Primary or secondary data: why the distinction decides your grade
· 5 min readThe same carbon footprint value does not carry the same weight depending on where it came from. Understanding what 'primary' means avoids building a declaration on sector averages.
In a battery compliance file, two companies can declare the same carbon footprint and be treated differently. The reason lies in where the data came from, not in its value.
The definition, in practice
Primary data is measured or calculated on the actual process concerned: the electricity consumption metered on the production line, the assay of a cathode batch, the measured yield of a furnace. It is specific to the company and the installation.
Secondary data comes from a generic source: a sector average, a default value from a reference framework, scientific literature. It describes a typical process, not yours.
Both are legitimate. The Regulation does not prohibit secondary data — it requires you to know which you are using and where.
Why it changes the outcome
Three concrete consequences.
Uncertainty. A declaration resting mostly on averages carries wider uncertainty. Against a regulatory threshold, wide uncertainty works against you: it moves you closer to the lower performance class.
Defensibility. Under inspection, primary data rests on a dated reading and an identified process. Secondary data rests on a bibliographic reference the inspector can challenge, or whose representativeness for your site can be disputed.
The improvement effect. A real investment — on-site renewable electricity, heat recovery, material yield — only shows up in your declaration if you measure. A sector average is insensitive to your progress: you pay for the investment without reaping the declarative benefit.
The reflex that makes the difference
Do not try to measure everything. Identify where secondary data costs you most.
Concretely: rank your life-cycle stages by contribution to the total footprint, then look at which rest on averages. The stage carrying 40% of your footprint on a default value is your first measurement project. The stage carrying 2% does not deserve the effort.
Provenance marking
Good practice is to store, alongside each value, an indication of its origin: measured, supplier, estimated, default. That marking has three uses:
- automatically compute the primary-data share of your declaration;
- identify what must be re-collected when a supplier changes;
- answer a verifier's question in minutes rather than days.
It is one more field in your data model. It is also what separates an auditable file from a spreadsheet that has to be reconstructed from memory.