GEO Pulse #92 - August 9, 2026

Issue #92 · August 9, 2026

This Issue

• Wikipedia (13%) + Reddit (12%) = 25% of all US ChatGPT citations — WSJ, NYT, Bloomberg not in top 20
• ChatGPT citations per answer fell 41% in 90 days: from 4.95 to 2.96 (SEO Clarity, 170M answers)
• Google's ARD spec launched June 2026 — none of the 11 founding members has implemented it yet
• SEMrush 50K-brand study: AI visibility is a topic-level game, not a domain-authority game

Wikipedia and Reddit Now Drive Over 25% of US ChatGPT Citations — WSJ, NYT and Bloomberg Don't Make the Top 20

May 2026 · 5W Public Relations Research

5W Public Relations analyzed US ChatGPT citation patterns and found that Wikipedia (13.15%) and Reddit (11.97%) together account for more than 25% of all ChatGPT citations — a combined share larger than the next 15 sources combined. The Wall Street Journal, New York Times, Bloomberg, and Financial Times don't appear in the top 20 at all. Forbes, at #18, is the only major US business publication that made the list.

LinkedIn moved from #11 to #5 in three months, now appearing in 14.3% of ChatGPT Search responses — overtaking traditional media at a pace that suggests structural platform preference, not just recency. Brands listed on G2, Capterra, Trustpilot, and Yelp see a ~3x citation multiplier. The finding that upends most PR strategies: earned media from high-authority branded outlets is producing fewer citations than community platform presence and directory listings.

Read the full report →

ChatGPT Citations Per Answer Fell 41% in 90 Days — And Brands Are The Most Affected

2026 · Lyra / SEO Clarity — 170M Answers Analyzed

A 170-million-answer analysis covering 3,000+ tracked brands found that citations per ChatGPT answer fell from 4.95 to 2.96 between December 2025 and March 2026 — a 41% drop. For brand-name queries specifically, the decline was steeper. OpenAI appears to be consolidating its answers around fewer, higher-authority sources rather than spreading citations widely. For brands that built GEO strategies around citation volume, this shift requires a recalibration toward citation quality and platform diversity.

See the data →

ARD & Agent Discovery

Google's ARD Spec: Zero Adoption Even Among Its 11 Founding Members

A census of all 11 ARD working-group members (Google, Microsoft, GitHub, NVIDIA, Salesforce and more) conducted June 18 found that none had published a discoverable ai-catalog.json file — including Google itself. The spec defines how AI agents find and verify website capabilities at runtime. Its adoption window is wide open.

SEMrush Research

AI Visibility Is a Topic-Level Game: 50,000-Brand ChatGPT Study

SEMrush analyzed 50,000 brands across ChatGPT responses and found that citation likelihood is determined at the topic level, not the domain level. A brand can have high domain authority and zero AI citations if it lacks topic-specific depth and third-party corroboration in the topics its buyers actually query.

Navarro's Take

"The 5W research on Wikipedia and Reddit dominating ChatGPT citations should end the debate about where to invest in GEO. The question isn't whether to optimize your own website — it's what percentage of your effort should go there versus off-site platforms. Right now, most brands are running at 90% owned content, 10% off-site. The citation data says that ratio should be inverted. The brands that figure this out in 2026 won't be playing the same game as everyone else in 2027."

— Ido Navarro, LLMagnet

GEO Pulse is published by LLMagnet.
Track your AI visibility score at ai-visibility.llmagnet.com