Where Analysis Ends and Judgment Begins: The LCA Responsibility Framework

In short: There is a specific failure mode in sustainability decision-making that no amount of methodological rigour prevents — a study whose calculations are entirely correct being used to justify a decision it was never designed to inform. A 2026 framework from the EcoResonance Institute maps exactly where that happens.

The structural risk hiding inside good analysis

Life Cycle Assessment provides a structured way to quantify and compare environmental pressures. What it does not do — and was never designed to do — is determine decisions. Interpretation, prioritisation, and the acceptance of trade-offs remain the responsibility of institutions and decision-makers.

In practice, that boundary blurs constantly. LCA results get treated as objective verdicts. Rankings and aggregated scores justify decisions while obscuring the trade-offs underneath. Responsibility quietly transfers from the executive who signed off to the model that produced the number.

Ekaterina Markelova's LCA Responsibility Framework (Zenodo, 2026, doi:10.5281/zenodo.18529664) names the consequence precisely: even technically correct LCAs can legitimise harm when values and non-negotiables remain implicit.

This is not an argument against LCA. It is an argument about governance — and it becomes materially relevant the moment LCA results start feeding CSRD disclosures and investment justifications.

What ISO deliberately leaves out

ISO 14040 and ISO 14044 provide a robust methodological basis for conducting LCA. They assume values are present in goal definition, system boundaries, and interpretation — but they do not require those values to be stated explicitly.

The result is that value choices remain fragmented across technical decisions, or embedded in later steps like weighting and aggregation where nobody examines them as value choices at all.

The framework's single structural addition is a pre-analytical values articulation step, placed before goal and scope definition. It asks the decision-maker to state what must not be traded away, and which life-supporting systems are considered essential to sustain or strengthen — before modelling choices narrow the space of possible outcomes.

Critically, values are treated as external to the model. They guide interpretation; they are not calculated by LCA. The framework does not modify ISO calculation steps and introduces no new impact indicators. Its contribution is governance, not methodology.

Two types of values

The framework distinguishes:

Protective values — what must not be traded away, even if optimising a particular indicator would improve. Avoidance of irreversible biodiversity loss. Prevention of critical water system degradation. Protection of human health. These are framed as normative commitments, not compliance criteria.

Regenerative values — what the organisation seeks to strengthen over time. Restoring watershed function. Improving soil fertility. Strengthening local human capabilities. These provide strategic direction particularly where impact reduction reaches technical or practical limits.

Three domains of responsibility

The framework separates who is accountable for what:

Science (S) — defines cause–effect relationships and impact pathways
Technicians and LCA modellers (M) — implement models, data, and calculations
Decision-makers (D) — set purpose, values, priorities, and accept responsibility for trade-offs and consequences

Across twenty levels, the framework maps how responsibility shifts upward as an assessment progresses. Of those twenty levels, the majority of decision-sensitive ones sit with decision-makers — including allocation rules, system model choice, weighting, ranking, aggregation, and final judgment.

The core principle: as the assessment progresses from framing to calculation to interpretation, the capacity of science to decide decreases while the responsibility of decision-makers increases.

Uncertainty changes type, not just magnitude

One of the framework's most useful insights for executives is that treating all uncertainty as a statistical problem is a category error. Uncertainty shifts in nature across the twenty levels:

Levels 1–7: structural and ethical. What is included, excluded, prioritised, or treated as negotiable. These concern responsibility allocation, not measurement error.
Levels 8–15: technical and scientific. Data representativeness, modelling choices, parameter values, cause-effect representations.
Levels 16–20: normative and governance-related. Interpretation, prioritisation, aggregation, and acceptance of trade-offs in context.

And uncertainty accumulates. Early framing choices define the space of possible outcomes and cannot be corrected by later numerical precision. As the framework states plainly: increasing data quality or refining characterisation factors cannot repair a misframed question.

This explains something practitioners observe constantly — highly detailed LCAs producing misleading guidance because early assumptions were misaligned with the decision context. It also explains why two LCAs of similar systems can reach different conclusions without either being methodologically incorrect.

Five misuse patterns worth recognising

The framework identifies recurring misinterpretations that arise not from calculation errors but from unclear boundaries between analysis and judgment:

Treating numerical results or rankings as objective decisions rather than conditional outputs dependent on assumptions, boundaries, and value choices
Focusing on aggregated scores while obscuring trade-offs across impact categories
Using studies designed for internal learning or hotspot identification as definitive comparative claims for external communication or policy justification
Interpreting unweighted or single-score results as value-neutral, despite implicit prioritisation embedded in indicator selection and system modelling
Shifting responsibility for controversial decisions to the model, database, or method

Pattern 3 deserves particular attention. The framework calls it a category error: a well-executed LCA designed for internal eco-design being deployed as a public comparative claim. The calculations remain sound. The use is illegitimate.

Where rankings and single scores become dangerous

The framework is direct about weighting, ranking, and aggregation: these are not neutral techniques. They are value-laden interpretative tools that compress information and reduce visibility of trade-offs.

At the weighting level, once results are weighted, trade-offs disappear and alternatives collapse into a single score. The framework identifies three risks: false universality (a particular value system presented as neutral), premature closure of debate, and responsibility outsourcing — the "the score decided" problem.

At the ranking level, declaring a winner when differences fall within uncertainty creates artificial certainty. Decision-makers lock in a choice because it ranks first, even when small assumption changes would reverse the order.

At the aggregation level, trade-offs become invisible and governance becomes fragile. The framework flags this as especially risky where irreversible impacts exist — biodiversity loss, toxicity persistence — because a composite index can show improvement while an unacceptable outcome proceeds underneath it.

Five questions for anyone receiving an LCA

The framework's most immediately usable output is a structured reading protocol. Decision-makers should begin with interpretation rather than results, asking:

Levels 1–2: What was the study designed to decide, and does that purpose match the decision currently under consideration? A mismatch here signals high misuse risk even when calculations are technically sound.

Levels 3–7: Which responsibilities were accepted, excluded, or implicitly shifted through choices of functional unit, impact categories, and system boundaries?

Levels 8–11: What system representation is being described, and does it correspond to the real-world change being considered? This includes the attributional versus consequential modelling choice — whether results describe the system as it is, or the consequences of change.

Levels 12–15: Which parts of the results are driven by scientific modelling and parameter choices, and where does interpretation or methodological convention begin to influence outcomes?

Levels 16–20: Where must the decision-maker, personally or institutionally, accept ownership of trade-offs? The framework is unambiguous here — if this question cannot be answered explicitly, the LCA is not yet being used responsibly.

Minimum practices for governance

The framework recommends that any LCA presented for decision support include explicit disclosure of Levels 1–7: values, goals, functional units, and system boundaries. System model choice should be explicitly justified against the decision context. Any use of normalisation, weighting, ranking, or aggregation should carry an explicit warning that these steps involve value-based judgment.

Weighted or aggregated results should be treated as decision aids rather than verdicts — supporting discussion, not replacing deliberation.

And critically: decision-makers should document, outside the LCA, why certain trade-offs were accepted, why others were rejected, and which considerations were treated as non-negotiable. This documentation is a governance function and cannot be delegated to the assessment.

How this connects to CSRD and existing frameworks

The framework is designed to be interoperable with ISO 14040/44, CSRD, EU Taxonomy–adjacent use, GHG Protocol, SBTi, TCFD, and TNFD — without defining compliance, thresholds, or eligibility. Its role is explicitly meta-level.

Under CSRD, organisations must disclose impacts, risks, and opportunities alongside governance processes and strategic responses. LCA increasingly informs those disclosures. The framework clarifies how LCA results relate to governance decisions by making explicit where values, priorities, and trade-offs enter — rather than allowing them to remain implicit inside technical choices or aggregated indicators.

For TCFD and TNFD, it helps link analytical assessments to governance by making explicit which risks are treated as material, which systems are prioritised, and where judgment rather than modelling determines the response.

Planetary boundaries, in this architecture, are treated as external value reference points rather than measurable endpoints — orientation for interpretation, not something LCA quantifies directly.

The reframe that matters

Perhaps the framework's most useful move is reframing residual impacts. Not as acceptable externalities or model limitations, but as responsibility signals — inputs to strategic discussion about risk management and potential restorative action, without implying such actions are certified by the assessment.

That reframe converts LCA from a defensive justification tool into a decision-support instrument. Which, as the framework argues, is what it was always meant to be: a decision-support instrument, not a decision substitute.

Frequently asked questions

What is the LCA Responsibility Framework? A governance architecture published by Ekaterina Markelova (EcoResonance Institute, 2026) that maps twenty levels of an LCA study, clarifying where scientific analysis ends and decision-making begins. It adds a pre-analytical values articulation step that ISO 14040/44 deliberately leaves implicit.

Does this framework replace ISO 14040 and 14044? No. It does not modify ISO calculation steps and introduces no new impact indicators. It builds on ISO-compliant practice, adding governance and interpretation rather than methodology.

What are protective and regenerative values? Protective values define what must not be traded away even if optimising a particular indicator would improve — such as avoiding irreversible biodiversity loss. Regenerative values define what systems the organisation seeks to strengthen over time, such as restoring watershed function. Both are treated as external to the model.

Can a technically correct LCA still cause harm? Yes. The framework's central warning is that even technically correct LCAs can legitimise harm when values and non-negotiables remain implicit — for instance when a study designed for internal eco-design is deployed as a public comparative claim.

Are weighting and aggregation neutral technical steps? No. The framework treats weighting, ranking and aggregation as value-laden interpretative tools that compress information and reduce visibility of trade-offs. Once results are weighted, alternatives collapse into a single score and trade-offs disappear.

How does this relate to CSRD reporting? Under CSRD, organisations must disclose impacts, risks and opportunities alongside governance processes. The framework clarifies how LCA results relate to those governance decisions by making explicit where values and trade-offs enter, rather than leaving them implicit inside technical choices.

How Sustek.co builds the governance layer

Most organisations have more environmental data than governance around it. The result is exactly what this framework warns against: technically sound analysis producing decisions nobody has explicitly owned. Sustek.co closes that gap in three stages:

Sustainability Pulse — Establishes the ESG baseline and identifies where data quality currently limits the reliability of any conclusion. Deploys the Sustrategize™ baseline diagnostic: maturity assessment, gap analysis, and circular economy potential mapping.

Sustainability Navigator — Structures the decision architecture alongside the technical roadmap. Double materiality assessment with values made explicit, circular economy value-at-stake in figures the CFO recognises, 4IR technology sequencing, and a board-ready Transformation Blueprint aligned to CSRD, GRI and TCFD.

Sustainability Command — Runs the governance continuously rather than annually. Quarterly Executive Impact Dashboards, investor-grade reporting, and ongoing SROI measurement through ValueFlow™, our social value quantification platform powered by MeasureUp.

Where a specific judgment requires specialist expertise — sector-specific LCA modelling, independent critical review, jurisdictional regulatory interpretation — Iconet® activates a network of 93+ vetted sustainability consultants on demand, without fixed structural cost. The Iconet® x AWS reStart program carries a verified SROI of 6.41:1, calculated using the same auditable methodology we apply for clients.

Sustek.co is an S.A.S. BIC certified under Sistema B Colombia, and part of Google for Startups and AWS Activate.

Can your organisation answer, right now, who owns the trade-offs in its last sustainability decision? If not, start with the diagnostic: sustek.co/services

Source: Markelova, E. (2026). "The LCA Responsibility Framework." EcoResonance Institute. Zenodo, doi:10.5281/zenodo.18529664. Framework references ISO 14040/14044, CSRD, EU Taxonomy, GHG Protocol, SBTi, TCFD, and TNFD.