How ChatGPT chooses sources and how to check your citations

ChatGPT Search selects sources using multiple factors aimed at relevant, reliable answers. OpenAI does not publish a formula that guarantees a citation. You can check whether your website is cited in an answer by inspecting its source links, then track visits from clicked links separately.

This article separates OpenAI's documented access requirements from findings in citation studies. The studies describe associations, which can guide content work but do not establish a guaranteed ranking recipe. For the broader search comparison, see SEO vs AEO .

Can I see when ChatGPT cites my website?

You can observe citations in answers you inspect. That gives you a sample, not a complete count of every answer that mentions your site. OpenAI's ChatGPT search guide explains how to open citations and the Sources list. That list can also include relevant links that were not cited in the answer.

Keep a small observation log and repeat the same procedure:

1. Choose questions your intended readers ask. Keep checks that name your brand separate from questions that do not name it.
2. Run each question in a new chat with web search. Record the exact prompt and date, plus the model and language. Note whether memory is enabled so you can interpret changes.
3. Inspect the answer's citations and open the linked sources. Log the cited URL and save the answer or a screenshot. Record a brand mention without a link separately; a link in the Sources list alone is not an inline citation.
4. Repeat on a schedule using the same prompts. Record when no citation appears. A single result confirms only that observed answer.

Track clicked referrals separately

OpenAI's publisher FAQ documents utm_source=chatgpt.com in referral URLs. In analytics that records campaign sources, filter for chatgpt.com and review landing pages over the same period. Exclude your own test clicks. This report measures attributed visits. It cannot count citations without a click or prove that a visit came from an inline citation.

Keep technical checks alongside this log. The IsAgentReady scanner checks published access and content signals. It does not detect ChatGPT citations or rankings, verify agent task success, or guarantee inclusion. A crawler request in server logs confirms a request, not a citation.

The Search Infrastructure Behind ChatGPT

ChatGPT Search can use external search providers and retrieve relevant web content. OpenAI describes rewriting a question into targeted queries, sometimes followed by further searches in its search documentation .

ChatGPT Search was originally built on Bing's search index . Two independent studies from mid-2025 found evidence that the paid version of ChatGPT also uses Google Search results. In May 2025, OpenAI also added Shopify as a third-party search provider for commerce queries. ChatGPT therefore does not depend on a single search backend.

The Three Crawlers You Need to Know

OpenAI operates three web crawlers, each with a different purpose. Crawler access affects whether OpenAI can retrieve your content for search. The breakdown from OpenAI's official crawler documentation :

OAI-SearchBot and GPTBot share crawl data. If your site allows both bots, OpenAI can use a single crawl for both purposes to avoid duplicate requests.

This robots.txt configuration allows all three crawlers:

# ChatGPT search crawling
User-agent: OAI-SearchBot
Allow: /

# AI model training (optional - blocking this does NOT affect search)
User-agent: GPTBot
Allow: /

# User-initiated browsing (robots.txt may not apply)
User-agent: ChatGPT-User
Allow: /

From OpenAI's Publishers FAQ : "Any public website can appear in ChatGPT search." You can therefore appear in search even if you opt out of AI training by blocking GPTBot. The crawlers are independent.

Domain Authority Is the Strongest Factor

One of the largest studies of ChatGPT citations comes from Wellows (7,785 queries, 485,000+ citations). It found that referring domains (backlinks) are the strongest predictor of whether ChatGPT cites a website.

Sites with 2,500 referring domains averaged 1.6 citations per query, and sites with 350,000+ referring domains averaged 8.4, about five times as many. Domain traffic is the second factor, but that correlation only shows up at very high volumes.

Search Engine Journal's analysis of the same dataset agrees: domain-level authority outweighs page-level metrics. In other words, the domain matters more than the individual page.

Google ranking position does correlate with ChatGPT citations. Positions 1-45 averaged 5 citations versus 3.1 for positions 64-75. The likely reason is that the same signals (backlinks, authority) drive both.

Long-Tail Sites Get Cited Too

The Wellows report found that the top 50 websites capture only 48% of all mentions. The remaining 52% goes to smaller, niche sites, so a site does not need to be Wikipedia to be cited.

The Profound AI Search Shift study found that only a small portion of ChatGPT's citations match Google search results, so ChatGPT's source selection is largely independent of Google's. A site that ranks poorly on Google can still be cited by ChatGPT if it has authority and clear content structure.

How to Structure Content for ChatGPT Citations

A study by Kevin Indig / Growth Memo (published via Search Engine Land ) analyzed 3 million ChatGPT responses and 30 million citations. These are correlations in what cited content tends to look like. The study does not show that any of them cause citations.

Front-Load Your Key Information

44.2% of citations come from the first 30% of content in a consistent "ski ramp" pattern. Information at the top of an article is cited far more often than content at the bottom. This resembles an encyclopedia-style opening, not the journalistic "inverted pyramid".

Use Q&A Heading Structure

Cited content skews heavily toward conversational Q&A structure. Among citations tied to questions, 78.4% came from H2 headings. Cited pages often have an H2 that asks a question, with the answer in the paragraph below. Writing headings as questions ("How does X work?" or "What is Y?") matches the way users phrase their queries to ChatGPT.

Pack in Specific Entities

Cited text averaged 20.6% proper nouns (versus 5-8% in typical English). Specific brands, tools, people, and place names reduce ambiguity and make content easier to verify. Write "Stripe, Shopify, and HubSpot use this approach" instead of "many companies use this approach."

Strike the Right Tone

Cited text clustered at a subjectivity score of 0.47. That sits between dry fact and emotional opinion, in the range of analyst commentary: fact plus interpretation. Text at a Flesch-Kincaid grade level of 16 was cited more often than dense academic prose at 19.1.

Content and Section Length

The Wellows report adds these findings:

• Total length: Articles under 800 words averaged 3.2 citations; over 2,900 words averaged 5.1
• Section length: 120-180 words between headings performed best (4.6 citations average)
• Expert quotes: Pages with expert quotes averaged 4.1 citations versus 2.4 without
• Statistical data: Content with 19+ data points averaged 5.4 citations versus 2.8 for minimal data
• Freshness: Content updated within 30 days is associated with about 3.2x more citations (Wellows)

FAQPage Schema and Citations

A study on ChatGPT visibility found a strong correlation between structured data and citation rates:

6.2% of ChatGPT-visible websites had FAQPage schema versus only 0.8% of non-visible websites , an 8x gap. Cited sites are far more likely to carry FAQPage schema, but the study does not show that schema makes a site get cited. FAQPage makes Q&A machine-readable. Structured data is one of several factors in AI agent readiness .

A 2026 Ahrefs study of 1,885 pages found that adding schema did not measurably change AI-search citations. Cited pages tend to be well-structured, but schema does not appear to drive the citations directly.

For e-commerce sites, OpenAI accepts structured product feeds (title, description, image, brand, SKU, price, availability, GTIN) for ChatGPT Shopping.

Who Gets Cited the Most?

Several studies have measured which domains dominate ChatGPT citations. The Ahrefs study (9.6 million queries) found the top cited domains in the U.S. are Reddit, Wikipedia, Amazon, Forbes, and Business Insider. Wikipedia is cited by ChatGPT at 16.3% (versus 12.5% on Perplexity and 8.4% on Google AI Overviews).

The Profound study (730,000 conversations, Q4 2025) adds these figures:

• Wikipedia appears in ~1 in 6 conversations with citations (18%)
• Reddit appears in 13%
• Reuters and NIH each at 4%
• Turn 1 is 2.5x more likely to trigger citations than turn 10, and 4x more than turn 20

The Visual Capitalist / Ahrefs analysis (78.6 million searches) found Reddit leads across all AI models with 40.1% citation frequency, followed by Wikipedia at 26.3%.

How ChatGPT Search Differs from Google

The table below compares the two, based on the figures above:

What You Can Do Today

Based on the research, these actions are ranked by strength of evidence:

• Include expert quotes and statistics . Content with 19+ data points averaged nearly 2x more citations
• Keep sections 120-180 words. This length between headings performed best in the Wellows report
• Update content regularly . Content fresh within 30 days is associated with about 3.2x more citations
• Use specific entities . Proper nouns, brand names, and tool names reduce ambiguity
• Server-side render your content . Do not assume ChatGPT's crawlers run client-side JavaScript. Learn more about how AI agents see your website through the accessibility tree

Summary

In the studies above, cited sites share high domain authority, well-structured content with Q&A headings, structured data, and fresh updates. ChatGPT's citations are not a copy of Google rankings.

Since 52% of citations go to sites outside the top 50 domains, specialized sites can be cited without top-50 authority. For all findings in context, see the analysis of Vercel's 2026 AEO report .

For other AI systems, read the companion posts on how Claude selects sources to cite and how Google AI Overviews selects sources .

Sources

• OpenAI Crawlers Documentation - Official crawler specifications and robots.txt guidance
• OpenAI Publishers and Developers FAQ - Official guidance on ChatGPT Search visibility
• OpenAI: Introducing ChatGPT Search - Official launch announcement
• Wellows: 7K Queries, 485K Citations - Large-scale ChatGPT citation study (7,785 queries)
• Search Engine Land: 3M ChatGPT Responses, 30M Citations - Content structure analysis
• Search Engine Journal: Top Factors Influencing ChatGPT Citations - Domain authority analysis
• Insightland: Structured Data and AI Search - FAQPage schema visibility correlation
• Ahrefs: 100 Most Cited Domains in ChatGPT - 9.6 million queries analysis
• Profound: How ChatGPT Sources the Web - 730,000 conversations citation analysis
• Visual Capitalist: Most Cited Websites by AI Models - 78.6 million searches cross-platform analysis
• Profound: AI Search Shift - ChatGPT source independence from Google
• OpenAI: Product Feed Specification - ChatGPT Shopping structured data