Product Emissions Calculator: Inputs, Checks and Honest Outputs
A person holds a ruler against a green unbranded casing at a worktable. Disassembled motor and electronic components, jars of material samples, blocks, a small meter, paper parts records and an abstract chart screen surround it. No legible brand, emissions amount, approved factor, customer identity, verification or certification appears. The scene illustrates matching a product to its component evidence, not a real study or functioning named software.
Diagram text description
A product emissions calculator turns a defined product model and source data into a greenhouse-gas result. It does not decide what the product's function is, which life-cycle stages count, whether a supplier's factor applies or whether the final figure has been verified. Those choices must be made and recorded by the study team before a credible number in kilograms of carbon dioxide equivalent (kg CO2e) can be used outside the calculation screen.
To use a calculator well, state the exact product and unit, set the boundary, prepare component and process activity data, inspect emission-factor sources, check allocation and missing data, then reproduce at least one result manually. Export the assumptions and evidence with the number. If the tool cannot show how it handled an excluded stage, conflicting supplier figures or a factor update, treat its output as a screening estimate until those questions are resolved.
This article helps you choose, run and audit a tool. A separate method guide explains product-footprint calculation in full. No generic online calculator can turn an unknown product into a certified result with one click.
The starting question is therefore not which tool produces the lowest figure. It is which tool lets your team explain a result and its limits. The GHG Protocol calculation-tools FAQ describes tools as aids to calculation rather than a mandatory route. Its guidance on activity data and suitable factors supports that distinction; it does not supply a complete product-study design.
Choose the tool by the decision it must support
Begin with the job the result must do. A quick screening model can show which inputs deserve better data. An internal improvement study may need repeatable stage results across product versions. A buyer disclosure needs an agreed unit, boundary and data period. A public comparative claim may require product rules and assurance beyond the calculator interface. The required evidence rises with the consequence of using the number.
| Intended use | Minimum decision question | Appropriate output status |
|---|---|---|
| Screening | Which sources may dominate? | Provisional estimate with visible gaps |
| Internal redesign | Did a change alter like-for-like stage results? | Versioned model with consistent method |
| Customer disclosure | What does the buyer's specification request? | Reviewed result with required unit and boundary |
| Public comparison | Are product rules and studies compatible? | Specialist review before a comparative claim |
Table 1, shown one row at a time. Every field is retained.
Row 1
- Intended use
- Screening
- Minimum decision question
- Which sources may dominate?
- Appropriate output status
- Provisional estimate with visible gaps
Row 2
- Intended use
- Internal redesign
- Minimum decision question
- Did a change alter like-for-like stage results?
- Appropriate output status
- Versioned model with consistent method
Row 3
- Intended use
- Customer disclosure
- Minimum decision question
- What does the buyer's specification request?
- Appropriate output status
- Reviewed result with required unit and boundary
Row 4
- Intended use
- Public comparison
- Minimum decision question
- Are product rules and studies compatible?
- Appropriate output status
- Specialist review before a comparative claim
A decision asks whether the calculator result is intended for external use. No leads to internal screening or redesign with visible assumptions; Yes leads to disclosure or comparison with agreed requirements. Below, the common audit controls run from goal, product unit and scope to source inventory, factor choices, validation and evidence export. These controls must be inspectable in order. Internal use is not permission to ignore data quality, and external use does not automatically create certification, verification or a right to make a comparative claim.
Diagram text description
This is a selection guide, not a universal legal threshold. Ask the recipient what method, stages and review they expect. A tool suited to internal scoping may be useful without producing a file suitable for customer publication.
Request a demonstration using one of your own representative product records, with commercially sensitive fields protected. Have the provider identify the intended decision, product unit and boundary, then show the source inventory, factor choices, validation and exported evidence in order. Record any missing control rather than treating a successful demonstration of a different task as acceptance. Screening, redesign, disclosure and comparison are separate intended uses, not automatic stages of certification.
Define the exact product and unit before entering data
Enter the product that actually exists. A SKU, material grade, revision and production location can change component masses and supplier data. If a calculator only asks for a generic product category, inspect what the category assumes before treating the result as product-specific. A variant with different electronics, packaging or finish may need a separate record rather than being averaged into one label.
The unit tells the model what the result is per. For a finished product, a functional unit can describe a quantified service with performance and duration; the reference flow is how much product delivers it. For an intermediate input whose use is unknown, a specified declared unit may be more practical. Both choices are more informative than an unlabeled “per item” number when alternatives have different capacity or useful life.
Write the unit in plain language beside the tool's field name. A supplier value per kilogram of material cannot be inserted as though it were per finished item. Record conversions, product quantity and the intended audience so a reviewer can reconstruct the denominator later.
A useful test is whether another analyst could tell which physical product the record describes without seeing its catalogue photograph. Give that analyst the revision, specification, quantity and service definition. If two variants share a name but have different material masses or lifetimes, resolve those differences before consolidating them. Otherwise a later comparison may attribute a changed denominator to an improvement in manufacturing.
Set the lifecycle stages instead of accepting defaults
Many tools let users select pre-production, manufacturing, distribution, use and end-of-life modules. A selected module does not mean its data are complete; an unselected module does not mean its emissions are zero. Decide whether the result is cradle-to-gate, cradle-to-grave or a clearly named partial footprint before entering quantities. ISO 14067 covers both full and partial product carbon footprints within the climate-change impact category.
Check every default stage checkbox. A browser tool may start with manufacturing alone; another may include a generic use phase. If your customer asks for a cradle-to-gate material result, use-stage assumptions may not belong. If the output is presented as a full life cycle for an electric product, use-stage energy might be essential. The boundary should be intentional, not inherited from software settings.
Save a scope note listing included and excluded processes, geography, data period and scenario assumptions. The result label must disclose any partial scope. An omission recorded in the scope is different from a missing input that should have been collected under the chosen scope.
Use a simple boundary register before configuring the interface: process, inclusion decision, reason, evidence state and responsible person. For a required process with no data, keep inclusion and incompleteness together. Do not move it outside the boundary merely to remove an error message. The public ISO abstract confirms that partial footprints are within its scope; it does not make an arbitrary two-line subtotal a conforming partial footprint.
Start with a component-level bill of materials
A product bill of materials is the input backbone. List each component, material type and mass per reference flow, including relevant packaging. A generic “one bottle” category may conceal a metal body, cap, elastomer seal and decorative coating with different suppliers and processes. Component-level data let the analyst identify which upstream factor or supplier product footprint belongs where.
The enclosing product record specifies revision and reference flow. Within it, component grade and mass connect to a matched supplier source with its own revision and review state. Function and quantity define what the result is per; a mass-based supplier value is not automatically a finished-item value. Retain relevant units and upstream boundary so a supplier component footprint and a separate generic material factor do not count the same production twice. No physical quantities, actual supplier performance or named software capability are invented.
Diagram text description
Check the match between supplier file and calculator field. A purchased component footprint might already include raw material and component manufacturing, while a separate generic material factor could count those same stages again. The unit also matters: a component PCF per piece cannot be multiplied by kilograms without converting through the actual quantity. Version and source dates should travel with the component.
Not every tiny part needs the same collection effort. Prioritize major contributors and uncertain inputs, but maintain a complete component register so an excluded or estimated item is visible. If the BOM changes in a later revision, preserve the older list with the result it produced.
Keep material identity separate from its evidence quality. A precisely weighed component may still use a weakly representative factor; a supplier footprint may have a clear source but the wrong product revision. Check both. Where a supplier result already covers extraction and component manufacture, mark those covered processes in the model so a second generic entry does not count the same contribution again.
Add manufacturing energy, scrap and shared processes
Factory emissions must be assigned to the defined product rather than copied wholesale from a site total. A calculation interface may request electricity, fuel, process gases, waste and production quantity. Its default allocation can be as consequential as its factor selection. Record what activity period and batch the data represent, then confirm that the reference flow receives the correct share.
- Record metered or invoiced activity with its unit and time period.
- Map each activity to the relevant operation and product or production batch.
- Separate directly measured product activity where possible; otherwise document the shared-process allocation rule.
- Check scrap, co-products and yield so input amounts reconcile with saleable output.
- Review whether a supplier or component factor already includes any factory process you are adding.
A sophisticated interface may automate this sequence, but its data mappings still need human review. If factory production mix changes, rerun the allocation with current quantities rather than applying a stale share. A reproducible model saves both the numerator activity and denominator output.
For a factory with several output lines, also confirm whether meter coverage is complete for the selected period. A disconnected submeter or excluded overnight load can distort the allocation before any emission factor is applied. Record who checked the source data against the production report.
For example, distinguish purchased material entering a batch from the mass remaining in finished items. A yield loss does not disappear because only the saleable unit is exported. Retain the relevant input, output and scrap records, and document which processes the selected factor already includes. This is a reconciliation task before it becomes a question about which allocation option to select in a menu.
Enter distribution, use and disposal only with scenarios
Downstream stages often rely on scenarios rather than direct measurements. Distribution needs route, mode, mass and distance. Product use may need expected life, energy or cleaning frequency. End of life needs a plausible treatment route and geography. Enter these only for the boundary selected, and label which values are measured, modeled or assumed.
The in-scope frame contains both data-available activities and a required-but-missing activity. The outside-scope frame contains an intentional exclusion with a reason. This separates boundary choice from completeness, so an empty required stage cannot become a zero or an exclusion just because the calculator cannot fill it. Delivery, use and disposal scenarios carry distinct assumptions and sources: route, mode, distance, service duration, operation and end-of-life treatment for the relevant geography. Keep their source dates and periods. No particular lifecycle boundary or actual downstream behaviour is asserted by the generic labels.
Diagram text description
Do not accept a generic “average product life” simply because the tool provides it. An appliance with ten years of use has a different service denominator from one replaced after two; a reusable item may require washing or refills. A scenario can be useful for sensitivity analysis, but it should not be presented as observed customer behavior. A default end-of-life split also may not represent the markets where the product is sold.
Ask whether changing a plausible downstream assumption would change the conclusion. If yes, report a range or limitation rather than a single seemingly exact result. The calculation screen should make the scenario recoverable in an exported record.
Maintain one scenario record per relevant downstream stage rather than one undifferentiated assumptions note. For delivery, preserve route and mode evidence; for use, preserve service duration and operating assumptions; for disposal, preserve treatment and market assumptions. Put the reference period or source date beside each. If none is resolved, an included-lines subtotal can still support debugging, but it cannot answer a full-life-cycle question.
Ask where every factor came from
An emission factor connects activity to an estimated greenhouse-gas amount. Its unit must match the input; its geography, technology, year, system boundary and gas-method basis should fit the product process. A tool may offer a library, import your own factor or automatically choose one. Automatic selection is a convenience, not a substitute for checking representativeness.
For example, one kilowatt-hour of factory electricity needs an applicable electricity factor for the relevant market and period. A factor for a different country's grid might be a poor proxy. A purchased material needs a factor for the actual material grade and production route. An “avoided emissions” factor is not the same as a factor for consumed activity.
Store the factor title, provider, version, unit, geography, reference year, boundary and reason for selection. If a better supplier-specific factor arrives later, rerun the result and retain the previous version. Do not quietly overwrite an old claim with a new factor and call the difference a product design improvement.
Inspect the factor denominator as carefully as its numerical value. A material production dataset per kilogram, a component result per piece and an electricity factor per kilowatt-hour are not interchangeable options in one drop-down list. The GHG Protocol FAQ warns that defaults can fail to reflect regional or process differences. Ask the method owner to explain the chosen match and any remaining limitation, rather than assuming a recent library update makes every selection appropriate.
Check arithmetic and units yourself
Check at least one line outside the tool. The basic pattern is activity multiplied by a compatible factor to produce a stage result. If 12 kWh of electricity is multiplied by an illustrative 0.5 kg CO2e/kWh factor, the result is 6 kg CO2e for that electricity line. Those numbers are hypothetical and are not a valid factor or measurement for any real product. The kWh units cancel; a calculator that outputs tonnes without a visible conversion needs scrutiny.
The equation uses an invented training factor: 12 kWh × 0.5 kg CO2e/kWh = 6 kg CO2e for the electricity batch line. It is not a real grid coefficient or product measurement. The kWh denominator cancels the activity unit. Selection still requires a suitable place, process, year, system boundary and greenhouse-gas method. A different test pairs material mass in kg with a factor per kWh; this is incompatible and the row must be rejected, not assigned zero or repaired by an unrelated gram conversion. Batch allocation to a product is shown separately in the worked example.
Diagram text description
Material and transport lines should receive the same dimensional check. A factor per kilogram must be multiplied by kilograms of the matching material per reference flow. A tonne-kilometre transport factor needs a mass-and-distance activity in the corresponding units. Unit errors can survive a polished interface because the final number may still look plausible.
Review gas coverage and the selected global-warming-potential method. CO2e is not merely a label added to a CO2-only result. A later method update can change a result even if activities are unchanged, so note the edition used.
A documented conversion changes scale within the same dimension: 200 g becomes 0.200 kg. It does not turn kilograms into kilowatt-hours. If the field contains a mass but the selected factor expects electricity, reject that row and obtain the appropriate mapping. Recording the expected unit at import is useful, but the review should also check the stored quantity after conversion and the unit attached to the exported result.
Review how the calculator allocates shared operations
Shared operations are a hidden source of error. A factory may run several products through one line and receive one electricity bill. A tool that divides that bill evenly by units might misstate products that use different processing time or mass. Ask whether the software lets you meter separately, assign activity to work orders or choose and record a justified allocation rule.
Do not assume that more detailed allocation is always better. The rule should be proportionate to the shared process's influence and consistent with the chosen product method. Document co-products, waste and any system expansion or allocation basis so a reviewer can understand why this product received its share. If an alternative reasonable basis changes the comparison, report that sensitivity.
When evaluating a calculator, inspect one shared-process example in its documentation. Does it reveal the original site amount, production denominator and resulting allocation? If not, the output may be hard to reproduce even if the front-end arithmetic is correct.
In the fictional test below, equal allocation is restricted to three identical output units from the stated batch. It is not advice to divide every factory total by item count. Products with different processing requirements need an appropriate method and supporting denominator. The accessible sources here do not establish a universal allocation hierarchy; consult the applicable product method and responsible specialist where co-products or materially different outputs require a more specific treatment.
Do not let blank fields masquerade as zero emissions
A missing field is not a measured zero. It can mean the stage is outside the declared boundary, the activity was confirmed absent, the data have not been supplied or a proxy has been used. Those states have different meanings and should lead to different actions. A calculator that silently converts blanks to zero can produce an attractive undercount.
| Input state | What it means | Appropriate record |
|---|---|---|
| Outside scope | Process intentionally excluded | Disclosed boundary and reason |
| Confirmed zero | No relevant activity after checking | Source or review evidence |
| Missing | Relevant quantity not yet available | Open data-gap owner and priority |
| Proxy | Representative substitute selected | Factor source, fit and uncertainty |
Table 2, shown one row at a time. Every field is retained.
Row 1
- Input state
- Outside scope
- What it means
- Process intentionally excluded
- Appropriate record
- Disclosed boundary and reason
Row 2
- Input state
- Confirmed zero
- What it means
- No relevant activity after checking
- Appropriate record
- Source or review evidence
Row 3
- Input state
- Missing
- What it means
- Relevant quantity not yet available
- Appropriate record
- Open data-gap owner and priority
Row 4
- Input state
- Proxy
- What it means
- Representative substitute selected
- Appropriate record
- Factor source, fit and uncertainty
The centre empty-field record branches to four distinct states. Missing means required evidence needs an owner and resolution and may block a complete result. Confirmed zero requires evidence that relevant activity was absent. Proxy means a documented substitute with source, fit and uncertainty rather than a measurement. Outside scope means an intentional boundary decision with a reason, not physical zero emissions. The responsible next action depends on the state: obtain data, retain absence evidence, review a proxy, or document scope. The same rule applies to an import mapping failure, but its technical owner differs from a supplier-data owner.
Diagram text description
Before accepting the total, export a data-gap report. Prioritize missing inputs likely to dominate or reverse the decision. A simple estimate with visible unknowns is more useful than a false complete number.
A reviewer should be able to distinguish an absent field caused by software import failure from a genuinely missing supplier response. Both need attention, but they have different owners and repair paths.
Do not solve every missing value with the same fallback. A missing supplier mass needs product evidence; an import mapping error needs a corrected interface or transformation; an uncertain scenario needs an explicit assumption and sensitivity assessment. Assign the right person to each. IEMA's practical guidance allows work to begin with imperfect data while calling for transparent gaps and improvement. That is not permission to present an incomplete inventory as complete.
Save factor versions and study assumptions
Calculators can change without the product changing. Factor libraries update, supplier files are corrected, a background dataset is replaced or a tool modifies its allocation logic. A result should therefore carry the calculation date, model version, factor versions, activity period and user decisions. If a product's reported footprint falls, the study team should know whether actual material or energy use fell or whether the input basis changed.
Keep the previous result and its source package rather than overwriting it. A change log can identify new data, corrected activity, revised supplier boundary, updated factor, changed production mix and altered downstream scenario separately. This is particularly important if the value was already shared with customers and must be explained.
Tool selection should include versioning and export capability. A downloadable PDF with only a final number is less useful than a machine-readable record of inputs, factors, allocations and unresolved gaps. An audit trail is not a luxury feature when product models are expected to be revised.
Test reconstruction, not just the presence of a version number. Select an older result and ask a colleague to recover its original BOM, activity period, coefficient and scenario. Then change one cause in a separate version and explain the difference. If the tool displays only the latest supplier dataset in every historical export, a change log alone will not recover what was actually calculated earlier.
Reconcile the stage subtotal before using the total
A total should reconcile to the stages and components included in the stated boundary. Inspect the line items and stage subtotals before accepting the headline. A supplier component PCF may include upstream transport, while a separate distribution entry may include the same leg. A BOM quantity may be recorded per production batch rather than per product. Two wrong entries can offset each other and leave a plausible-looking total.
Compare the calculator's stage list with a simple process map. Mark a source for every in-boundary stage. Check whether each component mass, energy amount and product yield corresponds to the chosen reference flow. Then sample one large factor and one allocation. If the calculator excludes a stage due to missing data, the total should say partial or unresolved rather than implying completion.
Reconciliation should finish with a reader-facing result label: numeric value, CO2e unit, product unit, boundary, method/data period and review state. Without those fields, a subtotal cannot safely travel into a product page or procurement file.
For an included-lines subtotal, list the counted lines and the unresolved processes beside it. A material line plus an allocated electricity line may reconcile perfectly while packaging, another factory operation or transport is absent. Arithmetic completeness and study completeness are different tests. In the example below, adding 1 and 2 correctly gives 3; it does not demonstrate that those are all the emissions of the product.
Test one line with a transparent hypothetical example
This fictional training example tests a calculator, not an actual product. The reference flow is one identical unit of demonstration product P-1. Component A is 200 g per unit. Three identical finished units share 12 kWh of batch electricity. Both factors below are invented teaching values, not measured supplier data or real electricity coefficients.
For the material line, convert 200 g to 0.200 kg using 1,000 g per kg. Multiply 0.200 kg by the fictional training factor of 5 kg CO2e/kg: the material contribution is 1 kg CO2e per unit. Preserve the original 200 g, the conversion and the factor denominator so the analyst can distinguish a scale conversion from a changed physical quantity.
For electricity, multiply the batch's 12 kWh by the fictional training factor of 0.5 kg CO2e/kWh. This gives 6 kg CO2e for the batch. Divide by three identical units under the documented equal-unit teaching assumption: 6 ÷ 3 = 2 kg CO2e per unit. The included-line subtotal is therefore 1 + 2 = 3 kg CO2e per unit. Retain the batch numerator, three-unit denominator and each separate contribution.
The batch frame divides fictional electricity emissions of 6 kg CO2e equally among three identical units, with 2 kg CO2e assigned to each. The batch used 12 kWh and an invented .5 kg CO2e/kWh coefficient. One unit also contains .200 kg of material with an invented 5 kg CO2e/kg coefficient, contributing 1 kg CO2e per unit. Its subtotal is therefore 1+2=3 kg CO2e per unit. This is a teaching allocation for identical units, not a universal rule. Retain the original activity, output denominator, sources and conversions, and reconcile any duplicated upstream coverage. Shipping, use and disposal remain unresolved for a full-life-cycle study; neither this drawing nor correct addition proves completeness, real product performance or verification.
Diagram text description
Now test two failures. If Component A's mass is missing, its material contribution is unresolved, not zero; do not silently carry forward the valid case's 1 kg CO2e per unit. If mass in kg is paired with a factor per kWh, the units are incompatible. Reject that row and obtain the correct mass-based factor or the appropriate activity input. A g-to-kg conversion cannot repair a mass-to-electricity mismatch.
Shipping, use and disposal remain unresolved for a proposed full-life-cycle study. The analyst must obtain the required stage evidence or make a justified, documented boundary decision. Until that happens, no complete life-cycle result is available: label 3 kg CO2e/unit only as the fictional included-lines subtotal. An unresolved required material input likewise blocks a complete result, even if the electricity line can still be reproduced.
Use these expected values as a small acceptance test. Check the stored units, row results, allocation and subtotal against them, then inspect the export. An output of 6 kg CO2e per unit loses the batch denominator; an output of 1,000 kg CO2e per unit from the material line misses the gram conversion. Passing this teaching vector proves neither real-factor suitability nor completeness of a real product study.
Export more than a single CO2e number
Export the calculation as a study record, not merely a number. Include product ID and revision, functional or declared unit, reference flow, included and excluded stages, component and activity quantities, factors with source and version, allocation, data gaps, scenarios, date and reviewer state. Stage subtotals make the result inspectable. A user should be able to open the export after a year and reproduce the same figure without guessing which default was applied.
Separate frames retain Version A and Version B. Each result carries the product unit, scope and review status and links to its own original quantities, supplier/activity evidence, factors, conversions, allocation denominator, scenarios, gap records and review. Record the reason for the newer version so a product revision, factor update or scenario change is not mistaken for physical improvement. The older package is not overwritten. Sharing can use audience-appropriate summaries with protected source access; these generic frames do not guarantee a tool’s export functionality, imply assurance or create a current product result.
Diagram text description
Separate audience layers. A buyer may see a concise value and boundary label; an analyst should have the detailed trail; a reviewer needs access to original supplier and meter records. A concise export is useful only if it points to its full evidence package. “3 kg CO2e” on its own does not say what was counted or what product service it represents.
Check how the tool handles confidential supplier data. Share only the appropriate summary with customers while retaining protected source records internally. A secure workflow can still be transparent about method, scope and assurance without exposing every commercial document.
A practical export test is to give a colleague the package without access to your current session. Ask them to identify the product revision, reproduce a selected subtotal and locate an unresolved assumption. Record what they could not recover. Keep the link between each result version and its source package, so a later factor or scenario change creates a new reconstructable result rather than silently replacing the evidence behind the old one.
Know when a calculator output is still preliminary
The word “calculated” does not mean “verified.” A browser tool may offer a preliminary estimate from user inputs, generic factors or selected lifecycle stages. That can guide data collection and scenario testing. It does not by itself demonstrate conformity with ISO 14067 or independent assurance. If a result will support a public comparison, the method, product rules and review requirements need separate examination.
Review the tool's own limitations. One calculator may support cradle-to-gate only; another may use a narrow factor library. Some will accept a supplier PCF and prioritize it over generic emissions or factors, which can be useful if the supplier value is appropriate and versioned. A choice made by the algorithm must be visible in the output if the result is to be audited.
Label the current status honestly: screening estimate, internally reviewed model, independently verified disclosure or another applicable state. Do not let a polished chart make an exploratory result appear more certain than its inputs allow.
Keep the review statement specific to the work performed. An arithmetic check, an internal methodological review and an independent verification have different subjects and evidence. Record who reviewed what, under which scope, and what remains open. IEMA recommends external verification for published footprint results; the recommendation in that guidance is not a universal legal requirement or proof that a particular output has been verified.
Why two calculator totals may not be comparable
Two calculator results can share the same kg CO2e unit and still be unsuitable for direct comparison. One might be per kilogram of material and the other per completed product. One may stop at the factory gate while another models ten years of use. Even if units and boundaries appear similar, factors, allocation, supplier data quality and product-category rules can differ. The GHG Protocol notes that additional product rules may be needed for robust comparative use.
Before publishing “lower than,” align function, performance, duration, reference flow, included stages, gas method, geography, data vintage and review state. Recalculate under a common method where possible. If alignment is not feasible, describe each result within its own scope and avoid a ranking. An attractive side-by-side dashboard is not itself comparison evidence.
A tool can accelerate consistent repeated calculations after the rules and data are settled. It cannot decide whether two products truly serve the same customer function. That remains an evidence and judgment task for the organization using the results.
Even a common standard name does not settle these differences. IEMA's guidance highlights continuing caution around boundaries, inventories and co-product treatment, while the GHG Protocol Product Standard page explains the role of more specific product rules for comparison. If one result cannot be reconstructed or aligned, use it within its own stated purpose instead of converting an unresolved mismatch into a green-versus-red product ranking.
The input-to-output review checklist
The following input-output audit uses the same fictional P-1 teaching record and invented factors. It is an expected-behaviour test, not an observed pass by a named calculator or an actual footprint.
| Case | Actual test inputs | Expected behaviour | Evidence and next owner/action |
|---|---|---|---|
| Valid included lines | A: 200 g/unit; 5 kg CO2e/kg. Batch: 12 kWh; 0.5 kg CO2e/kWh; 3 identical units. | Material 1 kg CO2e/unit; energy 2 kg CO2e/unit; included-lines subtotal 3 kg CO2e/unit. | Analyst retains gram conversion, fictional factors and batch denominator; reviewer reproduces both lines. |
| Missing material mass | A: mass absent; material factor 5 kg CO2e/kg. Electricity inputs unchanged. | Material unresolved; no complete product result. Do not replace missing mass with zero or reuse 200 g. | Product-data owner obtains the component mass and its source/revision, then analyst recalculates. |
| Incompatible unit | A: 0.200 kg/unit paired with 0.5 kg CO2e/kWh. | Reject the material row: kg cannot cancel kWh. No valid material result. | Method owner supplies the correct factor/activity mapping; analyst retests, retaining the rejected input. |
Table 3, shown one row at a time. Every field is retained.
Row 1
- Case
- Valid included lines
- Actual test inputs
- A: 200 g/unit; 5 kg CO2e/kg. Batch: 12 kWh; 0.5 kg CO2e/kWh; 3 identical units.
- Expected behaviour
- Material 1 kg CO2e/unit; energy 2 kg CO2e/unit; included-lines subtotal 3 kg CO2e/unit.
- Evidence and next owner/action
- Analyst retains gram conversion, fictional factors and batch denominator; reviewer reproduces both lines.
Row 2
- Case
- Missing material mass
- Actual test inputs
- A: mass absent; material factor 5 kg CO2e/kg. Electricity inputs unchanged.
- Expected behaviour
- Material unresolved; no complete product result. Do not replace missing mass with zero or reuse 200 g.
- Evidence and next owner/action
- Product-data owner obtains the component mass and its source/revision, then analyst recalculates.
Row 3
- Case
- Incompatible unit
- Actual test inputs
- A: 0.200 kg/unit paired with 0.5 kg CO2e/kWh.
- Expected behaviour
- Reject the material row: kg cannot cancel kWh. No valid material result.
- Evidence and next owner/action
- Method owner supplies the correct factor/activity mapping; analyst retests, retaining the rejected input.
The valid row demonstrates compatible arithmetic and an explicit production denominator. It does not complete shipping, use or disposal, which are unresolved for a full-life-cycle claim. The reviewer must keep the included-lines label with the 3 kg CO2e per-unit subtotal rather than interpreting a green validation indicator as a complete study. A later real-product test needs appropriate real factors and its own evidence; the fictional values must not migrate into that model.
For the missing case, the product-data owner must resolve the absent material quantity. Until then, the complete result remains blocked even though the electricity contribution is calculable. For the wrong-unit case, the method owner must determine whether the quantity, factor or field mapping is wrong. These are different repairs, and neither is solved by adding a generic uncertainty note to a numerical total.
Finish by tracing the chosen output backwards through its stages, quantities, factors, conversions and allocation. Check one influential record against its original evidence and test the export without relying on the live screen. Record pass, fail or unresolved separately for each action, with the next owner. Expected values in this table are not completed review results, approval history or evidence that a provider passed the exercise.
Keep product evidence separate from calculation approval
Keep product-evidence preparation separate from the calculation decision it supports. A useful handoff identifies the product and revision, lists its components and quantities, and keeps supplier documents and review status alongside those records. That organisation makes it possible to ask a precise question about a material or manufacturing line without pretending that a complete-looking file is already a finished footprint.
For a bounded pilot, select one product revision and one influential input. The data owner supplies the original document, its stated unit and period, the relevant component identifier and any correction history. The analyst checks whether it matches the chosen reference flow. The method owner then selects the appropriate factor, boundary and allocation, while the reviewer evaluates the resulting calculation and its limitations. Make each receiving responsibility explicit.
An updated supplier document should create a traceable revision rather than silently changing every historical result. Preserve the earlier source with the version that used it, record why the new information is preferred, and test what the receiving calculator actually imports or exports. If a field is dropped during transfer, resolve that handoff rather than inferring it can be reconstructed from a dashboard image.
This is a neutral workflow requirement, not a claim about a particular provider's current functionality. Evidence organisation does not supply missing factors, establish conformity, perform independent assurance or decide what a customer may claim. Evaluate those tasks separately using the required method and qualified responsibility. The practical outcome is a recoverable evidence package with a known next decision, not an unsupported promise of automatic calculation or certification.
What to do with the first result
Use a product emissions calculator as a controlled implementation of an explicitly defined study. Name the product and service, set the lifecycle boundary, bring source-backed components and activities, select representative factors and check how the tool treats allocation, missing data and revisions. Then reproduce sample lines, reconcile stages and export the evidence trail with a properly labelled result.
If the first output exposes major unknowns, that is useful information. Request better supplier data or a production record before making a strong claim. A transparent provisional number can guide improvement while the study matures. A hidden default that happens to produce a neat number is a reason to pause, not to publish.
The full product-footprint-method guide explains the underlying calculation in more detail. The next practical step here is to run one product through the input-to-output checklist and decide whether the result is only a screening estimate or fit for the intended recipient. That decision should be documented outside the calculator display.
Start with the most consequential unresolved issue rather than adding more product records to the same untested model. If the material mass is unknown, request it; if the denominator is wrong, correct the allocation; if a comparison needs compatible methods, arrange that review. Keep the previous version and document the decision reached. Scaling a calculator is useful only after its evidence and operating rules work on the first product.
Product carbon footprints
