Methodology · v2.9

PARPrix et distinctions

Composite product-awards recognition for {legal_name}. EXTRACTION RULES: Return type="composite_signal" with EXACTLY 3 sub-signals (use these exact names + 0-100 numeric values + a source_url backing each): (a) awards_count_normalized — 0-100 log-normalized count of distinct documented awards over last 10 years; band guide: 0 awards = 15, 1-3 awards = 45, 4-10 awards = 70, 11-30 awards = 85, 30+ awards = 95. EXAMPLE high pattern: documented 25+ iF Design + Red Dot wins over the last 10 years → value=85. EXAMPLE medium (mid-market brand): "5 documented Good Design Awards 2018-2024" → value=70. EXAMPLE low: "no public award entries" → value=15. (b) tier1_awards_share — 0-100 share of tier-1 (iF Gold / Red Dot Best of the Best / Compasso d'Oro / IDSA Gold / Good Design Best 100) over total; band guide: 0 = 20, 1-25% = 55, 25-50% = 75, 50%+ = 90. EXAMPLE high (premium brand): "12/20 awards are Red Dot Best of the Best" → value=90. (c) recent_3y_awards — 0-100 reflecting awards in last 3 years; band guide: 0 = 25, 1 = 55, 2-4 = 75, 5+ = 90. EXAMPLE high: "8 documented wins 2022-2024" → value=90. CRITICAL: each signals[].value MUST be a real assessment (numeric 15-95) based on cited evidence — NEVER return 0 placeholder values. If genuinely no public award data exists for any dimension return type="absent" with reason="not_disclosed" instead of composite_signal with zeros. Cite iF Design Awards / Red Dot / Compasso d'Oro / IDSA / Good Design Award archives per sub-signal. Server applies composite_signal_score — DO NOT compute any 0-100 score.