Methodology · v2.10

SCRRésilience de la supply chain

Documented supply-chain disruption events affecting {legal_name} in the last 36 months (rolling). EVIDENCE SOURCES: count ALL distinct disruption events within the last 36 months from BOTH (a) formal company disclosures (10-K risk factors / supplier-list updates / SEC 8-K material event filings / supplier-responsibility annual reports) AND (b) news-tier reporting from Reuters, Bloomberg, FT, WSJ, Nikkei Asia (APAC sourcing), South China Morning Post, equivalent national tier-1 outlets. News-tier coverage IS primary evidence — disruptions that affect production schedules typically reach tier-1 press before formal disclosure. EXTRACTION RULES: each event = date+url+severity+description naming the affected supplier node + disruption type + duration. Severity: multi-quarter halt = critical, single-quarter shortage = major, brief outage = moderate. Dedupe at the EVENT level (same incident covered by 5 outlets = 1 event — pick the most authoritative URL). Older events (2020-2022 COVID era for current rebuilds in 2025+) are historical context — drop them server-side via the 36-month window. If no disruptions in the last 36 months, return events=[] (NEVER absent). Server applies event_count_inverse on events WITHIN the 36m window.