The Empty Nature Attribute: Putting Land Condition on the Product

A farmer who has spent fifteen years rebuilding the water cycle on a piece of ground has almost no way of showing it to the person buying their lamb.

They can describe what they did. They can point at a scheme membership, which attests to a process they followed rather than to the state of the land now. They can show photographs. What they cannot do is hand over an independent read of the land itself and say: check it yourself, here is how.

That gap is about to become expensive, because the machinery for carrying exactly that kind of information is being built into products right now, and nobody has anything to put in it.

The passport is arriving with an empty field

The European Union has legislated Digital Product Passports into existence. Under the Ecodesign for Sustainable Products Regulation, a growing list of product categories will have to carry a machine-readable record of what they are and where they came from. Resolvable product identifiers, the plumbing that lets a scan on a shelf reach a structured record, are already ordinary retail infrastructure.

So the container is solved. The question is what goes in the part of it that is supposed to describe environmental performance.

For carbon, there is at least an established, contested answer. For nature, there is very nearly nothing. The field exists on the form and the honest options for filling it are a practice description, a certification that attests to process, or silence.

This is the same hole, one layer further down the supply chain, that we have written about in nature disclosure. TNFD names ecosystem condition as a core indicator of the state of nature and then declines to define the metric. ESRS E4 and GRI 101 inherit the silence. At corporate level that leaves a box on a report. At product level it leaves a field in a passport. Same absence, different container.

Why nobody fills it

Three reasons, and they compound.

Measuring land condition has been slow and local. Soil sampling, botanical survey and field walks produce good data about a small area at one moment, at a cost that does not scale to every field of every supplier. You cannot build a product attribute on a method that takes a week per farm.

What has been measurable has mostly been lagging. Soil organic carbon moves slowly and noisily. By the time it registers a change, the management that caused it is years in the past. An attribute that reports what the land was like several years ago is not much of an attribute.

And the claims risk is severe. In the UK the Digital Markets, Competition and Consumers Act now backs direct enforcement against misleading environmental claims. The CMA’s Green Claims Code is explicit that a claim must be substantiated and must not overstate. A producer who prints an unsupported nature claim on a package is exposed in a way they were not five years ago. Faced with that, the rational move is to say nothing, which is exactly what most do.

Any attribute worth having has to survive all three: cheap enough to run on every field, fast enough to reflect current management, and defensible enough that a regulator reading it finds it fairly stated.

What we built instead

A Land Health Credential is a statement of how a piece of land is functioning, issued to the person who stewards it and built to be handed on.

It carries, for a named property and a named year: the Land Health Score on the 1-to-9 scale from Collapsed to Flourishing, measured against what is possible in that land’s own ecoregion rather than a universal ideal; the four ecosystem process indices covering the water cycle, how sunlight becomes energy, how minerals are made available, and the relationships between living things; the trajectory across the monitored record; and carbon flux and stocks.

Because it comes from satellite observation and climate data rather than site visits, it runs on every field of every property at the same cost, and it reports the current year rather than a survey from three seasons ago. Because it is scored against local reference conditions, a hill farm is judged against what that hill could be.

But the measurement is the ordinary part. The part that makes it usable as a product attribute is that a stranger can check it.

Two questions, two mechanisms

A credential has to answer two quite different challenges, and most systems only handle the first.

Is this the document it claims to be? Every credential carries a fingerprint of its own contents and a signature made with a key only we hold. Alter one figure, one date or one property name and the check fails. Anyone can run that check in their own browser, and the file never leaves their device, because we hold no public register of our clients’ numbers.

Does the issuer still stand behind it? This is the question a signature cannot answer, and the one most schemes quietly ignore. A sealed document stays sealed forever, including long after its figures have been revised or the relationship has ended. So every credential also carries a short code that resolves against our register to one of exactly four answers: current, superseded, withdrawn, or not ours. Reissue a property’s assessment and the old code begins saying “there is a newer version” from that moment, everywhere it has already been printed, including on packaging that shipped last year.

That second mechanism is what turns a document into an attribute you can safely attach to a product. It means a claim can be withdrawn.

The landowner decides what is visible

None of this is worth anything if it forces disclosure on people who did not choose it.

Publication is off by default. A property appears in the register only when its owner opts in, and when they do they choose the depth: confirmation only, which names the property and says the credential is genuine and current and nothing further, or full disclosure, which publishes the figures and the signed credential itself. A record set to confirmation only physically does not contain the figures, so there is no setting anywhere that could accidentally reveal them, and it stays out of search results.

Wilder Wood Farm, our public demonstration property and the only one whose data appears publicly, is set to full disclosure. Its credential is live and anyone can resolve it: ecointel.io/v/wilder-wood-farm.

The reference you print is not the reference you check

There is a trap here that is worth naming, because getting it wrong is how a good scheme quietly breaks.

The code on a badge belongs to one assessment. Reassess the property and it changes, which is right for a badge: the badge shows a score for a particular year, so it is out of date anyway. But if that same code has been printed on packaging, the packaging now points at a superseded credential, and eventually at nothing anyone wants to read.

So a property that publishes in full also gets a stable reference: a readable address like ecointel.io/v/wilder-wood-farm that always resolves to its current credential. That is what belongs on a product, in an SKU record or on a supplier page. Reassessment changes what it points to rather than invalidating it, so a label printed today still resolves correctly in three years, and resolves to better information than it did when it was printed.

Stable references go only to properties publishing in full. A guessable name on a confirmation-only property would reveal that it holds a credential at all, which is exactly what its owner chose not to publish. There the opaque code is the protection, and it stays the only way in.

Built for the readers who are coming

Follow that link as a person and you get a page that says whether the credential is genuine, whose land it describes, and what it attests.

Request the same address as software and you get the structured record, with the complete signed credential embedded so it can be re-verified independently, without asking anyone’s permission and without the holder sending anything. It is a stable address, it is open to cross-origin requests, and it carries the standard vocabulary a machine reader expects.

That matters more each month. Increasingly the first reader of a product page is not a shopper but an agent working on their behalf: comparing suppliers, filling a procurement questionnaire, assembling a scope-3 estimate, checking whether a claim on a website is supported by anything. Those readers do not read marketing copy. They follow identifiers to structured records and check whether the numbers hold up.

A product line can carry its credential code the way it carries a batch number. An SKU record can hold it as an attribute. A supplier questionnaire can cite it. In every case the receiving party is one request away from the underlying evidence, and one request away from finding out that the credential was withdrawn.

What it deliberately does not say

A Land Health Credential describes the condition and trajectory of the land. It does not attribute that condition to any particular practice.

The distinction sounds pedantic and is the whole game. “The land behind this product scores 6, Recovering, and has improved over eight years of monitoring” is a statement about a place that we can evidence and a stranger can re-check. “This product improves nature” is a causal claim about the effect of a practice, and it needs a counterfactual, a control and an attribution method that satellite condition monitoring does not provide. The first is defensible. The second is the kind of sentence that attracts enforcement.

Nor is it species data. The community-dynamics index is a functional proxy inferred from how the vegetation behaves through the seasons, not a survey of what lives there. It is a genuine signal about ecological function and it is not a biodiversity metric, and we will not let it be read as one. Where a claim needs species evidence, it needs a species method, and eDNA and condition monitoring are complementary rather than competing.

And it is not assurance. A credential is indicative decision-support and the verification-ready evidence base an assurance process would draw on. “Certified” and “verified” stay reserved for when such a process has actually happened.

Those three restraints are not hedging. They are what makes the remaining claim worth printing, because a buyer’s lawyer can read it without flinching.

What this looks like for a producer

For a farm or grower, the credential is the answer to a buyer’s sustainability questionnaire that does not require re-describing your practices from scratch every time, and the thing you can put next to a product to say something specific about the land it came from.

For a timber or forestry business, where the land and the product are unusually tightly coupled and rotations are long, a dated series of credentials is a record of stewardship over a period long enough to matter, in a form a purchaser can check rather than take on trust.

For a buyer or a brand, it is a supplier-level nature attribute that arrives with its own evidence and its own withdrawal mechanism, which is a materially different risk position from a claim that rests on a supplier’s own description.

None of this requires the passport regulation to land first. The credential is useful the moment a buyer asks a question you would otherwise answer with an adjective.

The honest position

We are not claiming to have invented product transparency, and the identifier plumbing is not ours. What has been missing is not the container. It is a land-condition figure with an evidence base a stranger can interrogate, a currency mechanism that lets it be withdrawn, and enough restraint in what it claims to survive contact with a regulator.

That is what a Land Health Credential is. The infrastructure for carrying it is already being built into the products around you, and the nature field in it is empty.