top of page

Konitzny: How Query Fan-Outs and Source Retrieval shift over time

  • 15. Juli
  • 2 Min. Lesezeit

Key Takeaways:

Konitzny: How Query Fan-Outs and Source Retrieval shift over time
  • Konitzny (PeecAI) analyzed Query Fan-Outs and citation behavior of AI Search / LLM Chats and how they significantly shift over time

    • In May Reddit citations as well as Query Fan-Outs with "reddit" increased and dropped in June again

    • End of June Query Fan-Outs with "Official" (products / site / docs) jumped from 0.x% to 3.5% - in multiple countries --> government authorities, consumer protection sites and domains with "official" gain citation share

    • In May the open-access repository for scientific research papers arXiv increased in citations with GPT 5.5 and 5.6 from 1% to 4% - which is not reflected by a QFO change but rather "a broader adjustment in ChatGPT's retrieval behavior"

    • Mid of March Wikipedia started decreasing from 40% to below 10% in June/July

    • Differences in ChatGPT 5.5 vs. 5.6 Luna:

      • site-query in Query Fan-Outs 0.004% versus 43% in GPT 5.6 Luna

      • "Official" in Query Fan-Outs 1.4% versus 21% in GPT 5.6 Luna

      • Year-based modifiers in Query Fan-Outs 6.2% versus 26% in GPT 5.6 Luna

      • "Best"/"Top" in Query Fan-Outs 8.5% versus 5.4% in GPT 5.6 Luna

      • Quotation marks (in combination with site-query or without) e.g. for brand contact info, tariffs/product info, attributes (site:www2.hm.com/en_gb/productpage "Organic cotton" "Straight jeans" "Black" "Ladies" "H&M")

  • Ray points out that Google with Search went a similar path trying to counter spammy SEO tactics

    • In this context AI Search / LLM Chats try to focus on EEAT sources to create higher quality answers, to make it harder for spammy domains to earn citations and "game the algorithm" - and adds further patterns identified:

    • Hinckley observed ChatGPT adding "case studies" to Fan-Outs

    • Reynolds observed ChatGPT adding known brands to Fan-Outs

  • Consequently, you do need to monitor all parts of AI Search / LLM Chats (see DEJANs masterpiece "How AI Search works": from model perception to retrieval via Query Fan-Outs, citations etc.) to detect changes over time and adapt or rather finetune your strategy


" Tracking mentions, retrieved sources, Query Fan-outs, and retrieval behavior is becoming increasingly important to understand why visibility changes occur." - David Konitzny


Sources:

© 2026 David Epding.            Erstellt mit Wix.com.

logo_new_100_edited.png

David Epding ist AI Search (GEO & SEO), Data Analytics und Automation Manager mit über 10 Jahren Erfahrung in Technischem SEO mit breiter Expertise für LLMs und langjähriger Erfahrung in der Daten-Analyse.

bottom of page