목적 (Purpose)
Defines the principled boundaries (GHG Protocol) of the three downstream sold-product categories — 10 (processing), 11 (use), 12 (end-of-life) — and the three calculation structures (top-down / bottom-up / hybrid) with the criteria for choosing each.
Principled Boundaries per GHG Protocol
All three categories cover emissions caused downstream by products sold in the reporting year. Because downstream activity happens outside the reporting company's control, all three hinge on assumptions and scenarios about where products go, how they are used, and how they are treated.
- Category 10 — processing of sold intermediate products: the scope 1·2 emissions of third parties processing the company's intermediate products into final products. Applies only to sellers of intermediate products; harder when the final use of the intermediate is unknown.
- Category 11 — use of sold products: lifetime use-phase emissions of sold final products. Direct use-phase emissions (products that consume energy/fuel, are fuels/energy themselves, or contain/emit GHGs) are the minimum boundary; indirect use-phase emissions are optional. In principle, reporting-year sales × expected lifetime × use profile attributes the full lifetime emissions to the reporting year.
- Category 12 — end-of-life treatment of sold products: emissions from disposing of sold products (and packaging) at end of life. Requires assumptions on the mix of treatment paths (landfill, incineration, recycling), which varies with regional waste infrastructure.
Three Calculation Structures
These three categories do not fit a fixed template: assumptions, scenarios, and allocation differ by company and by product — even within one industry, differing product lifetimes and end-of-life routes change the calculation itself. So the first decision is not which factor to use but what structure to build: top-down (TD), bottom-up (BU), or HYBRID. They differ not in precision but in what level of data you start from; selection criteria follow below.
Factor choice comes after the structure and is a human methodological decision, not something that can be settled automatically. Downstream emissions are dominated by use-phase assumptions (lifetime, use intensity, regional grid mix), so an assumption not recorded alongside the value cannot be defended in assurance.
TD, Top-Down
Structure: start from one corporate-level aggregate (e.g., annual total sales volume) and allocate to categories 10·11·12 by multiplying per-unit intensity factors for processing, use, and end-of-life.
When to use: product-level detail (use profiles, end-of-life scenarios) is not yet available and only aggregates exist; screening-stage estimates to prioritize which category is material; or the product mix is homogeneous enough that a single intensity assumption does not distort much.
상향식 (BU, Bottom-Up) — 기본값 (default)
Structure: calculate per-category emissions per product/product line as individual values, then assemble by summing within each category. Typically product-level value nodes combine with sales-record references (summed by product/period conditions).
When to use: product-level use profiles, lifetimes, and end-of-life scenarios are available for precise calculation; per-product-line contribution analysis is needed (identifying reduction targets); or products are heterogeneous enough that a single intensity assumption distorts.
Prefer bottom-up where the data allows it — real downstream reduction starts from knowing which products emit how much, and a top-down total cannot tell you what to change.
혼합 (HYBRID)
Structure: bottom-up for major product lines with data, top-down allocation for the remaining long tail.
When to use: the transition period of progressive refinement starting from top-selling lines — expanding bottom-up coverage while still reporting the full boundary without gaps.
Shared Cautions
- All three categories produce annual results (no monthly dimension).
- References resolve at execution time, so re-running after source updates can change results — pin reporting values by adopting the run.
- Make assumptions (lifetime, use profile, treatment mix) explicit as value nodes so they appear in the formula text — verification must be able to trace where each assumption came from.