Declaring carbon neutrality is easy. Demonstrating it, to a standard someone else will accept, is a documentation exercise — and that is the part most organisations discover late.
PAS 2060 is the specification that has been widely used for demonstrating carbon neutrality. What it really asks for is not a claim. It is a file.
Note the moving ground: PAS 2060 has been the reference point for some years, and carbon-neutrality standards are being revised internationally. Check the current status and what your certifier expects before committing to a route. The evidence discipline below is the part that stays true regardless of which standard you end up under.
The four things being asked of you
Reduced to essentials, a demonstration of carbon neutrality asks you to establish and evidence:
What is in scope. The subject of the claim — an organisation, a site, a product, an event — and its boundary. This sounds administrative and is the decision everything else rests on.
What it emits. A quantified footprint, calculated by a stated methodology, from data you can produce.
What you are doing about it. A carbon management plan with a reduction target, a timescale, and the means — not an intention.
How the remainder is dealt with. Offsetting the residual, with the offsets documented and retired, and evidence of that retirement.
Then it is repeated. Neutrality is not achieved once; it is maintained, re-evidenced each period, which is what makes the manual version so expensive.
Where it goes wrong
Almost never at the claim. Almost always at the file.
The boundary shifts. A site is added, a business unit sold, a lease changes. If the boundary was never recorded precisely, this year's footprint is not comparable to last year's and nobody can say by how much.
The methodology drifts. Someone new calculates it slightly differently. Both figures are defensible; the trend between them is not.
The evidence is scattered. The footprint is in a consultant's spreadsheet, the offsets in an email, the reduction plan in a slide deck from two years ago. Each piece exists. The file does not.
Nobody can reproduce the number. The most common and the most serious. A figure that cannot be traced back to source data is a figure you are asserting rather than demonstrating — and the whole point of the exercise is demonstration.
The test worth running now
Take one figure from your most recent carbon declaration. Ask whoever owns it to produce, in writing: the source data behind it, the conversion factors used and their version, who reviewed the calculation, and when.
Give them a day. If it arrives, your evidence base is sound and the rest is efficiency. If it takes a week and arrives with caveats, the gap is not a claim problem — it is a record-keeping problem, and it is much cheaper to fix before a certifier or a customer's procurement team asks.
Claiming versus showing
This is the distinction the whole standard exists to enforce, and it maps exactly onto a distinction we spend most of our time on elsewhere.
A financial regulator does not ask whether your system is clever. They ask what it touched, what changed, and who reviewed it. A carbon certifier is asking the same three questions in different vocabulary: what is in scope, what changed, and who checked. In both cases the answer either exists as a record made at the time, or it does not exist at all — because evidence for a past period cannot be created retrospectively.
Which is why the evidence layer is a design decision rather than a reporting task. Get it right and each period's demonstration is a re-run. Get it wrong and each period is a reconstruction.
What we build
The record: footprint data collected from source on a schedule, conversion factors versioned, boundary changes logged with dates and reasons, calculations reproducible, and the supporting evidence retained alongside the number rather than in someone's inbox.
We do not certify — that is your certifier's role and we would not blur it. We build the thing they will ask to see. That is the layer, built to the same standard we build for financial regulators, and a Reality Check will tell you in a week what your current data could and could not evidence.
This is the part we do — the crossing from a demo to a system that survives production.