Source Quality Scoring

Metadata and content scoring, classification tiers, and what gets rejected or accepted.

Every search result goes through two scoring passes:

Metadata scoring (0–10)

Evaluates title, domain reputation, and snippet content.

  • High scores (7–10): Major research firms (Deloitte, McKinsey, Gartner), government domains, major publications
  • Medium scores (4–6): Industry publications, established company blogs with editorial standards
  • Low scores (0–3): Marketing agencies, generic tips content, unknown domains

Content scoring (0–10)

After fetching the actual page content, sources are classified:

  • Primary: Original research, surveys with methodology, official reports, press releases
  • Editorial: Quality journalism with original reporting, expert analysis with citations
  • Marketing: Service provider content, agency blogs, content with sales CTAs
  • Low quality: Thin content, repackaged information, generic advice

Sources classified as marketing or low quality are automatically rejected. Remaining sources need a content score of at least 6 to be included.

Join the Skribt waitlist.

Enter your email and we'll save your spot for early access. No credit card required.