TL;DR: The most damaging LCA failures aren’t calculation errors — they’re boundary condition mismatches that make two technically correct assessments incompatible with each other.
TL;DR: A functional unit defined at the wrong level of specificity can shift a packaging system’s carbon footprint result by 40–60% without any underlying data changing.
Where LCA Results Go Wrong Before the Calculation Starts #
The specification that drives more LCA failures than any other is functional unit definition — not data quality, not emission factor selection, not even system boundary setting, though all three matter. A functional unit that seems reasonable on its face can produce results that are structurally incomparable to any benchmark, any supplier’s declaration, or any prior study.
Per ISO 14044:2006 clause 4.2.3, the functional unit must quantify the performance of a product system in a way that allows comparison. The clause is technically clear. In practice, we see briefs that define the functional unit as “one packaging unit” without specifying fill weight, protection level, transport distance, or end-of-life route. Strip those out and you have a number that cannot be challenged, verified, or used.
The second compounding parameter is system boundary. ISO 14040:2006 requires explicit documentation of which life cycle stages are included. When a brand submits an LCA summary to us showing a carbon intensity of 1.2 kg CO₂e per unit, and our internal estimate for the same packaging reads 1.9 kg CO₂e, the gap is almost never the emission factors. In our experience reviewing incoming LCA documentation over the past three years, the gap traces to boundary exclusion roughly 70% of the time — specifically, upstream raw material extraction and end-of-life transport being silently dropped.
For paper-based packaging, upstream pulp and fibre processing typically contributes 35–55% of cradle-to-gate emissions. Exclude it and your result looks competitive. Include it and the trade-off between virgin coated board and recycled board looks completely different.
Supplier Qualification — What to Request and What the Response Reveals #
When we qualify a substrate supplier for LCA-relevant procurement, the first request is a product-specific Environmental Product Declaration (EPD) per EN 15804:2019+A2, not a general corporate sustainability report. The distinction matters because corporate reports aggregate across product grades. An EPD scopes the specific board grade, the specific mill, and the specific production year.
Ask for the EPD reference number and the verification body. Response time under 5 working days and a third-party verified EPD with a declared unit in kg CO₂e per tonne typically signals a supplier with a functioning LCA data infrastructure. A response that delivers a PDF of their ESG commitments instead tells you their emission factor documentation is not production-ready.
For printing inks, request Safety Data Sheets cross-referenced against REACH Regulation (EC) No 1907/2006, and separately ask for VOC content per kilogram of ink. UV-curable inks typically run 0–5% VOC by weight; solvent-based flexo inks range from 55–75% depending on formulation. Both the emissions profile and the allocation method for shared press runs need to be specified before any footprint calculation is valid.
One test we find consistently revealing: ask a potential substrate supplier for their Scope 3 upstream emission intensity for the specific grade you’re procuring. A supplier that can answer at grade level, not company level, has the data discipline to support product-level LCA work. A supplier that answers at company level is asking you to do allocation work they haven’t done themselves.
Cost-Performance Trade-offs in LCA Data Quality #
There is a real cost spectrum here, and the right position on it depends on what the LCA output will be used for.
| LCA Data Tier | Typical Cost Range (per study) | Emission Factor Source | Suitable For |
|---|---|---|---|
| Screening LCA (internal) | USD 2,000–6,000 | Ecoinvent v3 background data | Early-stage material comparison |
| Simplified EPD-backed LCA | USD 8,000–18,000 | Primary supplier EPDs + Ecoinvent | Supplier claims, tender qualification |
| Full ISO 14044-conformant study | USD 25,000–60,000 | Primary data + third-party verification | Public declarations, regulatory submissions |
| Product Carbon Footprint per PAS 2050 | USD 15,000–35,000 | Primary + secondary with uncertainty analysis | Carbon labelling, offset purchasing |
The counterargument to spending on full ISO 14044 conformance: if the packaging is not being publicly declared, and the primary use is internal material selection between two substrates, a screening LCA using Ecoinvent v3.9 background data is often sufficient and appropriate. Over-investing in data quality for a decision that only needs directional guidance is a real budget failure mode, not just the inverse.
Where the calculus changes is when an LCA result will be used in a claim to end consumers or submitted to a retailer’s supplier sustainability programme. At that point, third-party critical review per ISO 14044 clause 6.2 is not optional. Brands that submit unreviewed LCAs to retailer programmes risk disqualification, and we have flagged this directly to partners before they submitted.
Failure Mode Analysis — Allocation Errors in Multi-Output Packaging Processes #
Allocation is where most production-level LCA errors occur, and it is also the topic where opinions differ most sharply among practitioners.
The failure mode: a folding carton converting line runs multiple SKUs. The LCA for one SKU allocates 100% of the press setup energy and substrate waste to that job. Result: the carbon footprint per unit is overstated by a factor that scales with job size. On a 10,000-unit run versus a 100,000-unit run on the same substrate, the per-unit setup allocation difference alone can be 0.03–0.08 kg CO₂e per unit, depending on press configuration and substrate weight.
ISO 14044 clause 4.3.4 outlines three allocation approaches: physical causality (e.g., mass or volume), economic value, or system expansion. Physical allocation by output mass is the most common in packaging LCA. Economic allocation is common in chemical and co-product industries. System expansion is theoretically preferred by ISO but practically difficult when co-products aren’t well-defined.
Where practitioners diverge: some LCA consultants apply economic allocation for all shared press runs, arguing that a premium SKU with a higher revenue per unit should carry more of the shared environmental burden. Others reject this entirely, preferring mass allocation on the grounds that physical causality is more stable across market conditions. Our internal procedure — what we document under Form QE-09 in our LCA data collection workflow — applies mass-based allocation as default, with economic allocation flagged as a sensitivity scenario when a study involves mixed-value product lines.
The practical detection threshold: if two independently conducted LCAs of the same packaging format differ by more than 15% at the cradle-to-gate boundary, the first diagnostic step is to compare allocation method, not emission factors. Allocation method mismatch explains more divergence than emission factor vintage in our experience reviewing third-party studies.
One open question we are still tracking: how to handle waste trim allocation across a production line running both FSC-certified and non-certified board simultaneously. Current FSC chain-of-custody rules (FSC-STD-40-004) handle mass balance for fibre certification, but they don’t resolve the emission allocation question for shared waste streams. Our current practice is conservative — we allocate all trim waste emissions to the certified product to avoid any understatement — but we expect clearer guidance as the FSC’s LCA working group develops its supplemental methodology.
Specification Notes for Brand Partners #
When you brief us on a packaging project that requires carbon footprint data or LCA support, the first thing we need is a clearly scoped functional unit: what the pack contains, fill weight or volume, the protection performance required, and the intended end-of-life route in your primary market. Without these four parameters, any footprint estimate we provide is a rough directional figure, not a number you can use in a supplier declaration or retailer submission.
The most common gap we see in incoming briefs is the absence of end-of-life routing assumptions. A 350 gsm folding carton destined for kerbside recycling in Germany and the same carton going to general waste in Southeast Asia have different cradle-to-grave footprints. The difference in end-of-life stage emissions alone can be 0.04–0.12 kg CO₂e per unit depending on transport distance to processing.
Our standard timeline for a product-level carbon footprint summary using existing EPD data from qualified suppliers is 10–15 working days from receipt of a complete brief. If primary data collection is required from new suppliers, allow 25–35 working days. Studies requiring third-party critical review for public declaration add a further 3–6 weeks depending on reviewer availability.
What causes an LCA result to differ from a supplier’s declared footprint for the same material?
System boundary mismatch is the most common cause. If your LCA includes upstream forestry and pulp processing but the supplier’s EPD is declared gate-to-gate from the mill, the figures aren’t comparable. Always confirm the declared boundary scope (cradle-to-gate, gate-to-gate, cradle-to-grave) before drawing any comparison. A difference of 0.2–0.4 kg CO₂e per kg of board is entirely explainable by boundary definition alone.
How accurate are screening LCAs using Ecoinvent background data?
Directionally useful, but not declaration-ready. Ecoinvent v3.9 datasets for paperboard and corrugated are based on European average production. If your actual substrate comes from a mill with renewable energy supply, the Ecoinvent proxy may overstate emissions by 20–35%. For material selection decisions, this uncertainty is acceptable. For a public carbon label, it isn’t.
Can we use one LCA to cover multiple SKUs in the same packaging range?
It depends on how much the SKUs diverge in substrate weight and structure. If all SKUs use the same board grade and differ only in panel size within ±15%, a scaled functional unit approach is defensible. If some SKUs use a different board grade, coating, or add a foil lamination, each requires separate modelling — the emission intensity difference between a standard coated board and a foil-laminated variant can exceed 0.15 kg CO₂e per unit.
What is the minimum data quality required for a retailer sustainability programme submission?
Most major retailer programmes (Walmart Project Gigaton, Tesco’s supplier requirements, and similar) now require either a third-party verified product carbon footprint or an EPD from an accredited programme operator. Internal calculations without external verification are increasingly being rejected. Check the specific programme’s methodology requirements — several now reference GHG Protocol Product Standard as the baseline.
Is there a meaningful carbon footprint difference between FSC-certified and non-certified board of the same grade?
The physical emission profile of the board at the mill gate is generally within 3–7% between certified and non-certified production of the same grade from the same mill, because FSC certification governs chain-of-custody and forest management, not energy use in manufacturing. The carbon accounting difference comes from how you treat the biogenic carbon in the fibre — whether you count sequestered carbon as a credit depends on the LCA methodology you’re using, and this is an active debate in the field with no settled consensus.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.
