How to Measure Circularity: Methods Compared

Circularity has no equivalent of the tonne of carbon dioxide equivalent. Carbon reporting converged on one unit, which made comparison possible and gaming harder. Circularity has half a dozen metrics measuring genuinely different things, and choosing between them is a decision with consequences.

The useful way to compare them is not by which is best but by what each cannot see.

The five common metrics

The blind spots are not defects to be corrected. Each metric is measuring a specific thing correctly, and the error is treating any one of them as a summary of circularity.

Recycled content: the most reported, the most limited

It is popular because it is comparatively easy to evidence, appears in regulation with numerical thresholds, and is intuitive to a consumer. It measures only what went in.

A product with high recycled content that lasts two years and is unrecyclable scores well on the metric and poorly on any sensible view of circularity. Reported alone, it rewards buying recyclate rather than designing well.

Recyclability: a claim about a possibility

Recyclability rate describes what could be recovered under defined conditions. Those conditions include a collection system that reaches the product and a treatment route that accepts it, neither of which the metric verifies.

A product that is technically ninety percent recyclable and is discarded in general waste is recycled at zero percent. That gap between theoretical and actual is where most circularity reporting quietly lives — the underlying causes are covered in why recycling fails without product data.

The Material Circularity Indicator

MCI combines recycled input, recovered output and utility relative to an industry average into a single figure between zero and one. It is the most complete of the common metrics and its completeness is also its weakness.

A single number cannot tell a reader whether a mid-range score reflects good inputs and poor recovery or the reverse. Those are different problems with different fixes, and the metric that combines them hides which one you have — the calculation is set out in the Material Circularity Indicator.

Lifetime extension: the biggest effect, the hardest evidence

Doubling a product’s service life halves the material and manufacturing impact per year of use, which usually exceeds anything achievable through recycled content or recyclability. It is the most consequential metric and the least reported.

The reason is evidence. Recycled content can be evidenced at manufacture; lifetime can only be evidenced by observing products in service over years, which requires per-item records that most manufacturers do not maintain.

Which should you report?

Two metrics that pull against each other, so that optimising one is visibly at the expense of the other. Recycled content paired with expected lifetime is the most useful combination for most product categories.

That pairing is hard to game. Raising recycled content by using lower-grade recyclate tends to reduce durability, so a product improving on both is genuinely improving rather than shifting an impact into an unmeasured place.

What regulation actually requires

Specific thresholds rather than a general circularity score. Regulation (EU) 2025/40 sets recycled content percentages for packaging, Regulation (EU) 2023/1542 sets them per battery metal, and the ESPR, Regulation (EU) 2024/1781, will set parameters per product group through delegated acts.

None of these asks for a composite index. Your reporting obligations will be per-parameter, and a composite metric is something you choose to publish for comparison rather than something you must produce.

Where collection rate fits

Collection rate is the metric manufacturers report least and the one that most determines whether any of the others matter. A product designed for recycling, made from recyclate, that is never collected delivers nothing measurable at end of life.

It is reported rarely because it is largely outside a manufacturer’s control, sitting with municipal systems and consumer behaviour. That is a fair objection and it does not make the number less relevant, because a strategy that depends on collection while ignoring collection rates is resting on an assumption nobody has checked.

Take-back schemes are the lever a manufacturer does control, and the reason they appear in circularity strategies far more often than they succeed commercially is that they attack precisely this constraint.

How do these metrics get gamed?

• Recycled content raised with post-industrial scrap, which is production waste rather than recovered material.
• Recyclability assessed under ideal conditions that no real facility operates.
• MCI utility factors benchmarked against a weak industry average.
• Lifetime claimed from design intent rather than observed field performance.
• Collection rate reported for the scheme rather than for the product category.

The first is the most widespread and the most defensible-sounding. Post-industrial scrap is genuinely recycled material and is far easier to obtain than post-consumer, which is why a recycled content figure should always state which kind it is.

What about carbon as a proxy?

Some organisations report product carbon footprint instead of a circularity metric, on the reasoning that circular interventions reduce emissions and carbon is a comparable unit. It is a defensible simplification and it misses two things.

Resource depletion is not carbon. Extracting a critical raw material with a low energy cost can be a serious circularity problem and a trivial carbon one, and a carbon-only view treats it as solved. Waste volume and toxicity are similarly invisible in a carbon figure.

Carbon is a good proxy for the energy dimension of circularity and a poor one for the material dimension, which is precisely the dimension the passport exists to address.

The measurement problem underneath all of them

Every metric here depends on material composition, and most depend on evidence about what happens after sale. Both are the same product data problem in different clothing.

A manufacturer who cannot state composition accurately cannot compute any of these metrics defensibly, whichever they choose. That is why data quality precedes metric selection rather than following it.

The practical consequence is an ordering that most programmes get backwards. Choosing a metric first produces a reporting requirement that the underlying data cannot support, and the gap is then closed with assumptions that make the published figure unverifiable.

Establishing composition, recycled content provenance and end-of-life routes first produces a dataset from which several metrics can be computed, and the choice of which to publish becomes a communication decision rather than a constraint imposed by what happened to be measurable.