LCA Guide: How a Life Cycle Assessment Is Done

What is an LCA and what is it for?

A life cycle assessment compiles the inputs and outputs of a product system (raw materials, energy, emissions, waste) across its life cycle and evaluates their environmental impacts. The framework is set in ISO 14040 (principles) and the rules in ISO 14044 (requirements and guidelines).

In practice an LCA is used in three places: as the calculation behind an EPD, as the basis of an ISO 14067 product carbon footprint, and in product development to answer "which material or process drives the impact?". All three can be derived from one model; setting up that relationship at the start is how you avoid collecting the same data twice.

The four phases: goal and scope, inventory, impact assessment, interpretation

ISO 14040 defines LCA in four phases and states explicitly that they are iterative: a finding in interpretation can reopen the scope or the inventory. A good study plans for that loop; a poor one declares itself "finished" after the first pass.

1. 1Goal and scope definition — Why is the study being done, for whom, and which decision will it support? The functional (or declared) unit, system boundary, cut-off criteria, allocation approach, impact categories, data-quality requirements and limitations are put in writing here.
2. 2Inventory analysis (LCI) — Collecting inputs and outputs for every unit process: raw and auxiliary materials, energy, water, transport, waste and emissions. Primary (plant) data and secondary (database) data are separated and documented separately.
3. 3Impact assessment (LCIA) — Converting inventory flows into impact categories (climate change, acidification, eutrophication, resource depletion and so on) with characterisation factors. Which method and version is used is part of the result.
4. 4Interpretation — Identifying the significant issues (hotspots), the completeness, sensitivity and consistency checks, conclusions and limitations. The report is the product of this phase — a reasoned assessment, not a results table.

Goal and scope: functional unit, system boundary, cut-off criteria

Three decisions made in the scope phase determine the result. The functional unit makes the function the product provides measurable (e.g. "insulating 1 m² of wall to a given thermal transmittance for 50 years"). EPDs under EN 15804 mostly use a declared unit instead (e.g. 1 tonne of steel, 1 m³ of concrete) — a quantity reference, not a function. Confusing the two is the most common reason two EPDs get compared wrongly.

The system boundary states which processes are in the model: cradle to gate (A1–A3), cradle to grave (A–C) or including module D. EN 15804+A2 defines the module structure for construction products and which modules are mandatory; the scope widened compared with +A1, which is why studies planned on an "A1–A3 only" assumption come back from verification.

Cut-off criteria say which minor flows may be neglected. EN 15804+A2 limits this to 1% of mass and of primary energy per unit process and a total of at most 5% per module, and the neglected flow must also be shown not to be environmentally significant. "We left it out because it was small" is not, on its own, a sufficient justification in verification.

Inventory: primary data, secondary data and traceability

Two kinds of data meet in the inventory phase. Primary data comes from the plant: production volumes, raw-material recipes, electricity and fuel consumption, water, waste, transport distances. Secondary data comes from background databases: production of the raw materials, the electricity grid mix, transport vehicles, waste treatment. Which flow is represented by which kind of data must be stated clearly in the report.

The reference year and the production volume are the common denominator of all plant data: a year's consumption is divided by the same year's production. Mixing years, estimates with no meter or invoice behind them, and energy that cannot be split by product in multi-product plants are the three issues that cost the most time in verification.

• Every figure carries its source: meter, invoice, stock record, production report. The first question in verification is "where did this come from?".
• Build a mass balance: raw material in should roughly equal product plus waste plus losses out. If it does not, a flow is missing or a unit is wrong.
• Check the energy balance separately: electricity and fuels should add up to the plant's known consumption. Calorific-value assumptions are documented.
• Write background data with its version: database name, version, geography and reference year of the dataset. "European average" is not a source on its own.
• Leave missing data missing: if an estimate is used, state that it is an estimate, what it rests on and what it affects. A gap filled silently is found in verification.

Allocation: sharing burdens in multi-product processes

When a process yields more than one product (main product plus by-product, or different products from the same line), inputs and emissions have to be shared between them. ISO 14044 defines an order for this, and the order cannot be skipped.

1. 11 · Avoid allocation — Where possible, subdivide the process so that each product gets its own inputs; or expand the system boundary to include the function the by-product displaces.
2. 22 · Allocate by physical relationship — If it cannot be avoided, share inputs and outputs in a way that reflects the physical relationship between the products (mass, energy content and so on). The choice of physical quantity is justified.
3. 33 · Allocate by other relationships — If no physical relationship can be established, another relationship such as economic value may be used. The period the prices refer to and the effect of price fluctuation are shown through sensitivity analysis.

Impact assessment (LCIA): categories and characterisation

In the LCIA phase the hundreds of flows in the inventory are reduced to a small number of impact categories with characterisation factors. Which method and version is used is an inseparable part of the result: the same inventory gives a different number with a different characterisation set.

For construction product EPDs, EN 15804+A2 defines the mandatory core indicators and their characterisation method (the set based on the JRC's Environmental Footprint). Global warming potential is not one number but four separate indicators: total, fossil, biogenic, and land use and land use change. This is one of the main reasons the GWP figures of +A1 and +A2 EPDs cannot be compared directly.

The indicator list is a summary; the full list and units are in EN 15804+A2 and the relevant PCR.

Interpretation: hotspots, sensitivity and limitations

Interpretation is the phase that makes the result readable. ISO 14044 asks for three checks: completeness (are all significant flows and data included?), sensitivity (how sensitive is the result to which assumptions and data?) and consistency (are the method, data quality and assumptions the same throughout the study?).

In practice the most valuable output is the hotspot analysis: which material, process or module the total impact comes from. It is the only thing the manufacturer can act on. Sensitivity analysis shows which assumption moves the result — the electricity mix, transport distance, allocation ratio and end-of-life scenario come up most often.

• The limitations section is mandatory and must be honest: flows that could not be represented, where proxy data was used, how representative the reference year is.
• If a comparative assertion is made (one product is "better" than another), the additional requirements of ISO 14044 apply; if it is to be disclosed to the public, a critical review panel is required.
• Focusing on one impact category (usually GWP) and ignoring the others hides burden shifting; the report shows all categories, the emphasis is made in the interpretation.

Data quality: the ISO 14044 requirements

ISO 14044 requires data-quality requirements to be defined in the scope phase and reported. They are not a checklist but the frame that shows the reader how reliable the study is. In an EPD verification, the verifier looks at exactly these headings.

Critical review and verification: ISO 14044 versus ISO 14025

There are two separate control mechanisms and they are often confused. The ISO 14044 critical review assesses whether an LCA study conforms to the standard as a method; for comparative assertions disclosed to the public it must be carried out by an independent panel. EPD verification, under ISO 14025, confirms that the declaration conforms to the PCR and EN 15804 and that the data is traceable; it runs under the programme operator's rules.

Neither says the product is "environmentally friendly". Review and verification confirm that the study follows the rules and that the figures are traceable; the evaluation is made by the reader.

Choosing software and a database

LCA software and the background database influence the result but do not determine it; the scope and the plant data do. The criteria for the choice: the databases accepted by the target EPD programme, support for the EN 15804+A2 characterisation set, the size of the product portfolio (one product or a parametric family), the capacity of the in-house team and the licence model. This page recommends no brand; to see which tool approach fits your portfolio, use the free selector.

Through a verifier's lens

In the LCA files that arrive for verification, the error is rarely in the arithmetic; the software adds up correctly. The problems accumulate in three places: scope decisions that do not match the PCR, plant data that cannot be traced back to its source, and assumptions that do not appear in the report.

The practical test of a good study is this: can the reader of the report find, in the same report, the answer to "where did this number come from and why this value?" for every figure? If yes, verification closes in one round; if not, every round is a new list of questions.

Common mistakes

• Starting data collection before the scope is closed. If the declared unit or system boundary changes later, part of the data collected is wasted.
• Confusing the functional unit with the declared unit. Comparing two EPDs with different units is the most common reading error.
• Mixing data from different years. Consumption and production must belong to the same reference year.
• Skipping the allocation hierarchy. Going straight to economic allocation is not accepted without showing why avoidance and physical relationship could not be applied.
• Writing background datasets without a version. Without the database name, version and dataset geography the result cannot be reproduced.
• Not stating the characterisation method. The +A1 and +A2 sets give different results; which one was used must be written.
• Filling gaps silently. A value not marked as an estimate is the finding that costs the most time in verification.
• Looking only at GWP. Burden shifting in the other categories stays invisible.

Frequently asked questions

Sources and basis

This page does not replace the standard texts; it explains the structure of ISO 14040/44 and EN 15804+A2 and verification practice. When planning a study, always work from the current text of the standard and the target programme operator's PCR.

• ISO 14040 — Environmental management, life cycle assessment: principles and framework ↗
• ISO 14044 — Environmental management, life cycle assessment: requirements and guidelines ↗
• ISO 14025 — Type III environmental declarations ↗
• ISO 14067 — Carbon footprint of products ↗
• EN 15804:2012+A2:2019 — Core rules for EPDs of construction products (CEN)

Last reviewed: 11 October 2026