Methodology · v2.8

CRPComplaint Response Speed

Speed of {legal_name}'s response to publicly-posted customer complaints. EXTRACTION RULES: Return time_responsiveness with events (each: date = complaint posted timestamp + response_at = company first public response timestamp + severity reflecting complaint magnitude + url to the original complaint). IMPORTANT — TIME WINDOW: events MUST be DATED WITHIN THE LAST 36 MONTHS from today. Search the MOST RECENT Trustpilot reviews, Twitter/X support replies, BBB pages first; older events get dropped server-side as all_events_outside_time_window and the index bails. ACCEPTED SOURCES (mix freely — only 2 events needed for preliminary, 5 for confident): (a) Trustpilot company page — look for reviews with a visible "Reply from {company}" date stamp, sample the 3 most recent; (b) Twitter / X — search the verified company support handle replies (typically @<company>Support, @<company>Help, or equivalent) within the last 12 months; (c) BBB complaint pages where the company posted a reply (page shows both timestamps); (d) Reddit posts in r/{company} or similar with a verified company-rep reply (look for the [V] / verified flair or company-named accounts); (e) Downdetector outage threads where the company posted an acknowledgement. Sample at least 3 events across at least 2 platforms; do NOT cherry-pick fast-resolution cases. For widely-engaged consumer brands the Trustpilot reply timeline alone is typically enough — but verify dates are recent. Server applies time_responsiveness_score (median lag → bucket).