Why 90% of ChatGPT's Citations Never Rank on Page One of Google

By David Bellairian, Founder of AIDiscover

I hear a version of the same sentence from almost every brand I talk to: "We rank well on Google — why doesn't ChatGPT ever mention us?" It's a fair question, and the honest answer surprises people: ranking well on Google and getting cited by ChatGPT are much less related than most of us assume.

The number that changes the conversation

According to an analysis by NetRanks of how ChatGPT, Claude, and Perplexity select sources, roughly 90% of the URLs ChatGPT cites in its answers rank at position 21 or lower on Google — meaning almost none of them are on page one, or even page two, of a traditional search. If your mental model is "rank higher on Google, and eventually ChatGPT will notice," that model doesn't match the data.

This isn't a flaw in ChatGPT, and it isn't a flaw in Google. They're solving different problems. Google's ranking algorithm is built to sort the entire web for a query and surface the ten (or so) best matches, weighted heavily by decades of link authority and click behavior. ChatGPT isn't ranking a list for a person to scan — it's synthesizing a single answer from whatever sources its retrieval system considers reliable and relevant right now, and citing a handful of them along the way. Those are different jobs, run on different signals, and it turns out they pick different winners.

What ChatGPT is actually pulling from

Look at where citations concentrate and the picture gets clearer. Per Otterly AI's 2026 citation report, ChatGPT's top cited domains are Reddit, Wikipedia, and Amazon — not top-ranking commercial websites. Community platforms and forums account for over half of all AI citations across engines (52.5% versus 47.5% for brand-owned domains, per the same report). ChatGPT's brand-citation rate specifically sits around 44.7% — meaning more than half the time, when ChatGPT names a source, it's not a brand at all.

That doesn't mean brand sites never get cited. It means the path to being cited runs less through your own domain's Google ranking and more through how consistently your brand is discussed across the wider web — reviews, forum threads, YouTube videos, press coverage, comparison posts written by other people. NetRanks found that sources appearing consistently across four or more platforms are about 2.8x more likely to show up in ChatGPT's answers than sources that exist in only one place. Interestingly, in their analysis, YouTube mentions showed the single strongest correlation (0.737) with a source later getting cited by ChatGPT — higher than backlink counts.

The mistake this leads to

The mistake I see most often is brands pouring their entire AI-visibility budget into their own site: rewriting meta descriptions, adding more schema markup, publishing more blog posts — the SEO playbook, just aimed at a new target. That's not wasted effort, exactly, but it's aimed at the 15–20% of the problem that's actually technical. The bigger lever is the one most companies never touch: showing up, consistently and credibly, in the places AI models already trust — Reddit threads where your category gets discussed, review platforms, YouTube, trade press, and forums where real people are already comparing your options.

What to do instead

A few concrete shifts, based on what the data actually rewards:

Stop treating your own website as the only surface. Your site still matters for AEO (winning the direct-answer box) and for converting the traffic you do get. But for GEO specifically, your presence off your own domain — reviews, forum answers, YouTube mentions, third-party comparisons — is doing more of the work than your on-site content.

Go where the citations already concentrate. If Reddit and Wikipedia dominate citations in your category, a credible, non-spammy presence there (answering real questions, not dropping links) is worth more than another landing page. This has to be done authentically — AI systems and the communities themselves both penalize obvious self-promotion.

Track more than one engine. Google AI Overviews, ChatGPT, and Perplexity don't behave the same way. Google AI Overviews shows the strongest brand-domain preference of the three (nearly 60% of its citations go to recognizable brands, per Otterly), while ChatGPT and Perplexity lean harder on community content. A strategy tuned only for Google AI Overviews will underperform in ChatGPT, and vice versa.

Measure citation, not just rank. Set up a simple habit: ask ChatGPT, Perplexity, and Google's AI mode the 10–15 questions your best customers actually ask before they buy, and note honestly whether you show up and how you're described. That's a more accurate scoreboard for this problem than your Google Search Console rankings.

The gap between "we rank well" and "we get recommended" isn't a bug to be patched with better keywords — it's a sign that a second, parallel discovery system has grown up alongside search, with its own rules. Brands that notice this early have a real head start over ones still optimizing purely for the old scoreboard.

If you want an honest read on where your brand actually stands across ChatGPT, Perplexity, and Google's AI Overviews — not just where it stands on Google — book a free strategy session with AIDiscover and we'll show you exactly what these engines currently say about you.

---

About David Bellairian

David Bellairian is the founder and CEO of AIDiscover, an AI-visibility and brand-discovery agency in Los Angeles, and the creator of AIDiscover's Entity Optimization methodology for keeping brands recognizable to AI systems like ChatGPT, Perplexity, and Grok. Before AIDiscover, he led marketing operations for a SaaS platform, increasing lead quality by 40% through AI-driven automation. Find David on LinkedIn, X, Instagram, or read his full bio.

---

Sources referenced in this article:

• NetRanks, "How ChatGPT, Claude & Perplexity Choose Which Brands to Cite" (2026)

• Otterly AI, "The AI Citation Economy: What 1+ Million Data Points Reveal About Visibility in 2026"