A number on a compliance certificate looks authoritative. It is specific, it is signed, and it sits inside a document that says the standard was met. But a number is only the visible tip of something larger, and the real question — the one that decides whether it holds up when an incident, an insurer or a court comes asking — is quieter: what did you measure, with what, and can you prove it?
The anatomy of a defensible measurement
Behind every reading you can actually rely on sits a chain of custody. Break any link and the number still looks the same on the page — it just stops meaning what everyone assumes it means. The links are:
- The instrument. Calibrated against a traceable reference, and known to be within calibration on the day.
- The method. The right measurement technique for the question being asked, applied correctly, with its known error modes accounted for.
- The conditions. The environment and season recorded, because the same earth measured wet and dry is not the same number.
- The competence. A person who understands what they are measuring and can recognise a result that is wrong rather than merely reading the display.
- The record. An unbroken trail from the raw reading through to the signed certificate, so the number can be traced back to its origin.
Provenance is simply the sum of those links, kept intact. It is unglamorous, and it is the whole game.
Calibration and traceability
Start with the instrument, because it is the link most often taken on faith. A meter that has drifted out of calibration does not warn you — it gives you confident, precise, wrong numbers. That is why calibration has to be traceable: the instrument checked against a reference that is itself checked against a higher one, back to a national standard, on a known schedule. The principles here are the same ones that underpin ISO/IEC 17025. Without that traceability you do not have a measurement; you have a reading, and there is a difference.
The method matters as much as the meter
IEEE 81 exists for a reason: the same grounding system, measured by different methods or with the electrodes in the wrong place, yields different answers — and only some of them are true. Fall-of-potential, clamp-on and continuous techniques each answer a slightly different question and each has conditions under which it misleads. So the method, and the judgement to apply it correctly, are part of the result, not a detail beneath it. A number reported without the method behind it is only half a fact.
A raw number is not assured data
This is the distinction that matters most as monitoring spreads. It is now easy to put a box on a site that emits earthing numbers around the clock. That is genuinely useful — but a stream of readings is not, by itself, assurance. If nobody can say how the sensor was calibrated, what it actually measures, under what assumptions, and how the record is kept honest, then the data is a set of claims with no chain of custody behind them. It may be perfectly accurate. It may also be quietly wrong, and there would be no way to tell. Continuous data earns the word “assurance” only when the same provenance discipline is wrapped around it as around a manual test.
The signature is the accountability
All of which is why the accredited signature, not the sensor, is the thing a certificate really certifies. A signature from an engineer who stands behind the measurement chain is a person accepting accountability for every link in it — the calibration, the method, the conditions, the record. That is materially different from a tick-box or an auto-generated number that no one is answerable for. When the questions get hard, “the system said so” is not an answer; “I measured it, this is how, and here is the record” is.
Why provenance is the next compliance frontier
The direction of travel is clear. As incidents are litigated and insurers scrutinise claims more closely, measurements without provenance will increasingly fail to persuade the people who matter — assessors, regulators, courts. The systems that win will be the ones that can show their work: not just the number, but the traceable, method-sound, recorded, signed chain behind it. This is an engineering view rather than a legal one, but it is where the burden of proof is heading.
The number is the easy part — anyone can produce one. What separates a measurement you can defend from one you merely reported is everything behind it that no one usually sees.