Methodology · v2.6

PSDurabilité produit

Percentage reduction in {legal_name}'s product-related GHG emissions since 2015 (or the company's analog baseline year if explicitly disclosed). EXTRACTION RULES: Cite the company ESG / sustainability / climate report (PDF) plus at least one independent verifier (CDP disclosure, Climate Action 100+ benchmark, MSCI ESG profile, IPCC-aligned sector report). Use the headline scope-1+2 (or product-lifecycle) reduction figure most prominently disclosed. Return exact_numerical{value: <pct_reduction_since_baseline>, unit: "percent", source_urls: [ESG report PDF + verifier URLs]}. Server applies ratio_band (≤20 → 25, 20-40 → 50, 40-60 → 70, 60-80 → 85, 80+ → 95) — DO NOT compute any 0-100 score. DISTINGUISH TWO CASES (anti-fabrication, CLAUDE.md #4): (a) NO documented baseline+achieved-reduction figure exists for {legal_name} (not-found / 0-as-placeholder) → return type="absent" reason="not_disclosed", do NOT emit value=0 or any placeholder number — a missing figure is a data gap (bail), NOT a real measured zero; (b) the company DISCLOSES a reduction that is small, zero, or NEGATIVE (emissions flat or INCREASED since baseline, e.g. value=-3) → that is a REAL measured value and a genuine bad-news datapoint: emit it as a negative/low percent so it scores the deserved bottom band 25, NOT absent. Do not launder a disclosed worsening as a data gap.