In May a customer's assurance provider asked to review the data behind the reuse figures we supply them. We said yes immediately, partly because we had nothing to hide and partly because we genuinely wanted to know whether our reporting would survive contact with someone whose job is to doubt it.
They spent a day. Here is what happened, written up because almost nobody publishes this and it would have been extremely useful to us a year earlier.
They did not start with carbon
The first three hours were entirely about traceability. Can you show me, for this specific container serial, where it came from, what you did to it, and where it went? Repeated for a sample of fourteen serials drawn from our own outbound records.
The logic is straightforward once you see it: an emissions factor applied to a number you cannot substantiate produces a substantiated-looking number that means nothing. The factor is the easy part. The count is the audit.
- Fourteen serials selected by them, not by us, from twelve months of outbound records.
- For each: intake photograph, prior-contents log entry, grading decision and grader, wash record with cycle and temperature, leak test result, outbound photograph and destination.
- For three that had been granulated: the weight ticket from the receiving processor, matched to a date and a batch.
- For two that had been rebottled: the disposition of the old bottle, again with a weight ticket.
We produced thirteen of fourteen without difficulty. The fourteenth had a gap — a container washed in a week when the wash log had been kept on paper during a controller replacement, and the paper record was legible but not scanned. That is the kind of thing that costs you a finding, and it should.
The baseline question
The second thing they wanted was our baseline, in writing, with a justification for why it was the right counterfactual. We baseline against the manufacture of an equivalent new container — same capacity, same construction, cradle-to-gate.
The reviewer pushed on this properly. Why not baseline against the customer's previous supplier? Why not include the landfill emissions of the container that was displaced? Why a blended steel factor rather than a regional one?
| Choice | Our position | Effect on the number |
|---|---|---|
| New-container manufacture as counterfactual | That is what would have been purchased | Neutral — it is the defensible one |
| Exclude displaced landfill emissions | Too uncertain to claim | Conservative — lowers our figure |
| Blended BOF/EAF steel factor | We do not know which mill made each cage | Conservative — EAF-only would raise it |
| Include our own wash water and energy | Honest accounting requires it | Conservative — lowers our figure |
| Attribute backhaul freight at marginal cost | The leg ran regardless | Favourable — and we flag it explicitly |
Four of the five choices work against us and we said so. The reviewer's comment was that conservative choices are easy to accept and the one favourable choice was fine precisely because it was declared rather than buried.
Where we nearly came unstuck
Language. Not data — language.
A quote template from eighteen months earlier described a figure as carbon saved. Our quarterly report described the same calculation as burden avoided relative to a new-container baseline. Those are not the same claim, and the reviewer, quite correctly, wanted to know which one we meant.
Saved implies a removal. Avoided implies a counterfactual. You are doing the second thing and describing it with the first word, and that is how greenwashing findings start.— The reviewer, and she was right
We changed the quote template that afternoon and swept the website the following week. It is why every environmental figure on this site now says avoided rather than saved or removed, and why the reporting page spells out the difference.
What we changed permanently
- Every wash record is captured digitally at the point of work. No paper interim, ever, even during equipment changes.
- Weight tickets from receiving processors are filed against a batch number the day they arrive, not at quarter end.
- One approved phrase for the claim, used identically in quotes, reports and marketing: avoided relative to a new-container baseline.
- A written statement of baseline and coefficients that goes out with every report, so the reader can disagree with our arithmetic rather than having to trust it.
- An annual internal sample audit of fourteen serials, run the same way, before anyone external asks.
The advice, if you are on the other side of this
If you are a customer intending to report a reuse programme, ask your supplier for a serial and follow it. If they cannot produce the chain in an afternoon, the number they gave you will not survive review, and you will find that out at the worst possible moment.