GEO Metrics: How to Measure AI Search Visibility | Patrick Schofield

With Patrick Schofield, Lead relevance engineer at iPullRank · Streamed live · 55 min

What this episode covers

Patrick Schofield explains iPullRank's three-tier reporting model for AI search (input, channel and performance metrics) and findings from a study of about 70,000 URL and query pairs: top-five rankings strongly predict citation, and entity coverage matters more than word count.

Key takeaways

  • ✦Measure AI search in three tiers: input, channel and performance metrics.
  • ✦Treat AI search as its own channel, since rankings alone don't show citations.
  • ✦Balance entities against word count; long pages with few entities rarely get cited.
  • ✦Test on a small set of URLs before scaling changes.

Key moments

  1. 24:32GEO is SEO, new game plus
    “GEO is effectively SEO new game plus.”
  2. 27:28Three-tier reporting model
    “We always fall into this three tier reporting model.”
  3. 39:54Top-five rankings and citations
    “A 90 or 91% higher citation probability if your URL was ranking in the top five.”
  4. 41:23Word count vs. entities
    “Pages that have high word counts, but low entity counts, receive almost no citations.”
  5. 46:52Start with small tests
    “I grab five to 10 URLs, right? Small sample size.”

Quotes are from the episode transcript, lightly cleaned of auto-caption errors.

Questions answered in this episode

What are the three tiers for measuring AI search?
Input metrics show content alignment, channel metrics measure how visible and cited content is, and performance metrics track business impact through the lens of AI referred visitors.
Why can conceptual depth hurt AI citations?
Specific questions need specific answers, and content that covers many related entities bogs down retrieval, making it harder for AI systems to pinpoint the exact answer.
Do you need to rank to get cited by AI?
Ranking helps considerably, but AI looks for quality resources, so you do not inherently need to rank to receive a citation and traffic.

More episodes on AI Search & LLMs