Methodology · v2.10

PRAffidabilità del prodotto

Reliability ratings for {legal_name}'s products from RECENT (last 36 months) CATEGORY-APPROPRIATE independent durability/quality testers. EXTRACTION RULES: Return rating_aggregate with one entry per source (each: platform, score, scale, source_url, review_count). SOURCE SELECTION (agnostic, category-adaptive — do NOT restrict to electronics/automotive benchmarks): pick the independent testers RELEVANT TO THE PRODUCT CATEGORY. Examples (illustrative): electronics/automotive → Consumer Reports reliability surveys, SquareTrade / Allstate Protection Plans claim-rate reports, J.D. Power dependability studies, iFixit Repairability Index; appliances/general goods → Stiftung Warentest, Which?; apparel/footwear/textiles → durability tests + OEKO-TEX / longevity reviews; food/supplements → ConsumerLab; furniture/homeware → sector testing labs / longevity reviews. Plus manufacturer warranty-claim data and Statista product-satisfaction surveys where available. REQUIRED: every cited rating MUST have a publication date within the last 36 months from today_iso. Drop ANY source older than 36 months even if no recent replacement exists — return rating_aggregate with fewer entries (or absent) rather than back-fill with stale data. EXPLICIT DROP LIST: Rescuecom Computer Reliability Report (defunct after 2020) is NOT an acceptable source. Use the published reliability/dependability rank score on its native scale; do NOT translate or normalize. Server takes the trimmed, log-review-count weighted median × 20 — DO NOT compute any 0-100 score. If NO category-appropriate independent tester covers {legal_name}'s products, return type="absent" reason="not_disclosed" (clean not_applicable) — do NOT fabricate a rating.