A factory does not have one carbon footprint. It has as many as it has product families, and the allocation rule between them is a commercial decision disguised as an accounting one.
| Line item | What we measure | Common failure |
|---|---|---|
| Line electricity | Sub-meter per line, 15-minute intervals | Metering the whole plant, then dividing by output |
| Process heat / fuel | Metered volume × fuel factor | Estimating from production hours |
| Compressed air | The largest hidden load in most plants | Omitted entirely |
| Material input | Mass balance per SKU | Counting purchased mass as consumed mass |
| Embedded tooling | Amortised capital carbon | Ignored, or allocated to year one only |
| Outbound logistics | Distance × mode × load | A single national average |
Compressed air deserves its own line. In many plants it is 10–20% of electricity and nobody meters it, because leaks are invisible until someone puts a meter on the compressor.
Three rules cover almost every case. Each gives a different answer for the same plant; what makes one of them correct is that it was declared before the measurement and is reproducible by a third party.
| Rule | Use when | Fails when |
|---|---|---|
| By mass | Products are physically similar | One product is heavy and cheap, another light and expensive |
| By economic value | Value drives the process burden | Prices move quarterly and the factor moves with them |
| By machine hour | Production is the dominant driver | Shared utilities are large and un-metered |
| Stage | Days | From you | You get |
|---|---|---|---|
| Scope | 2 | Product families, allocation preference | Boundary decision |
| Document review | 4 | Meter list, mass balance, factor sources | Findings log |
| Verification | 3 | Plant access, production records | Sample results, reconciliation |
| Technical review | 2 | Nothing | Statement per family |
An injection-moulding line makes three housings from the same polymer. Electricity is 78% of the line footprint, and the three SKUs run for 60, 25 and 15 percent of machine hours. The mass-weighted result would over-credit the heaviest part; machine hours track the actual load.
| SKU | Machine hours | Line energy | kgCO₂e / unit |
|---|---|---|---|
| A | 60% | proportional | 0.41 |
| B | 25% | proportional | 0.28 |
| C | 15% | proportional | 0.19 |
Usually no. Most plants already have line-level or machine-level metering in the SCADA system; it is simply not exported into the carbon model.
Yes, as a design-stage estimate with a clearly stated uncertainty. It is not a label until it ships.
Annually, or whenever the allocation rule or product mix changes by more than the stated tolerance.