LCA Methodologies in Practice: A Guide to Sustainable Product Development

Choosing a life cycle assessment (LCA) methodology is a decision about which standard the result has to survive. ISO 14040 and 14044 set the frame, the EU Product Environmental Footprint fixes the rules per product category, PACT governs how carbon data moves between companies, and the GHG Protocol Product Standard sits between them. This page routes a use case to the standard that fits it.

What an LCA methodology fixes: ISO 14040/14044 and the four phases

A methodology does not tell you what a product emits. It fixes the choices that decide what the number means. ISO 14040 sets the principles and framework, ISO 14044 the requirements a study has to meet, and together they define four phases.

Goal and scope definition names the question, the functional unit and the system boundary. Life cycle inventory (LCI) collects the flows across that boundary. Life cycle impact assessment (LCIA) converts them into impact categories such as climate change, acidification or water use, using the characterisation factors of a named method: the EU Environmental Footprint set (EF 3.1), ReCiPe or CML, with normalisation an optional step on top. Interpretation tests whether the result carries the conclusion. A step-by-step walkthrough covers each phase.

Three of those choices move the result more than any data source. The functional unit decides what is compared: one litre of paint and one square metre of coated surface can rank the same two products differently. The system boundary decides how much is counted, cradle-to-gate stopping at the factory gate, cradle-to-grave including use and end of life. Attributional LCA allocates the impacts a product carries today, consequential LCA models what changes if one more unit is produced. A supplier asking for a cradle-to-gate attributional figure and a design team asking what a material switch would change are not asking the same question.

Which standard applies: ISO, PEF, PACT and the GHG Protocol Product Standard

ISO is the frame the others build on. What separates them in practice: what gets measured, how prescriptive the rules are, and whether the figure is built to be handed to someone else.

The routing rule I use is short: the standard follows the audience, not the product. An internal design decision needs the ISO frame and a documented dataset choice, nothing more. A customer data request follows whatever the customer names, increasingly PACT in manufacturing supply chains. A public claim has to survive third-party review, so use PEF where a PEFCR exists; in construction and a growing set of other categories the reviewed output a buyer expects is an environmental product declaration (EPD). For EU disclosure, keep the product study consistent with the corporate inventory, because that is where the two get compared.

Using LCA results in product and investment decisions

An LCA earns its cost when it changes a decision before the design is frozen. A footprint that lands mostly in the use phase points at efficiency in the product, not at the supply chain. One that concentrates in a single purchased material points at substitution or a supplier switch, and shows how much of the total is addressable at all.

The same table gets read a second way in investment. I work as a venture partner as well as an adviser, and in a committee an LCA is not an environmental document. It is evidence for a DNSH threshold, for how exposed a portfolio company is to a customer requirement that has not arrived yet, and for whether a redesign belongs in the operating plan or the capital request. That reading decides whether a footprint gets funded.

From footprint to disclosure: where LCA results land in EU reporting

A product footprint rarely stays in the product file. Under the ESRS it feeds the Scope 3 categories of the corporate inventory, and product studies with inconsistent boundaries produce an inventory that will not reconcile. For EU Taxonomy alignment the same data supports substantial contribution and do-no-significant-harm evidence at activity level. For consumer-facing claims, the empowering consumers directive (EU) 2024/825 applies from 27 September 2026 and bars unsubstantiated generic environmental claims, while the separate Green Claims Directive has stalled. A footprint that was never built to be reviewed is the wrong input for all three.

Where LCA results break in practice

Studies rarely fail on the model. They fail on three things.

Secondary datasets. A background database chosen for coverage rather than regional fit puts most of the footprint into numbers nobody in the company can defend. The share of primary data behind the hotspots is what decides how far the figure carries. When a customer asks where the electricity mix came from, the tool default is not an answer.

Allocation. Multi-output processes and recycled content are where two defensible methods produce results far enough apart to change the conclusion. If the choice is not written into goal and scope, the software makes it silently.

Functional unit drift. The unit set in phase one becomes a different unit by the time the result is presented, usually because a marketing comparison needs a different basis than the engineering one.

Before I accept a supplier's LCA figure, five things have to be visible: the functional unit as the study words it, the system boundary the study actually stops at, the vintage of the background data, the allocation rule written into goal and scope rather than left to the software, and the review status. A figure missing any of them is usable internally and nowhere else.

I have built ESG software as well as used it, which is why I read a tool output as a set of embedded assumptions rather than a result. Most of what fails review was decided by a default nobody changed.

Questions about LCA methodologies