Life cycle assessment is standardised into four phases, which gives it an orderly appearance. In practice the first phase determines the outcome more than the remaining three combined, and it is the phase most often rushed because it involves no data.
The Four
| Phase | What happens | Where the leverage is |
|---|---|---|
| Goal and scope definition | Functional unit, boundaries, allocation rules | Very high |
| Inventory analysis (LCI) | Compile inputs and outputs | Effort-intensive |
| Impact assessment (LCIA) | Convert flows to impact categories | Mostly mechanical |
| Interpretation | Conclusions, sensitivity, limitations | High, often truncated |
The Functional Unit Is the Whole Argument
Everything is expressed per functional unit, and choosing it choosing what the study can conclude.
Whether measured per tonne of green coffee, per hectare, per cup delivered, or per unit of caffeine, these metrics produce different rankings of the same farming systems. A high yield intensive system looks good per tonne and poor per hectare. A low input system looks the reverse.
Neither is wrong. They answer different questions, and a study that does not state which question it answers has produced a number that can be quoted in support of almost anything.
Allocation Is the Contested Step
When a process yields more than one output, which is nearly always the case in agriculture, the emissions have to be divided between them.
Palm oil yields crude oil, kernel, kernel meal and empty fruit bunches. Coffee yields beans, husk and pulp. Dividing by mass, by economic value, or by energy content produces materially different footprints for the main product.
ISO sets a preference hierarchy: avoid allocation by subdivision or system expansion where possible, then allocate on underlying physical relationships, then on other relationships such as economic value.
In practice economic allocation is widely used because physical relationships are often unclear, and economic allocation means the footprint moves when prices move, which surprises people the first time a recalculation shifts the answer with no change in farming.
Boundaries Decide What Counts
Cradle-to-farm-gate stops at the farm boundary. Cradle-to-grave continues through processing, transport, use and disposal.
For agricultural carbon work, cradle-to-farm-gate is usually appropriate: it captures fertiliser and pesticide manufacture, on-farm energy, and field emissions, which is where the agricultural footprint sits.
The requirement is stating it. A farm-gate figure compared against a cradle-to-grave figure is not a comparison, and this mismatch appears in real corporate reporting more often than it should.
Interpretation Is Where Studies Get Weak
The fourth phase requires sensitivity analysis, completeness checks and explicit limitations. It is the phase most often reduced to a summary.
A study reporting a single number with no sensitivity analysis has not shown whether the conclusion survives plausible alternative assumptions, such as those regarding allocation, emission factors, or the functional unit. Where it does not survive, that is the most important finding in the study, and it is the one a truncated interpretation phase omits.
Why This Matters for Carbon Work
Buyers increasingly want both an LCA footprint and a carbon-methodology reduction, as discussed elsewhere on this site. Anyone producing both needs to know which ISO choices were made, because the same underlying farm data can yield quite different published figures depending on decisions made in phase one.
Document those decisions. When two numbers disagree, the explanation is almost always in the goal and scope definition rather than in the data.
Turn climate goals into a verified portfolio
TREEO connects the full carbon cycle, encompassing eligibility, simulation, real time monitoring, and registry ready reporting, while combining expert consulting with dMRV technology so the evidence exists before anyone asks for it.
