Many corporate carbon programmes begin with a procurement decision: pick a system first, work out how to use it later. The usual result is a live system that nobody fills in, factors nobody owns, and reports that still depend on outside consultants. The more reliable order is the reverse: define boundaries, inventory the data, and only then choose tools.
Step one is the boundary. State which entities, plants and products are in scope, and how far Scopes 1, 2 and 3 extend. Without a written boundary, no data has an owner and reporting bases drift between departments.
Step two is the data inventory. For each class of activity data within the boundary, list the source: meters, fuel invoices, purchasing records, logistics documents, with the measurement method, frequency and owner. This step exposes most problems, such as a plant whose electricity data exists only in the finance ledger and is not split by production line.
Step three is factor governance. Record the source, version and applicability of every emission factor, and log every change. CBAM, product footprints and the corporate GHG inventory have different factor requirements, so one activity dataset may map to several factor sets; design for that from the start.
Step four is the evidence chain. Every figure must trace back to a source document, and every edit must be logged. This is not only for verifiers; six months on, the company itself needs to see why a number was calculated the way it was.
Only step five is tooling. With the first four in place, the company can say precisely what it needs: which data arrives automatically and which is keyed in, what formats reports must take, who may change what. Choosing a system then becomes a test of whether it can carry the rules already set, not a comparison of feature lists.
Pilot with one product or one plant, run all five steps, produce a first verifiable result, then roll out across the group. Companies that complete the first cycle in three months are not rare. The key is writing the checklist first.