I own the carbon methodology here, and the question I get most often from customers' sustainability teams is some version of: why does your number look smaller than your competitor's?
Because we subtract
The avoided side of the calculation is the same for everyone and it is not controversial.
| Component | Quantity | Factor | CO₂e |
|---|---|---|---|
| Virgin HDPE bottle | 52 kg | 1.9 kg CO₂e/kg | 99 kg |
| Galvanised cage steel | 27 kg | 2.3 kg CO₂e/kg | 62 kg |
| Steel pallet base | 17 kg | 2.3 kg CO₂e/kg | 39 kg |
| Moulding and assembly energy | per unit | — | 14 kg |
| Total avoided | 214 kg |
Manufacturing burden avoided, per 275-gallon unit
The incurred side is where methodologies diverge, and it is the side that is easy to leave out because nobody is checking.
| Activity | CO₂e | Share |
|---|---|---|
| Inbound and outbound freight | 38 kg | 43% |
| Wash water heating | 18 kg | 21% |
| Yard electricity | 12 kg | 14% |
| Granulation and baling (allocated) | 10 kg | 11% |
| Caustic and chemistry, net of recovery | 6 kg | 7% |
| Buildings and other | 4 kg | 4% |
| Total incurred | 88 kg | 100% |
What reconditioning costs us, per unit
214 minus 88 is 126. That is the only figure we quote anywhere else on this site.
The freight line is the honest one
Forty-three per cent of our own footprint is moving containers around. That is uncomfortable to publish and it is the most actionable number we have.
It is the entire reason we certified 38 independent partner yards instead of building depots. Depots concentrate stock and then pay to distribute it; partner yards put stock where demand already is. Freight miles per tote fell 41% against our 2019 baseline, which moved the 38 kg figure down from 52.
| Year | Avg. haul per tote | Freight CO₂e per unit | Net avoided |
|---|---|---|---|
| 2019 (baseline) | 612 mi | 52 kg | 112 kg |
| 2021 | 478 mi | 44 kg | 120 kg |
| 2023 | 401 mi | 39 kg | 125 kg |
| 2025 | 361 mi | 38 kg | 126 kg |
How the freight term has moved
If a supplier's avoided-emissions figure has no corresponding incurred figure, you are reading a marketing claim with a decimal point on it.
What we also refuse to claim
Four things, each of which would make the number look better.
- No use-phase modelling. Your filling and shipping depend on your product and lanes. We have no basis for estimating them, so we do not.
- Conservative end-of-life credit. We credit recovered regrind at 37 kg per bottle. We do not claim avoided-burden credits for virgin resin displaced downstream, which some methodologies do and which would add roughly 60 kg.
- No projected cycles. A container we stopped seeing is credited with observed cycles only. Assuming it kept performing is the single largest source of overstatement in reuse reporting.
- Stated uncertainty. Coefficients are cradle-to-gate industry averages, accurate to roughly ±15%. We do not quote to three significant figures as though it were measured.
What the uncertainty actually means
People ask what ±15% implies for a reported figure, so here it is worked through for a 400-unit annual purchase.
| Figure | Low | Central | High |
|---|---|---|---|
| Avoided per unit | 107 kg | 126 kg | 145 kg |
| Avoided, 400 units | 42.8 t | 50.4 t | 58.0 t |
| Equivalent vehicle miles | 110,000 | 129,000 | 149,000 |
400 reconditioned units, with the uncertainty band
We report the central figure and state the band. A report that says "50.4 tonnes" without the band invites a reviewer to treat it as measured, which it is not, and the moment that is noticed the whole document loses credibility.
We have been wrong four times
Since 2019 we have corrected this methodology four times. Twice after a customer's sustainability team challenged a coefficient, once after we found we were double-counting granulation energy, and once when the galvanising factor we were using turned out to be for cold-rolled rather than hot-dip.
Every correction moved the number down. That is probably not a coincidence — the errors you do not notice are the ones that flatter you.
What this is and is not
It is a first-party report built on our own asset records and certified scale tickets, with published methodology. It is not a third-party assured LCA and we do not describe it as one.
Used accurately in a filing it holds up. Described as verified it will not, and I would rather say that here than have it discovered in an audit. The full working is on our carbon math, and the certificate format is on diversion reporting.
Written by
Theo Marsh
Materials & reporting
Owns the granulator, the baler and the diversion certificates. Writes the carbon methodology and defends the coefficients.