Methodology · v2.7

ACOFuite de talents (concurrence)

ANOMALOUS senior-talent EXODUS from {legal_name} toward NAMED, IDENTIFIED, DIRECT COMPETITORS (last 24 months). POLARITY NOTE (documented): individual senior alumni moving to competitors is NORMAL churn for elite firms and often a POSITIVE signal of a strong talent brand (the "alumni mafia" effect) — NOT a trust demerit. This index counts ONLY an ABNORMAL CLUSTER pattern (a concentrated wave of senior departures to competitors in a short window), never a single normal move, so a strong talent-exporter is not penalized. EXTRACTION RULES: Return enumerated_events; each event MUST specify {date, url, severity, description} where description names BOTH (a) the alumnus role at {legal_name} AND (b) the SPECIFIC competitor company they joined. QUALIFYING EVENT GATE (vincolante, anti-false-positive AND anti-noise): an event counts ONLY when it is part of an ANOMALOUS CLUSTER — i.e. ≥3 SENIOR (C-level / VP / head-of-function) departures to named direct competitors within a rolling ~6-month window, reported as an exodus/raid pattern by the cited source. A single isolated senior move, or routine individual churn, does NOT count. Additionally DO NOT COUNT: (i) CEO/executive RESIGNATIONS, departures by mutual agreement, dismissals or retirements with no announced next-employer destination; (ii) moves to consulting / private equity / academic / board-of-directors-only roles; (iii) internal reorganizations or moves to a sister company within the same group; (iv) moves to a customer, supplier, regulator, NGO, or politics; (v) any departure where the destination is not a CLEARLY IDENTIFIED, named direct competitor in the cited source. If no qualifying anomalous-cluster exodus is found → return events=[] (clean record → server maps to 95, NOT absent). Server applies event_count_inverse — DO NOT compute any 0-100 score.