefix changes scale, not the footprint boundary.
Diagram text description

The downward arrows are two successive inverse conversions, not two alternative steps from the top. The same denominator and gas-equivalent basis remain attached.

When publishing a small number, make the symbol legible. “tCO2e” indicates metric tonnes of carbon-dioxide equivalent; “kgCO2e” indicates kilograms. An unqualified “tons of carbon” can mean elemental carbon mass, CO2 mass or CO2e depending on context. Those are not interchangeable. Write the full name once in a report and use a consistent abbreviation thereafter.

The footprint subject supplies the denominator

The carbon footprint's subject tells the reader what sits below the mass figure. For a company, an absolute inventory commonly covers the organization's selected operations over a year. For a product, the reported result must relate to a defined product system and unit of analysis. For a service, it may relate to a stated service outcome. These denominators make the same kg CO2e numerator useful for different decisions.

Table 1: scroll horizontally if needed.

Subject Possible result label What must be defined
Organization tCO2e per reporting year Entity boundary, scopes and period
Intermediate material kg CO2e per kg material Grade, production stage and declared amount
Finished product kg CO2e per functional unit Service delivered, quality, lifetime and stages

These examples are patterns, not prescribed values for every study. A company could also report emissions intensity per unit of sales, and a product could be reported per individual unit where that is meaningful. The key is to state the denominator before comparing or aggregating. Multiplying a per-product figure by annual production requires a verified count of comparable products and care over which impacts are already in the company inventory.

What is a functional unit?

A functional unit describes the quantified service a final product provides, including performance and duration. It is not merely a convenient package count. The GHG Protocol Product Standard says the function of the product and the reference flow needed to deliver it are central to a product inventory. When the function is known, the functional unit becomes the basis for gathering material, energy, use and disposal data.

Consider two reusable bottles. A result per bottle may be a poor comparison if one lasts for hundreds of uses while the other is intended for a much shorter life. A functional unit could specify delivery of a defined amount of beverage over a stated number of uses under stated cleaning assumptions. The exact unit would need to fit the study's goal and product type; it should not be chosen simply to favor one design.

Function includes the service level and its duration.
Function includes the service level and its duration.
Diagram text description

The branches describe the functional unit, not two quantities to add. Reference flow follows only after a product system can meet both conditions.

A good functional unit makes alternatives answer the same buyer question. If the purpose is to compare lighting, the unit might describe light output over a defined service life rather than one bulb. Without that match, even perfectly calculated kg CO2e results can lead to a wrong purchasing decision.

What are declared units and reference flows?

A declared unit is an explicitly stated quantity, such as one kilogram of an intermediate material, when the final service context is not known or is outside the study. A reference flow is the quantity of the studied product needed to deliver the chosen functional unit. The two concepts connect the reported denominator to the physical inputs used in the life-cycle model.

Table 2: scroll horizontally if needed.

Term Question it answers Example without a claimed result
Declared unit How much material is being described? One kilogram of a specified resin grade
Functional unit What defined service is being provided? A stated lighting output for a defined lifetime
Reference flow How much product delivers that service? The number of lamps and replacements needed

The definitions protect against casual cross-study comparisons. A kilogram of resin is not the same functional proposition as an assembled product that uses it. Nor can a buyer compare two lighting products on a “per lamp” basis if one has a much shorter service life without accounting for replacements. Record the unit in the study design and keep it consistent in later inventory updates.

A fictional lamp comparison makes the distinction concrete. Define the service as a specified light output and quality for 10,000 operating hours. Assume lamp A lasts exactly 10,000 hours and lamp B exactly 5,000 hours while delivering that same service. The reference flow is then one A or two B lamps. These lifetimes are teaching assumptions, not product ratings. A study must also include the relevant electricity, replacements and end-of-life treatment within its chosen boundary.

A declared result per lamp can feed that comparison, but it cannot finish it. If A has an illustrative manufacturing contribution of 6 kg CO2e per lamp and B has 4 kg, the manufacturing contributions for the service are 6 kg for A and 8 kg for two B lamps. B looks lower per item and higher for this particular manufacturing contribution per service. Neither result establishes the full lifecycle ranking because the use-stage electricity and other included stages have not been calculated.

Different product quantities can provide the same defined service.
Different product quantities can provide the same defined service.
Diagram text description

Fictional lifetimes with identical light performance and quality; no failures or repair. This is reference-flow arithmetic, not measured product performance. Use electricity and other stages remain outside the pictured quantity.

The example assumes replacement at the stated lifetime with no failures, repair or residual service. Different assumptions can change the reference flow. If the service period ends part-way through a second lamp's life, the applicable study method must determine how that situation is treated. Simply rounding or allocating a fraction to obtain a preferred ranking would hide a methodological choice. The functional unit specifies the task; the reference flow expresses what is needed to deliver it under the stated assumptions.

Per-item ranking can reverse after matching the service.
Per-item ranking can reverse after matching the service.
Diagram text description

Fictional manufacturing contributions:1×6=6 versus2×4=8 kg CO2e. These are not complete lifecycle totals; electricity and other included stages could alter the overall ranking.

Why one item may be a bad comparison

“One item” feels intuitive because procurement systems count units. But one item can be heavier, perform better, be repaired more often or last longer than another. A comparison that ignores those differences may reward a product that simply shifts impacts outside the chosen per-item snapshot. The correct denominator depends on what decision the reader is making.

Table 3: scroll horizontally if needed.

Per-item shortcut Missing comparison issue
One battery against one battery Capacity and delivered energy may differ
One garment against one garment Fabric mass, washes and useful life may differ
One package against one package Product quantity or protective performance may differ

Per-item reporting can still be valuable for inventorying a specific SKU. It becomes risky when presented as a like-for-like environmental superiority claim without checking service equivalence and boundaries. If a supplier sends kg CO2e per item, ask for item specification, mass, expected use, method and included life-cycle stages. If those cannot be aligned, treat the numbers as separate descriptions rather than a ranked comparison. A comparison can be resumed later if the teams agree a common service definition and collect the missing use data.

Boundaries: gate and grave are not equal

A carbon footprint has a system boundary as well as a mass unit. Cradle-to-gate normally covers raw materials and production through the factory gate; cradle-to-grave extends through distribution, use and end of life under the study's defined scenarios. A partial footprint can be useful, but it should be called partial and its cutoff should be visible. ISO 14067 covers both full and partial product footprint quantification in a climate-change-only impact category.

Imagine two products each labelled 5 kg CO2e. One figure may cover only materials and manufacturing; the other may include many years of electricity use. The identical numerical unit does not make the two results equivalent. A later product design change can also move emissions between stages, for example by increasing manufacturing impact to reduce use-phase energy. Only a consistent boundary reveals the trade-off.

A gate result stops before downstream life-cycle stages.
A gate result stops before downstream life-cycle stages.
Diagram text description

The last arrow extends the study boundary; it is not an instruction to add a second complete footprint. Comparable results need aligned inclusions and methods.

When preparing a label, put the boundary near the number, not buried in a separate file. A clear expression is “kg CO2e per defined product unit, cradle-to-gate, method and data year stated.” The exact method may require more detail, but those words immediately prevent the most obvious mismatch.

Let the units cancel in a calculation

Unit checking is a practical quality control in footprint calculations. If electricity use is measured in kilowatt-hours and the selected factor is in kg CO2e per kilowatt-hour, multiplying activity by factor gives kg CO2e. The kWh cancels. If the data arrive in megawatt-hours, convert them first or use a matching factor. Otherwise a thousand-fold scaling error can slip into an apparently polished result.

The same reasoning applies to materials and transport. Kilograms of a material multiplied by kg CO2e per kg material produce kg CO2e; tonne-kilometres require an emission factor per tonne-kilometre. Record the activity unit, factor unit, output unit and conversion. A factor that is geographically or temporally inappropriate remains a poor choice even if its dimensions cancel correctly.

Dimensional analysis also reveals whether the result is an absolute amount or an intensity. If the calculation is for a batch, divide by the defined product or functional quantity only after checking the batch allocation. The denominator is a method choice, not an optional formatting detail added at the end.

A small worked example

For example, suppose a deliberately hypothetical production step uses 10 kWh of electricity and the illustrative emission factor is 0.4 kg CO2e per kWh. Multiplying 10 kWh by 0.4 kg CO2e/kWh gives 4 kg CO2e. The kWh units cancel. This is a unit demonstration, not a current electricity factor for any specific country, supplier or grid.

If those 4 kg CO2e were allocated to a batch of two otherwise equivalent products under a justified method, the step would contribute 2 kg CO2e per product. That is not necessarily the entire product footprint: materials, transport, other manufacturing energy, use and end of life may still need assessment depending on the chosen boundary. If the products are not equivalent, simply dividing by two could be wrong.

Activity and factor units cancel to the result unit.
Activity and factor units cancel to the result unit.
Diagram text description

Illustrative factor, not a regional recommendation. The four kilograms cover one included electricity activity, not a complete product footprint.

The example shows why “what unit?” cannot be answered by the numerator alone. Activity records, factor units, allocation and final denominator all have to line up. Keep the source value and conversion in the calculation trail so another reviewer can reproduce it.

Absolute tonnes versus per-unit intensity

An absolute footprint and an intensity answer different management questions. A company may report 2,000 tCO2e for a defined year and also 0.5 kg CO2e per unit of output. The former speaks to total climate impact inside the boundary; the latter can help compare production efficiency. A falling per-unit intensity does not prove total emissions fell if output rose substantially.

Table 4: scroll horizontally if needed.

Measure Typical question Common misreading
Absolute tCO2e How much was emitted in the period? Confusing a changed boundary with an operational reduction
kg CO2e per item What is the footprint of this defined item? Comparing unlike items or stage coverage
kg CO2e per functional unit What impact supplies this defined service? Changing use or lifetime assumptions between studies

Report both where each adds value, with the denominators in plain sight. A product footprint should not be substituted for the company inventory, and an annual corporate intensity should not be sold as a full product life-cycle result. The two datasets can inform each other, but only through a carefully documented mapping.

Check the gases and GWP edition

CO2e calculations depend on a gas basket and a set of global-warming-potential values. If a study includes methane and nitrous oxide while another counts only CO2, identical activity data can produce different labelled totals. A new method edition can also update factors or GWP values. Differences in the resulting kg CO2e may then reflect method changes rather than a physical improvement.

For comparison, record which greenhouse gases were covered, the standard or method version, the GWP time horizon and any factor database version. These details belong in supporting documentation even if the consumer-facing label must be short. If a supplier sends an unexplained figure, seek the method record before treating it as equivalent to a second supplier's result.

Do not overstate precision. Product activity data and supplier estimates may be uncertain. Expressing a result to four decimal places does not establish that its underlying inputs are that accurate. A rounded number with clear scope and method can be more useful than a precise-looking number without them.

The conversion itself can be expressed without borrowing an unsupported numerical GWP. For each gas, multiply its mass in kilograms by the selected characterisation value in kg CO2e per kg of that gas. The resulting contributions share a CO2e basis and can then be summed within the inventory boundary. Adding raw kilograms of methane and nitrous oxide first would discard the different weighting that gives CO2e its meaning.

A factor already expressed in kg CO2e per kWh normally contains the relevant greenhouse-gas aggregation for its stated coverage. Applying a second gas-specific GWP to that aggregated factor would not make it more complete; it would change the calculation incorrectly. If a dataset instead provides separate gas masses, the analyst needs the appropriate conversion for each before aggregation. The source's unit and method determine which operation is needed.

Read the dataset unit before choosing the transformation.
Read the dataset unit before choosing the transformation.
Diagram text description

The left needs gas-specific weighting; the right is already characterised for its stated coverage. Applying a second GWP to an aggregated CO2e factor is incorrect.

This is why a changed GWP edition should be identified as a method change when interpreting a time series. Recalculating an unchanged gas inventory with different weights can change the reported CO2e total without any change in the physical emissions. Keep that explanation separate from changes in activity, product design or energy use, so readers can understand what actually moved.

Do not subtract offsets from the ISO 14067 figure

ISO 14067 is a product carbon-footprint quantification standard. Its published scope addresses the climate-change impact of the studied product system, including principles for a full or partial footprint; carbon offsetting and communication of CFP information are outside its quantification scope. That distinction matters when someone tries to turn a kg CO2e product result into a marketing claim by subtracting external credits.

A purchased credit may have a role in a separate claim framework, but it should not silently alter the measured life-cycle footprint that the product study calculated. Keep the product-system emissions and removals with their own methodology and report any external action separately under the rules governing that action. A “net zero product” label requires more than a corrected arithmetic line.

The article about net corporate emissions deals with broader balance claims; this unit guide's boundary is narrower. The product result's unit still needs to name a functional or declared denominator, and the analysis remains grounded in the actual life-cycle stages studied.

How to label a defensible footprint

A defensible footprint label can be assembled in four stages:

  1. Name the subject: organization, facility, product, material or service, with an exact specification.
  2. State the numerator: amount in g, kg or metric tonnes CO2e, plus the gas basket or method where needed.
  3. State the denominator: reporting year, declared material quantity or functional unit with its performance and duration.
  4. State the boundary and method: life-cycle stages, geography, data period and the standard or calculation version.

A readable amount still needs its denominator and boundary.
A readable amount still needs its denominator and boundary.
Diagram text description

Fictional electricity-only label used later in the worked example. The figure annotates two fields; the section retains method version, period and the full study record.

For example, “kg CO2e per one kilogram of specified material, cradle-to-gate, using the documented 2026 supplier-data set” conveys more than “low-carbon material: 3”. It is still not a verification certificate, and the underlying report should explain allocations and uncertainty. The reader can at least see what physical quantity was studied and which stages were included.

The label should be tested by someone who did not build the calculation. Ask that reviewer what they think the result includes. If they assume a full life cycle when the study ends at the gate, or interpret CO2e as CO2-only, revise the wording. A short label is successful only when it does not create a wrong inference. Preserve the detailed supporting study so the visible label can remain short without losing its evidence trail.

The shorthand that misleads

“Kg CO2 per product” is a tempting shortcut. It can omit whether other greenhouse gases were converted to CO2e, whether “product” means one item or a specified function, how long the item is assumed to work and which life-cycle stages are included. None of those questions is pedantry: each can change the meaning of the number. The shorthand may be understandable within one team but hazardous on a customer-facing page.

It also invites false comparisons. A lighter-looking figure can be achieved by stopping at the factory gate, choosing a shorter product lifetime or excluding difficult supplier stages, not by reducing real emissions. If two studies use different methods, do not force an apparently precise ranking. Present each as its own result until the bases can be reconciled.

Repair the label by adding the missing denominator and boundary. “2 kg CO2e per specified item, cradle-to-gate” is materially clearer than “2 kg carbon”. If the product function and lifetime matter to the decision, add them too and provide a link to the study record when the final content is published.

Read a result label through two numerical checks

A useful label lets the reader reconstruct the scale of a result. Consider a fictional electricity record of 0.01 MWh, used to make two equivalent items. The illustrative factor is 0.40 kg CO2e per kWh. Convert the activity first: 0.01 MWh × 1,000 kWh/MWh = 10 kWh. Multiplying by the factor gives 4 kg CO2e for the electricity activity. Dividing by two items gives 2 kg CO2e per item for this included electricity contribution.

There are three different quantities here: purchased electricity, its estimated emissions, and the allocated contribution per item. Writing only “2 kg” loses both the gas-equivalent basis and the denominator. Writing “2 kg CO2e per item” still needs the boundary: in this example it covers the specified electricity contribution only. It does not include materials, freight, use or disposal, and it is not a complete product footprint.

Conversion and allocation are separate operations. Electricity contribution only; equal allocation assumed.
Conversion and allocation are separate operations. Electricity contribution only; equal allocation assumed.
Diagram text description

Convert 0.01 MWh to 10 kWh, then apply the fictional 0.40 kg CO2e/kWh factor to obtain 4 kg CO2e. Dividing that electricity contribution equally between two assumed equivalent items gives 2 kg CO2e/item. Equal allocation is an explicit teaching assumption, not a method validated by dimensional arithmetic. This is not a complete product footprint.

Table 5: scroll horizontally if needed.

Step Operation Result and meaning
Convert activity 0.01 MWh × 1,000 10 kWh electricity
Apply matching factor 10 kWh × 0.40 kg CO2e/kWh 4 kg CO2e included contribution
Allocate to equivalent output 4 kg CO2e ÷ 2 items 2 kg CO2e per item, electricity only

The unit conversion does not justify the allocation. The two items are assumed equivalent and equally responsible for this electricity in the teaching example. If one takes twice as much machine time, or shares a meter with unrelated output, dividing by two may no longer represent the process. A correct cancellation of units can reveal a thousandfold error, but it cannot establish that the chosen allocation reflects the product system.

The second check concerns intensity. Imagine one period produces 1,000 equivalent units at 2 kg CO2e per unit. The corresponding included total is 2,000 kg CO2e. In a later period, 1,500 equivalent units are produced at 1.6 kg CO2e per unit, giving 2,400 kg CO2e. All figures are fictional and use the same assumed boundary and method. The intensity falls by 20%, while the absolute total rises by 20%.

Both statements can be true. The activity increased by 50%, which more than offsets the lower amount per unit. Calling this simply a “20% carbon reduction” would leave the reader unable to tell which quantity fell. A clearer statement gives both the intensity and total, names the output denominator and explains the change in volume. If the product mix or boundary changed, that explanation would need another step before the periods could be compared.

Lower intensity can coexist with a larger total.
Lower intensity can coexist with a larger total.
Diagram text description

Same assumed boundary and equivalent units: intensity falls20%, output rises50%, total rises20%. The lanes are periods, not contributions to add. No real company improvement claimed.

Notice what prefix conversion cannot fix. Expressing the totals as 2.0 and 2.4 tonnes instead of 2,000 and 2,400 kilograms changes neither the direction nor the percentage. The denominator and boundary carry the meaning; the prefix only changes the display scale. Similarly, switching from an item denominator to revenue introduces prices and currency into the intensity measure. It is a different indicator, not a unit conversion of the original physical result.

When assessing a supplier statement, try these two checks on one disclosed line. First follow activity units into the factor and result. Then follow the result into its denominator and included stages. A missing step identifies a precise question for the supplier, such as the MWh conversion, the output allocation or the lifecycle cutoff. That question is more useful than asking for a more impressive-looking total.

What the unit does not tell you

The unit for carbon footprint is commonly mass of CO2e, but the full answer is amount per precisely defined subject. For a product, read the functional or declared unit and life-cycle boundary beside kg CO2e; for an organization, read the entity boundary, scopes and reporting period beside tCO2e. Convert grams, kilograms and tonnes only after the gas coverage and denominator match.

Before using a figure to choose a supplier or publish a product comparison, ask three practical questions: What exact service or quantity is the result for? Which stages and gases are included? Can the calculation trail reproduce the amount and factor units? If any answer is missing, the number may still be useful internally, but it is not yet a sound like-for-like comparison.