Look for repeated phrases, consistent upvotes, and multiple perspectives converging on the same obstacle. Longevity matters: if a discussion resurfaces across months, it indicates ongoing demand. Save user verbatims and map them to intents, not just keywords, to catch nuanced needs.
Scrape only where rules allow, or export using built-in tools, and always credit communities. Normalize usernames, dates, and thread depth. Remove sensitive data, deduplicate similar answers, and tag quotes so editors can attribute honestly while preserving valuable context.
Score opportunities using traffic potential, competitive gaps, and internal expertise. Balance quick wins against cornerstone pieces that can anchor clusters. Use a simple matrix to prioritize what to draft next so publishing stays consistent instead of reactive or random.