Comparative LCA and critical review: when is it mandatory?

ISO 14044 has a specific term: comparative assertion. It is a claim about the environmental superiority or equivalence of one product against another performing the same function, disclosed to the public. Any study meeting that definition carries an additional set of obligations.

The panel requirement and why it exists

For comparative assertions disclosed to the public, ISO 14044 requires critical review by an independent panel of at least three experts. A single verifier is not sufficient. The reasoning is that comparative assertions have commercial consequences: a claim that directly affects a competitor's market position demands a higher evidential bar.

In practice that raises cost and timeline markedly. Many companies therefore run the comparison as internal decision support and do not disclose it — entirely legitimate, and it does not trigger the panel requirement. What triggers the requirement is public disclosure, not the comparison itself.

Functional equivalence

For a comparison to hold, the two products must be shown to deliver the same function, at the same performance level, for the same duration. This is the most frequently violated condition. Comparing two insulation boards of different thermal resistance per square metre, or two coatings of different service life per unit, is technically invalid.

• The functional unit must include the performance level — not 'one m² of wall' but 'one m² of wall achieving R=4 m²K/W'
• Reference service lives must be equalised or normalised on an annual basis
• The same background database and the same version must be used for both products
• The same LCIA method and characterisation factors must be applied
• System boundary, allocation and cut-off rules must be identical on both sides

The same-data-source rule

Using primary installation data for your own product and a sector average for the competitor introduces systematic bias. Primary data is almost always better than an average because it reflects actual performance and carries no conservative assumptions. That asymmetry invalidates the comparison.

The correct approach models both sides at the same data quality level. If you cannot obtain the competitor's primary data, you must model your own product on sector data as well — or not publish the comparison.