PCF Frameworks Compared — Methodology and Selection Approach
A practitioner’s reference to the standards that govern product carbon footprints: ISO 14040/14044, ISO 14067, the GHG Protocol Product Standard, EU PEF v3.1, the WBCSD Pathfinder Framework, and the ISO 14025 / EN 15804 EPD route. Covers what each framework is for, the five decision axes that actually separate them, how they nest rather than compete, how to choose one, and — the part most comparisons skip — how to satisfy several at once without producing incomparable numbers.
There is no single product carbon footprint standard, and that is not an accident. The frameworks in this comparison are layers in a stack, not rivals in a race: one supplies the foundation method, one makes it carbon-specific, one governs supplier-to-customer data exchange, one is a regulatory regime, and one is the route to a verified environmental product declaration. Choosing well means understanding what layer your PCF needs to satisfy — and reconciling carefully when it needs to satisfy more than one.
Use ISO 14067 for a standalone product footprint; the GHG Protocol Product Standard when the PCF feeds Scope 3; WBCSD Pathfinder when exchanging a cradle-to-gate PCF with a customer; EU PEF for a regulated claim; ISO 14025/EN 15804 when the deliverable is a verified EPD.
Why There Are Multiple PCF Frameworks
The proliferation of product carbon footprint frameworks confuses newcomers because it looks like duplication. It is not. Each framework answers a different question about the same underlying inventory, and they were written by different bodies for different audiences over two decades.
- The foundation standards — ISO 14040/14044 — define life-cycle assessment in general: goal and scope, inventory, impact assessment, interpretation. They are method, not carbon.
- The carbon-specific standard — ISO 14067 — narrows that general LCA method to the single impact category of climate change, adding rules for biogenic carbon, land-use change, and carbon-specific data quality.
- The corporate-reporting bridge — the GHG Protocol Product Standard — frames the same product footprint so it can roll up into an organisation’s Scope 3 Category 1 inventory.
- The data-exchange protocol — the WBCSD Pathfinder Framework (PACT) — standardises how a supplier hands a cradle-to-gate PCF to a customer so it can be aggregated without re-modelling.
- The regulatory regime — EU PEF v3.1 — prescribes a fixed method (including a mandatory recycling formula) for environmental claims in the EU single market.
- The declaration route — ISO 14025 Type III EPDs, operationalised for construction by EN 15804 — turn an LCA into a verified, comparable, third-party-registered declaration.
Read that way, the “competing standards” problem dissolves. A European building-products manufacturer might legitimately need all six: ISO 14044 as the method, ISO 14067 for the carbon rules, EN 15804 for the EPD it must publish, Pathfinder to send the figure to a customer, the GHG Protocol Product Standard so the customer can book it as Scope 3, and PEF if it makes a marketing claim. The discipline is not picking one — it is keeping the results these frameworks produce clearly labelled so they are never silently compared across incompatible rules.
Provider directory
Need help keeping this up all year?
See who does this work. Every listing names the standards it works to, and paid placements are labelled. Including Attain Zero and Pangolin Associates.
Browse 6 carbon accounting & inventory providers →Do this work? A listing is US$390 a year. Get listed →
The Framework Landscape at a Glance
The master matrix below is the centrepiece of this page: seven frameworks across the dimensions that determine whether two PCFs are comparable. Read it as a reference to return to, not a sequence — the deep treatment of each axis follows in § decision axes.
| Framework | Role in the stack | Default boundary | Recycling rule | Primary-data emphasis | Third-party verification | Comparability target |
|---|---|---|---|---|---|---|
| ISO 14040/14044 | Foundation LCA method | Defined per goal & scope | Practitioner choice (closed/open-loop) | Required where it affects results | Critical review for comparative assertions | Within a single study |
| ISO 14067 | Carbon-specific PCF standard | Cradle-to-gate or cradle-to-grave (declared) | Inherits ISO 14044; carbon tracked through loops | Required; data-quality assessment mandated | Recommended, not mandatory | Within a declared PCR |
| GHG Protocol Product | Corporate value-chain bridge | Cradle-to-grave (full lifecycle preferred) | Document recycled-content & end-of-life where material | Prioritised; uncertainty assessment expected | Assurance recommended | Year-on-year, within a company |
| EU PEF v3.1 | Regulatory claim regime | Cradle-to-grave, fixed by PEFCR | Mandatory Circular Footprint Formula | Prescribed by category rules (DNM/company-specific) | Mandatory for public claims | Across products in a category |
| WBCSD Pathfinder | Supplier-to-customer exchange | Cradle-to-gate (business-to-business) | Cut-off (recycled content) default | Primary supplier data strongly required | Encouraged for exchanged PCFs | Across a value chain |
| ISO 14025 / EN 15804 | Verified declaration (EPD) route | Modular (A1–A3 to A–D under EN 15804) | Module D (loads & benefits beyond boundary) | PCR-prescribed; background data permitted | Mandatory third-party verification | Across EPDs under one PCR |
| ecoinvent system models | Background LCI dataset | Supplies upstream unit processes | Cut-off / APOS / consequential variants | Background (secondary) by definition | Database-level, not study-level | Within a chosen system model |
The single most useful column in this matrix is comparability target. Two PCFs are only comparable if they share that target — two EN 15804 EPDs under the same product category rule are comparable; an ISO 14067 cradle-to-gate figure and a PEF cradle-to-grave figure are not, even for the same product. Comparability is a property of the framework boundary, not of the number.
The Five Decision Axes That Actually Separate Them
Strip away branding and the frameworks differ along five axes. Two PCFs that align on all five are genuinely comparable; a mismatch on any one breaks comparability even when the headline numbers look similar.
1. System boundary — where the assessment starts and stops
The boundary is the largest single driver of divergence. Cradle-to-gate covers raw-material extraction through the factory gate and is the default for business-to-business exchange (Pathfinder) because it ends where the supplier’s control ends. Cradle-to-grave adds distribution, use, and end-of-life, and is preferred by the GHG Protocol Product Standard and mandated by most PEF category rules. EN 15804 takes a third path entirely: a modular boundary (information modules A1–A3 for product stage, A4–A5 for construction, B for use, C for end-of-life, and the separate module D for loads and benefits beyond the system boundary). A cradle-to-gate number and a cradle-to-grave number for the same product are different physical quantities — comparing them is the most common framework error.
Boundary depth is set by the framework and the product category rule, not chosen freely per study. The detailed boundary mechanics live on the cradle-to-gate and cradle-to-grave methodology pages.
2. Allocation and recycling — splitting shared and circular burdens
All the frameworks inherit the ISO 14044 allocation hierarchy (avoid → physical → economic), but they diverge sharply on recycling. Pathfinder and the ecoinvent cut-off model default to cut-off (a product carries only the burden of the material it contains); EN 15804 routes recycling credits to module D; and PEF mandates the single fixed Circular Footprint Formula. In the Master Brain, the cut-off recycled-content convention sits at a fixed position in the special-case allocation ranking — 7 — which is why it serves as the reproducible default for exchanged PCFs. The full treatment is on the PCF allocation methodology page; for this comparison the point is that recycling convention is framework-determined, and a cut-off result and a module-D result are not interchangeable.
3. Data quality and primary-data rules
Frameworks differ in how hard they push toward primary (supplier-measured) data versus secondary (database) data. Pathfinder is the most demanding on primary data because an exchanged PCF that leans on generic database values transmits its uncertainty straight into the customer’s value-chain total. ISO 14067 mandates a documented data-quality assessment but does not fix a single scale. PEF prescribes data-quality requirements per category rule, distinguishing situations that must use company-specific data from those where a default may be used. The background dataset most studies rely on — ecoinvent — is secondary by definition, which is why its system-model choice (below) matters so much.
4. Verification and comparability
Verification rigour rises with the claim’s public weight. ISO 14067 recommends but does not mandate third-party verification; the GHG Protocol Product Standard recommends assurance; EPDs under ISO 14025 / EN 15804 require mandatory independent verification; and PEF requires verification for public claims. Crucially, verification establishes that a PCF was computed correctly within its framework — it does not make two differently-scoped PCFs comparable. A verified cradle-to-gate figure and a verified cradle-to-grave figure are both correct and still not comparable.
5. Attributional versus consequential
The deepest axis is the modelling mode. Almost all PCF reporting — ISO 14067, GHG Protocol Product, Pathfinder exchange, EN 15804 EPDs — is attributional: it partitions a fixed pool of observed emissions among products. The ecoinvent consequential system model, by contrast, models the marginal effect of producing one more unit and is built on substitution. Mixing the two — pulling consequential background data into an attributional PCF — is a subtle but serious error that can produce implausibly low or negative figures. The attributional/consequential distinction is developed on the material substitution methodology page.
Framework by Framework
Each framework in one place: what it is, where it sits in the stack, and when to reach for it.
ISO 14040 / 14044
The foundation LCA standards. They define the four-phase method (goal & scope, inventory, impact assessment, interpretation) and the allocation hierarchy that every framework below inherits. Not carbon-specific and not a reporting format — the grammar, not the sentence. You rarely “report under” ISO 14044 alone; you report under a framework that cites it.
ISO 14067
The dedicated product carbon footprint standard. Takes ISO 14044’s general method and narrows it to climate change, adding carbon-specific rules: biogenic carbon (−1/+1), land-use change amortisation, carbon storage permanence, and a mandatory data-quality assessment. The default choice for a standalone, defensible product carbon footprint.
GHG Protocol Product Standard
Frames a product footprint so it aggregates into corporate Scope 3 Category 1 reporting. Prefers full cradle-to-grave coverage and emphasises consistency with the company’s wider inventory. Reach for it when the PCF must reconcile with corporate value-chain accounting, not stand alone.
EU PEF v3.1
A prescriptive regulatory method for environmental claims in the EU single market. Fixes the boundary, the recycling formula (the Circular Footprint Formula), and the data-quality requirements through product-category rules (PEFCRs). The most constrained framework — by design, to make claims comparable and dispute-proof within a category.
WBCSD Pathfinder (PACT)
A data-exchange protocol, not an accounting standard in its own right. Standardises how a supplier computes and transmits a cradle-to-gate PCF so a customer can aggregate it without re-modelling. Demands primary data, defaults to cut-off recycling, and discourages substitution — all to keep exchanged figures additive across a value chain.
ISO 14025 / EN 15804 (EPD)
The route to a verified, registered environmental product declaration. ISO 14025 sets the Type III EPD framework; EN 15804 operationalises it for construction products with a modular boundary (A1–A3 … C, plus module D). Mandatory third-party verification and a governing PCR make EPDs the most rigorously comparable PCF format — within a single PCR.
How They Nest: The Inheritance Stack
The clearest way to hold the landscape in mind is as a vertical stack of inheritance. Each layer adds rules on top of the layer beneath; nothing replaces what is below it.
Goal & scope, inventory, impact assessment, interpretation, and the allocation hierarchy. Everything above inherits this verbatim.
Narrows LCA to climate change; adds biogenic carbon, land-use change, permanence, and carbon data-quality rules on top of the foundation.
Pathfinder standardises supplier-to-customer transfer; the GHG Protocol Product Standard makes the footprint roll up into corporate Scope 3. Both sit on the carbon layer.
Fixes boundary, recycling formula, and data-quality rules through PEFCRs so claims are comparable within a category. Constrains the freedom the layers below allow.
Turns the LCA into a registered, third-party-verified declaration under a product category rule. The modular boundary and mandatory verification make EPDs the most comparable PCF format within one PCR.
The instinct to ask “which framework is best?” is the wrong question, because the frameworks occupy different layers and a serious PCF programme uses several. The question that actually determines your number is “which boundary, which recycling rule, and which data-quality bar apply to this deliverable?” Two analysts using the same framework but different boundaries will disagree more than two analysts using different frameworks at the same boundary. Lead with the deliverable’s purpose, let that fix the boundary and recycling rule, and the framework choice usually follows.
Choosing a Framework: A Decision Path
Framework choice is driven by the deliverable’s purpose, not by preference. The path below maps the common purposes to the framework whose rules they invoke.
Allocation choices feed every framework above. See the companion PCF allocation methodology for the decision that most often moves the number →
The Interoperability Problem: One Inventory, Many Frameworks
The question practitioners actually face is rarely “which framework?” — it is “I have to satisfy two or three at once; how do I avoid producing incomparable numbers?” The good news is that the shared ISO 14044 foundation means the underlying inventory is largely reusable. The risk is that the divergent boundary, recycling, and data-quality rules quietly produce results that look like the same metric but are not.
- Build one foreground inventory, label every result. Model the product once at the finest boundary you need, then derive each framework’s figure as a clearly-labelled cut of that inventory — never a single number claimed to satisfy all.
- Hold the boundary constant before comparing. Only compare results that share a boundary. A Pathfinder cradle-to-gate and a PEF cradle-to-grave for the same product are different quantities by construction.
- Apply each recycling rule separately. Compute the cut-off result for Pathfinder, the module-D result for the EPD, and the Circular Footprint Formula result for PEF as distinct outputs. Do not seek one recycling number that “passes” all three.
- Keep one consistent GWP basis and biogenic convention. Use the same AR-basis GWP set and the same biogenic treatment across all framework cuts so the only differences are the framework rules, not the underlying science.
- Match background system model to foreground. Keep ecoinvent cut-off background with attributional foreground; never let a consequential background leak into an attributional report.
- Record the data-quality bar per framework. Pathfinder’s primary-data expectation and PEF’s category-specific requirements differ; document which dataset met which bar so a verifier can trace each figure.
Done this way, a single modelling effort yields a family of correctly-labelled, framework-specific figures. The failure mode — and the one verifiers and customers catch — is collapsing that family into one number and presenting it as universally valid. It never is.
Worked Example: One Product, Four Framework Results
The same product, modelled once, reported under four frameworks. All figures are fixed — this is an illustrative audit record at the snapshot date of this methodology revision, not a live calculation, and the inputs are deliberately simple to isolate the effect of framework rules.
A 1 kg aluminium component uses 40% recycled and 60% primary metal. Assume an illustrative primary aluminium burden of 10.0 kg CO₂e/kg and recycled (remelt) burden of 0.5 kg CO₂e/kg; downstream transport, use, and end-of-life add an illustrative 2.0 kg CO₂e/kg over the full lifecycle; and at end of life 90% of the component is recyclable, displacing primary metal. The four frameworks split these facts differently — and publish four different numbers for the same physical part. (Burdens are illustrative placeholders, not Master Brain factors — a live PCF would draw cradle-to-gate metal figures from the embodied-carbon factor library.)
| Framework | Boundary | Recycling treatment | Reported figure (kg CO₂e/kg) |
|---|---|---|---|
| Pathfinder (cradle-to-gate, cut-off) | Extraction → gate | Cut-off: 0.60×10.0 + 0.40×0.5 = 6.20 | 6.20 |
| ISO 14067 (cradle-to-grave, cut-off) | Full lifecycle | 6.20 cradle-to-gate + 2.00 downstream | 8.20 |
| EN 15804 EPD (A1–A3 + module D) | Modular, product stage shown | A1–A3 = 6.20; module D credit reported separately | 6.20 (A1–A3); D shown apart |
| End-of-life credited (avoided burden) | Cradle-to-grave + EoL credit | 8.20 − (0.90×10.0 × credit share) → lower net | < 8.20 (credit-dependent) |
The cradle-to-gate figure (6.20) and the cradle-to-grave figure (8.20) differ purely because of boundary — the downstream 2.00 kg CO₂e/kg is in one and not the other. The EN 15804 EPD reports the same 6.20 for the product stage but pushes the recycling credit into a separate module D rather than netting it, so its headline product-stage number is not directly comparable to the end-of-life-credited figure, which nets the credit into a single lower number. Four valid figures, four different rules, one physical part — and the only way to compare any two is to first confirm they share a boundary and a recycling convention.
None of these is “the” footprint of the part. Each is the footprint under a stated set of framework rules. The lesson the example exists to teach is the one a verifier applies first: before comparing two PCF numbers, confirm they share a boundary and a recycling convention — otherwise the comparison is meaningless however precise the figures look.
Comparability Traps
Each trap below is a case where two PCFs look comparable but are not — the failure mode that framework-comparison work exists to prevent.
| Trap | Why it misleads | How to avoid |
|---|---|---|
| Boundary mismatch | Comparing a cradle-to-gate figure with a cradle-to-grave figure for the same product — the lower one merely stops earlier. | State the boundary with every figure; only compare like boundaries. |
| Recycling-convention mismatch | A cut-off PCF and a module-D or end-of-life-credited PCF treat the same recycling loop differently. | Confirm both use the same recycling rule before comparing; never net across conventions. |
| Framework-label illusion | Two figures both labelled “ISO 14067” can still diverge if they used different boundaries or PCRs. | Comparability requires a shared boundary and category rule, not just a shared framework name. |
| Verified ≠ comparable | Assuming two independently-verified PCFs are comparable. Verification confirms correctness within a framework, not across frameworks. | Treat verification as an integrity check, not a comparability guarantee. |
| Background system-model mismatch | One PCF built on ecoinvent cut-off, another on consequential — different upstream rules invisible in the headline number. | Record the background dataset and system model; match it to the foreground mode. |
| Mixed GWP basis | One PCF on AR5 GWPs, another on AR6 — the same inventory, a different metric. | Disclose the GWP set per figure; reconcile to one basis before comparing. |
| Module-D netting | Treating an EN 15804 EPD’s module-D credit as if it were part of the A1–A3 product-stage number. | Keep module D separate; compare product-stage to product-stage. |
| Consequential figure in an attributional comparison | A substitution-credited (often negative) figure compared against partitioned attributional figures. | Never mix modelling modes in one comparison; label any consequential result. |
Documenting Your Framework Choice
Because the framework choice silently sets the boundary, recycling rule, and data-quality bar, the choice must be recorded as explicitly as any calculation input. A PCF that survives verification carries the following framework-decision record.
- Declared framework and version. The exact framework and edition (e.g. ISO 14067:2018, PEF v3.1, EN 15804+A2), plus the governing product category rule where one applies.
- Stated boundary. Cradle-to-gate, cradle-to-grave, or the EN 15804 modules included — named, not implied.
- Recycling convention. Cut-off, end-of-life/avoided burden, module D, or the Circular Footprint Formula, with the framework that mandates it.
- GWP basis and biogenic convention. The AR-basis GWP set and the biogenic-carbon treatment, held constant across any multi-framework cuts.
- Background dataset and system model. The LCI source (e.g. ecoinvent) and the system model (cut-off / APOS / consequential), confirmed consistent with the foreground.
- Data-quality bar met. The primary/secondary data split and the framework’s data-quality requirement, documented per material input.
- Comparability statement. An explicit note of what this PCF may and may not be compared against — the single most useful line for any downstream reader.
The comparability statement deserves emphasis: it is the artefact that prevents a correctly-computed PCF from being misused downstream. A figure that travels without a statement of its boundary, recycling rule, and framework will eventually be compared against something it should not be — and the resulting claim becomes a CO₂e comparison that fails under scrutiny.
How the Framework Choice Flows into the Calculator
A PCF calculator can compute each framework’s figure once the framework rules are set — but, exactly as with allocation, it cannot choose the framework for you. The division between mechanical computation and analyst judgement is the same one that governs every methodology on this platform.
What a calculator handles
Applying the chosen boundary to include or exclude downstream stages; applying the framework’s recycling rule (cut-off, module D, or the Circular Footprint Formula); holding the GWP basis and biogenic convention constant; and producing the family of clearly-labelled per-framework figures from one foreground inventory.
What you must decide first
Which framework the deliverable invokes; the boundary that framework and its PCR require; the recycling convention; whether the study is attributional or consequential; the background dataset and system model; and the data-quality bar each material input must meet. None are computable — they are documented judgements that the framework choice fixes.
The boundary mechanics behind these decisions are detailed on the ISO 14067 cradle-to-gate and cradle-to-grave methodology pages, and the allocation and recycling rules on the PCF allocation methodology page. In every case the figure a calculator returns is only as comparable as the framework decisions documented before the inputs were entered.
Framework choice and allocation are the two judgements that set a PCF before any number is computed. The allocation methodology covers the decision that most often moves the result; the cradle-to-gate methodology sets the boundary these framework rules sit within.
PCF methodology series. Core methods: cradle-to-gate, cradle-to-grave, allocation, frameworks comparison, sensitivity & hotspots. Sector methods: food, apparel & textile, electronics, packaging, beverages.
Frequently Asked Questions
Neither — they answer different questions and are routinely used together. ISO 14067 is the carbon-specific method for computing a defensible product footprint; the GHG Protocol Product Standard frames that footprint so it aggregates into corporate Scope 3 reporting. Many organisations compute a footprint that is ISO 14067-aligned in method and GHG-Protocol-aligned in reporting boundary. The choice is not which one to adopt but which rules each one sets for your specific deliverable. They share the ISO 14044 foundation, so the underlying inventory is largely common to both.
No, and trying to is the central mistake this page exists to prevent. The frameworks diverge on boundary, recycling rule, and data-quality requirements, so a single number cannot legitimately satisfy all of them. What you can do — and should — is build one foreground inventory and derive each framework’s figure as a clearly-labelled cut of it. A Pathfinder cradle-to-gate figure, a PEF cradle-to-grave figure, and an EN 15804 EPD product-stage figure for the same product are different quantities by construction. Report them as a labelled family, not as one universal value.
Mainly boundary and recycling. PEF mandates a cradle-to-grave boundary fixed by the product category rule and a specific recycling formula (the Circular Footprint Formula); Pathfinder defaults to a cradle-to-gate boundary and cut-off recycling for business-to-business exchange. So the PEF number includes downstream use and end-of-life that the Pathfinder number stops short of, and the two treat recycling loops under different rules. Both can be correct simultaneously. The divergence is a feature of the frameworks’ different purposes, not an error in either calculation.
No — this is one of the most common misconceptions. Verification confirms that a PCF was computed correctly within its own framework, boundary, and category rule. It says nothing about whether two PCFs share those things. A verified cradle-to-gate figure and a verified cradle-to-grave figure are both correct and still not comparable, because they cover different parts of the lifecycle. Treat verification as an integrity check on a single figure, not as a licence to compare it against another verified figure with different scope.
An EPD (under ISO 14025, operationalised for construction by EN 15804) is a verified, registered, third-party-checked declaration governed by a product category rule, with a modular boundary that reports the product stage and end-of-life separately and pushes recycling credits into a distinct module D. A corporate product carbon footprint under ISO 14067 or the GHG Protocol Product Standard is typically a single climate-change figure framed to support reduction decisions or Scope 3 reporting, and it does not require third-party registration. The EPD route is more rigorous and more comparable within its PCR, but heavier to produce; the corporate-PCF route is more flexible and faster but carries weaker comparability guarantees.
Because “background” data still carries embedded allocation and recycling rules, and those must match your foreground. ecoinvent ships in cut-off, APOS, and consequential system models — and each treats every recycled and co-produced input in the supply chain differently. A cut-off foreground PCF built on a consequential background is internally inconsistent: the foreground partitions a fixed pool while the background models marginal substitution. Choosing the system model is therefore a methodological decision, not a database setting. For attributional PCFs, the cut-off system model is the usual match.
Framework choice sets the rules; allocation is one of the decisions those rules constrain. Every framework here inherits the ISO 14044 allocation hierarchy, but they differ on the recycling convention specifically — Pathfinder defaults to cut-off, EN 15804 uses module D, PEF mandates the Circular Footprint Formula. So you choose the framework first, which narrows your allocation and recycling options, then make the allocation decisions within that constraint. The allocation decision itself — mass versus economic, co-product versus waste, the recycling fork — is covered in full on the dedicated PCF allocation methodology page.