A product footprint often arrives as one polished total, although the calculation begins by drawing boundaries. Life-cycle assessment follows linked stages from raw-material acquisition through production, use and disposal. A factory-gate study may answer a useful question, but a reader should not assume that it represents the whole chain. Before comparing totals, the reader needs the service being measured, the stages included and the conditions under which the product is used.
The functional unit defines the service. Counting one shopping bag is weaker than comparing the delivery of a stated load over a stated number of trips. A durable item may require more material at first yet perform the service many times. The consumer receives that service rather than merely possessing an object. Without a common function, a comparison can reward a lighter object while ignoring how often it must be replaced.
The system boundary defines the processes. Analysts decide whether to include extraction, transport, factory energy, washing, repair, recycling and final disposal. They may exclude a small flow when data are unavailable, but the exclusion should be visible. Two studies can use sound measurements and still reach different rankings because their boundaries answer different questions. A changed boundary alters the model without making either measurement automatically false.
Every model also contains an assumption about behaviour. How far will the product travel? How much energy will it consume during use? How many times will it perform the function? A result can be approximately right for one pattern of consumption and misleading for another. Worldwide averages help when local information is absent, but they can hide electricity sources, transport distances, repair habits and waste systems that vary by place.
Conditional does not mean useless. It means that the number belongs to a stated decision. A careful report exposes its function, boundary, data sources, exclusions and uncertainty before displaying the final total. A careful reader asks whether those choices match the proposed action. More decimal places cannot repair a mismatched scope; a transparent model is valuable because another person can understand, challenge and improve it.