ISO 14067: Carbon Footprinting Without the Checks

ISO 14067:2018 defines how to calculate product carbon footprints (CFPs). Companies use it to quantify GHG emissions across life cycles, from resource extraction through manufacture, use, and disposal.

The standard comes from the ISO 14060 family of GHG standards. Published September 2018, it replaced ISO/TS 14067:2013 and narrowed its scope significantly.

What ISO 14067 Actually Is

ISO 14067 is a subset of life cycle assessment that looks only at climate change. Full LCA (ISO 14044) examines multiple environmental impacts: acidification, eutrophication, resource depletion, water use, toxicity. Carbon footprinting strips that down to a single metric: kilograms of CO₂ equivalent.

This creates a problem. You can optimise for carbon whilst making other environmental impacts worse. A product might have low carbon emissions BUT high water consumption, or minimal GHG emissions BUT significant ecosystem damage from raw material extraction.

ISO 14067 doesn’t claim to be comprehensive environmental assessment. It explicitly states it “addresses only a single impact category: climate change” and “does not assess any social or economic aspects or impacts, or any other environmental aspects”.

The Verification Gap

Here’s where ISO 14067 differs fundamentally from Environmental Product Declarations.

EPDs require mandatory third-party verification:

Independent verifier checks all calculations
Site visits to validate primary data
Multiple rounds of corrections (typically 2-3)
Verification takes 5-20 weeks
Verifier must be accredited
Results published in public registry

ISO 14067 makes verification optional:

Clause 8 states “a critical review facilitates understanding and enhances the credibility of CFP”
Then adds “if any” review shall follow ISO/TS 14071
No requirement for site visits
No mandatory external checking
No public registration
You can calculate a CFP, report it internally or externally, and never have anyone verify your numbers

This isn’t a minor detail. Without mandatory verification, two things happen:

Companies can make calculation errors (allocation mistakes, boundary omissions, data gaps) that never get caught
Companies can make methodological choices that favour their products without independent scrutiny

ISO 14044 (full LCA standard) requires panel review for comparative assertions disclosed to the public. ISO 14067 mentions this requirement exists BUT doesn’t enforce it. The standard says results “may be used for comparisons provided that, at a minimum, the requirements of Annex B… are met” without explaining how anyone checks compliance.

CFP vs EPD: Critical Differences

Aspect	Carbon Footprint (ISO 14067)	EPD (ISO 14025 + EN 15804)
Scope	Climate change only	15+ impact categories
Verification	Optional	Mandatory third-party
Product Category Rules	Optional (via ISO/TS 14027)	Mandatory
Public disclosure	Not required	Required (registered database)
Site visits	Not required	Typically required
Validity period	None specified	5 years maximum
Timeline	2-3 months possible	5-8 months typical
Cost	Lower (no verification)	Higher (£5,000-15,000+)
Comparability	Poor (different methodologies)	Better (PCR standardisation)
Credibility	Depends on voluntary review	Built-in through verification

EPDs use Product Category Rules (PCRs) that specify exact methodologies for product types. All concrete EPDs follow the same PCR, all window EPDs follow another PCR. This enables meaningful comparison within categories.

ISO 14067 allows PCR use through ISO/TS 14027 BUT doesn’t require it. Without PCRs, every CFP study can use different system boundaries, allocation methods, data sources, and assumptions. Results become incomparable.

CFP vs Full LCA

Full LCA under ISO 14044 examines products across multiple dimensions:

Climate change (greenhouse gas emissions)
Acidification (SO₂, NOₓ leading to acid rain)
Eutrophication (nutrient pollution of water bodies)
Ozone depletion (CFCs, halons)
Photochemical ozone creation (smog formation)
Resource depletion (abiotic and biotic)
Water use
Land use
Ecotoxicity (freshwater, marine, terrestrial)
Human toxicity (carcinogenic and non-carcinogenic)
Particulate matter formation

ISO 14067 strips all that away. You get carbon numbers BUT miss everything else. A biofuel might have low carbon emissions BUT destroy peatland ecosystems. Recycled aluminium has lower carbon than virgin BUT the recycling process might generate significant water pollution. ISO 14067 won’t tell you about these trade-offs.

For comprehensive environmental assessment, you need full LCA. For regulatory compliance or narrow decision-making focused solely on carbon, ISO 14067 works. Just recognise what you’re missing.

Core Methodology

System Boundaries

Every CFP study draws a boundary around what gets counted. This decision fundamentally shapes results.

Cradle-to-gate: Raw materials through factory gate (common for business-to-business products) Cradle-to-grave: Full life cycle including consumer use and disposal Cradle-to-cradle: Includes recycling and recovery back into production Gate-to-gate: Just one stage (manufacturing only, excluding upstream materials)

ISO 14067 requires life-cycle-wide boundaries for complete CFPs. Partial CFPs can use narrower boundaries BUT must disclose excluded stages.

The boundary choice matters enormously. For electronics, use-phase electricity often dominates total footprint. A cradle-to-gate study would miss 80%+ of emissions. For packaging, manufacturing and materials dominate whilst use phase contributes negligible emissions.

Without verification, boundary choices go unchallenged. A company could legitimately exclude high-impact stages, report the partial CFP, and most audiences wouldn’t spot the omission.

Functional Unit

Instead of “per product”, CFPs use functional units that define service provided. This enables comparison between products that deliver the same function differently.

Examples:

Not “one light bulb” BUT “10,000 hours of illumination”
Not “one car” BUT “100,000 km of passenger transport”
Not “one paint tin” BUT “coverage of 50 m² to specified durability”

The functional unit choice affects results dramatically. LED bulbs look terrible per unit (high embodied carbon) BUT excellent per 10,000 hours (long lifespan). Choosing the wrong functional unit invalidates comparisons.

Methodological Foundation

Data Quality

ISO 14067 lists data quality requirements: time coverage, geographical coverage, technology coverage, precision, completeness, representativeness, consistency, reproducibility. These sound rigorous BUT the standard doesn’t specify minimum thresholds.

What counts as “sufficient” completeness? 80%? 95%? The standard doesn’t say. What level of precision is needed? Again, undefined. This flexibility allows organisations to self-assess data quality without objective benchmarks.

EPD verifiers check data quality against PCR requirements. CFP studies without verification rely on self-reporting. Two studies might claim “high data quality” whilst using completely different standards.

Data types:

Primary data: Measured from your actual operations (most reliable)
Secondary data: From databases like ecoinvent (generic averages)
Proxy data: Estimates from similar processes (least reliable)

The more secondary and proxy data you use, the less your CFP reflects actual environmental performance. A study using 90% database values tells you about average industry performance, not your specific product.

ISO 14067 requires documenting data sources BUT doesn’t mandate minimum primary data percentages. EPD PCRs often require 75%+ primary data for foreground processes. CFPs can use whatever mix they choose.

Allocation

When industrial processes produce multiple products simultaneously, how do you split environmental burdens? A refinery produces diesel, petrol, kerosene, LPG, and petrochemicals from crude oil. Which product bears what share of emissions?

ISO 14067 follows ISO 14044’s allocation hierarchy:

Avoid allocation by subdividing processes (often impossible)
Physical allocation based on mass, energy, or other physical property
Economic allocation based on revenue or market value

Different allocation methods produce different results. Physical allocation might assign 40% of refinery emissions to diesel. Economic allocation might assign 35% because diesel sells for less per tonne than specialty chemicals.

Both methods comply with ISO 14067. Neither is objectively “correct”. The choice depends on study goals BUT fundamentally shapes results. Without verification, allocation choices go unchallenged.

Biogenic Carbon

Materials from living organisms (crops, wood, biofuels) absorb CO₂ during growth then release it during use or decay. Accounting for these flows creates complexity.

ISO 14067 requires separate reporting:

Biogenic uptake (CO₂ absorbed during plant growth)
Biogenic emissions (CO₂ released during use/decay)
Fossil emissions (from processing biogenic materials)

Example: Wood pellet fuel

Uptake: -10 kg CO₂/kg (tree absorbed carbon)
Processing: +2 kg CO₂/kg (fossil fuel for harvesting, drying, transport)
Combustion: +10 kg CO₂/kg (biogenic release)
Net: +2 kg CO₂/kg (fossil only)

This separate reporting matters for policy. Some carbon accounting frameworks treat biogenic emissions as neutral, others don’t. ISO 14067 provides the data BUT doesn’t make the policy choice.

The standard allows various approaches to land-use change emissions. Converting peatland to palm plantations releases centuries of stored carbon. How much gets attributed to the first harvest cycle versus amortised across future harvests? ISO 14067 permits multiple methods, each giving different answers.

Electricity Treatment

Grid electricity poses challenges. Generation mixes vary by location and time. UK grid at 2pm on a sunny, windy day might be 60% renewable. At 6pm on a cold, still evening it might be 10% renewable.

ISO 14067 permits several approaches:

Consumption mix (average grid emissions where electricity gets used)
Supplier-specific (emissions from your actual electricity supplier)
Contractual instruments (Renewable Energy Guarantees of Origin, Power Purchase Agreements)

Each produces different numbers. UK grid average: ~200 g CO₂/kWh (2024). Your specific supplier if mostly gas-fired: ~400 g CO₂/kWh. Your contractual supply if you buy wind REGOs: ~0 g CO₂/kWh.

Which is correct? All three comply with ISO 14067. The question is what you’re trying to measure. Physical reality (grid mix)? Your supplier relationship? Your financial support for renewables through certificate purchases?

This matters enormously for electric vehicles. An EV charged on UK average grid has ~40% lower lifetime emissions than equivalent petrol car. Charged on 100% renewable certificates, it claims ~70% reduction. The car is physically identical. Only the accounting choice changed.

Recycling and Circularity

Recycling breaks linear product flows. When you recycle aluminium, does the original can producer get credit for enabling recycling? Or does the new product using recycled content?

ISO 14067’s Annex D describes several approaches:

Cut-off method: Recycled material enters the system burden-free. Original product gets no credit for recyclability. The next product benefits from using recycled content.
Avoided burden: Original product gets credited for recycling potential. Recycled content carries the burden of virgin production it avoided.
Closed-loop: For stable recycling systems (like aluminium cans), assume perfect loop. Each cycle bears average production burden.

Example: 1 kg aluminium can

Cut-off approach:

Virgin production: 15 kg CO₂
Using recycled content: 0.5 kg CO₂ (just remelting)
Current can using virgin: 15 kg CO₂
Future can using recycled: 0.5 kg CO₂

Avoided burden:

Current can using virgin: 15 – 14.5 = 0.5 kg CO₂ (credited for avoiding future virgin)
Future can using recycled: 15 kg CO₂ (bears full burden of what it avoided)

Same physical reality. Opposite accounting results. Both valid under ISO 14067.

This creates perverse incentives. Under cut-off, making products hard to recycle carries no penalty. Under avoided burden, using recycled content looks worse than virgin. Neither perfectly reflects environmental reality.

EPD PCRs specify which method to use within product categories. ISO 14067 without PCRs lets each study choose. Comparison becomes meaningless.

Impact Assessment

Converting diverse GHGs into single metric requires characterisation factors. These multiply each gas by its global warming potential (GWP):

CO₂: 1 (reference)
Methane (CH₄): 28-36 depending on IPCC version
Nitrous oxide (N₂O): 265-298
SF₆: 23,500-25,200

ISO 14067 doesn’t mandate which IPCC assessment report to use. Fifth Assessment Report (AR5) uses GWP over 100 years. Sixth Assessment (AR6) revised factors based on updated science. Earlier reports used different values.

Same emissions, different characterisation factors, different results. A CFP calculated using AR4 (2007) shows different numbers than identical assessment using AR6 (2021).

The standard also permits shorter or longer time horizons (20-year, 500-year GWPs). Methane has GWP of 85 over 20 years BUT 28 over 100 years. Short-lived gases matter more over shorter timeframes.

These choices are policy decisions disguised as technical details. They fundamentally alter which products appear preferable.

What ISO 14067 Explicitly Excludes

The standard doesn’t cover:

Carbon offsetting (purchasing credits to compensate emissions)
Communication (how to report or market CFP results – see ISO 14026)
Social impacts (labour conditions, community effects)
Economic impacts (costs, employment, development)
Other environmental impacts (everything except climate)

Most significantly, it excludes carbon offsetting from calculations. You can’t subtract purchased carbon credits from your CFP. The standard quantifies actual product emissions, not net emissions after financial transactions.

This is technically correct (offsetting doesn’t change physical emissions) BUT means CFP and corporate net-zero claims use different accounting. A company can claim net-zero through offsetting whilst their products have high CFPs.

The Comparability Problem

ISO 14067’s Annex A states limitations frankly:

“Because of these limitations, the results of a quantification of the CFP in accordance with this document are often not a sound basis for comparisons.”

Often. Not always. Often.

The standard continues: “However, these results may be used for comparisons provided that, at a minimum, the requirements of Annex B… are met.”

Annex B requires:

Identical product category rules
Equivalent functional units
Equivalent system boundaries
Equivalent data quality
Equivalent allocation procedures
Equivalent impact assessment methods
Equivalent assumptions

In practice, meeting all these requirements means studies must be conducted jointly or one party must adopt the other’s exact methodology. Independent CFP studies rarely meet these criteria.

Without verification, who checks whether Annex B requirements are met? Nobody. Companies can claim comparability whilst using incompatible methods.

Reporting Requirements

CFP study reports must include:

Goal and scope (what you assessed, why, intended audience)
System boundary with diagram
Functional unit definition and justification
Data quality assessment
Allocation procedures with justification
Cut-off criteria (what got excluded as negligible)
Impact assessment method (which IPCC version, time horizon)
Results with biogenic carbon reported separately
Uncertainty analysis
Limitations and assumptions
Interpretation of significant issues

This sounds comprehensive. BUT “must include” doesn’t mean “must be public”. Internal reports can contain all required elements whilst revealing nothing externally.

EPDs get published in searchable databases. Anyone can download full technical details. CFPs might remain entirely confidential.

Even when published, CFP reports often omit commercially sensitive data. You see results BUT not the underlying inventory. Without that data, independent verification becomes impossible.

The Reality of CFP Use

ISO 14067 serves several legitimate purposes:

Internal decision-making: Comparing design alternatives within one organisation where methodological consistency is maintained.
Hotspot identification: Finding high-emission stages in your supply chain. Doesn’t require perfect accuracy, just identifying where emissions concentrate.
Trend monitoring: Tracking your own products over time to measure reduction progress. Comparability across time matters more than absolute accuracy.
Regulatory compliance: Where regulations reference ISO 14067 specifically. The standard provides the required calculation method.
Supplier engagement: Asking suppliers for component footprints. Imperfect data beats no data for supply chain management.
What ISO 14067 doesn’t support well:
Marketing claims: Without verification, claims lack credibility. Consumers can’t distinguish rigorous studies from optimistic guesswork.
Product comparison: Different methodologies make comparison meaningless. “Lower carbon than competitor” only works if both used identical methods.
Financial instruments: Carbon pricing, taxes, or trading schemes need verified data. Unverified CFPs aren’t sufficiently reliable.
Regulatory enforcement: Authorities can’t penalise non-compliance without ability to verify reported values.

CFP vs EPD: When to Use Which

Use CFP (ISO 14067) when:

You need quick carbon screening (2-3 months)
Budget is limited (£3,000-8,000 typically)
Results stay internal for decision-making
You’re comparing your own design alternatives
Regulatory requirement specifically references ISO 14067
You don’t need public credibility (yet)

Use EPD (ISO 14025/EN 15804) when:

You need public credibility and third-party verification
Making comparative public claims
Operating in construction sector (often mandatory)
Need to demonstrate compliance rigorously
Willing to invest time (5-8 months) and money (£8,000-20,000)
Want comprehensive environmental assessment (15+ indicators)
Results will appear in tenders or procurement

Use full LCA (ISO 14044) when:

You need to understand trade-offs between environmental impacts
Carbon alone gives misleading picture
Designing products where multiple impacts matter
Academic research or detailed environmental assessment
No specific regulatory framework requires narrow focus
You have time and expertise for complex analysis

Practical Implementation

If you’re conducting a CFP study:

Phase 1: Define scope carefully

Be honest about boundaries. Excluding difficult stages doesn’t make emissions disappear.
Choose functional unit that enables meaningful comparison.
Document why you made each methodological choice.

Phase 2: Prioritise data quality

Use primary data where emissions concentrate.
Database values are fine for minor contributors.
Document data sources transparently.
Acknowledge where you used estimates.

Phase 3: Model systematically

Use proper LCA software (openLCA, SimaPro, GaBi).
Don’t build CFPs in spreadsheets unless very simple.
Check mass balances (inputs should roughly equal outputs).
Verify energy calculations against known benchmarks.

Phase 4: Interpret honestly

Uncertainty often exceeds ±20%. Claiming 3.47 kg CO₂ suggests false precision.
Identify hotspots (stages contributing >10% of total).
Run sensitivity analysis on key assumptions.
Don’t overstate certainty.

Phase 5: Report transparently

Include all required elements per Clause 7.
Explain limitations clearly.
If using for comparison, verify Annex B compliance.
Consider voluntary verification to build credibility.

Common Errors

Boundary errors: Excluding significant life cycle stages (especially use-phase for energy-using products).

Allocation errors: Using economic allocation when physical would be more appropriate, or vice versa.

Double counting: Including both recycling credit at end-of-life AND using recycled content at beginning-of-life.

Data mixing: Combining recent data with outdated emission factors, or mixing geographical contexts inconsistently.

Temporal mismatch: Using current grid factors for products manufactured years ago.

Comparability claims: Stating “lower carbon than X” without verifying methodological equivalence.

False precision: Reporting 2.347 kg CO₂ when uncertainty is ±0.5 kg.

Missing biogenic carbon: Failing to report biogenic flows separately as required.

Software and Databases

Conducting credible CFPs requires:

LCA software:

openLCA (free, open-source)
SimaPro (commercial, widely used)
GaBi (commercial, industry standard)
Umberto (commercial, process-focused)

Life cycle databases:

ecoinvent (most comprehensive, subscription)
GaBi databases (integrated with software)
ELCD (European Life Cycle Database, free)
US LCI (US-focused, free)
Agribalyse (agricultural products, free)

Database quality varies significantly. ecoinvent provides well-documented, peer-reviewed data. Free databases often lack transparency about data sources and quality.

The ISO 14067 Dilemma

ISO 14067 finds itself in an awkward position. It wants rigour (detailed methodology, documentation requirements, transparency) BUT makes verification optional.

This creates two classes of CFP:

Serious CFPs: Conducted rigorously, independently reviewed, meet all requirements, transparently documented.
Compliance CFPs: Follow the letter of the standard (all boxes ticked) BUT choose assumptions favouring desired outcomes, never verified, remain opaque.

Both claim conformity with ISO 14067. Outsiders can’t distinguish between them.

EPDs solved this by mandatory verification. ISO 14044 solved it by requiring panel review for comparative assertions. ISO 14067 acknowledged the problem (Annex A’s limitations) BUT didn’t enforce solutions.

Future Development

Climate policy increasingly requires quantified emissions data. EU’s Carbon Border Adjustment Mechanism (CBAM), product carbon footprint regulations, and green procurement criteria all demand calculated values.

ISO 14067 provides methodology BUT its optional verification creates loopholes. Future revisions will likely face pressure to:

Mandate verification for specific applications
Require PCR use for comparative studies
Set minimum data quality thresholds
Improve alignment with EPD standards
Address emerging issues (scope 3 boundaries, carbon removal accounting)

Until then, CFPs under ISO 14067 range from excellent to unreliable with no easy way to tell which is which.

Key Takeaways

ISO 14067 defines carbon footprinting methodology BUT makes verification optional. This creates credibility gap compared to EPDs.
CFPs measure only climate change. Full LCA examines 15+ environmental impacts. Carbon optimisation can worsen other impacts.
Methodological choices matter enormously. System boundaries, allocation, electricity treatment, recycling approaches all fundamentally shape results.
Comparability is problematic. Different CFP studies usually aren’t comparable unless using identical methods.
Use CFPs for internal decisions, EPDs for public claims. If you need third-party credibility, the optional verification of ISO 14067 isn’t sufficient.
The standard acknowledges its own limitations. Annex A explicitly states results “often” aren’t suitable for comparison.

ISO 14067 provides valuable methodology for carbon accounting. Used carefully with appropriate caveats, it supports better environmental decisions. Used carelessly or manipulatively, it generates misleading numbers that drive poor choices.

The difference between good and bad CFPs comes down to intellectual honesty, methodological rigour, and transparent documentation. Without mandatory verification, those qualities depend entirely on the organisation conducting the study.

Choose accordingly.