Improve your skills with weekly insights from Growth Memo experts. Subscribe for free!
When an AI answers a question using your content, it will usually quote you with a source link. What he doesn’t do, 62% of the time, is say your name. The link is there. The mention of the brand is not. This is what I like to call a ghost quote: the AI using your content does not mention you in the response.
This week, I share:
- Why being cited and being mentioned are two different outcomes that require different strategies.
- Which LLMs name brands and which treat them as anonymous sources.
- The query format and content type that produces 30x more brand mentions.
A word from Kevin: I’m a big fan of HubSpot Counter-current marketing. I had Kieran, one of the co-hosts, on my Tech Bound podcast in 2023. Now they’ve launched a newsletter with clever experiments, new perspectives, and practical lessons on what works now. So I thought of sending out a friendly shout: Check it out.
This analysis is based on 3,981 domains in 115 prompts, 14 countries and four AI search engines (ChatGPT, Google AI Overviews, Gemini, AI Mode), using data from Semrush AI Toolkit. Each appearance is labeled as “cited” (source link present) and/or “mentioned” (the brand name appears in the response text). The gap between these two states is the problem of phantom citations.
1. 62% of your brand’s LLM citations are functionally invisible
Most brands assume that being mentioned means being seen. The data says otherwise.
74.9% of domains were cited and 38.3% mentioned. 61.7% of citations are ghost citations: the domain gets a source link but no name recognition in the answer text.
Only 13.2% of appearances turn into both a quote and a mention. Not a single area was cited but not mentioned at all, or vice versa.
2. Each LLM shows different behavior
The four AI engines process citations and mentions in fundamentally different ways:
- Gemini names brands in 83.7% of appearances, but only generates a citation link 21.4% of the time. It functions more as a conversationalist drawing on brand awareness.
- ChatGPT is the opposite: it cites 87.0% of the time but only mentions brands in 20.7% of responses, functioning more like an academic article with footnotes.
- Google AI (AIO) previews fall in the middle but lean toward quoting.
- Google’s AI mode offers about 17% more brand mentions than ChatGPT in its results, but also performs closer to an academic article than its Gemini sibling.
For brands, this means that Gemini visibility and ChatGPT visibility are not the same thing. (This data set clearly showed that there wasn’t much overlap with ChatGPT citations/mentions and Gemini citations/mentions for the same prompts.) Optimizing for one doesn’t help for the other. There is no single “AI visibility metric.” There are at least 4 different behavioral systems operating in parallel.
3. Strong brands are named in the text
A clear pattern emerges among domains appearing three or more times: content aggregators and academic sources are cited several times but almost never mentioned.
- Medium.com has been cited 16 times for the same prompts on three different engines and named zero times.
- Wikipedia.org was cited 27 times and mentioned in only two answers, both times for the same conversational query (“What is the most dangerous creature in the world? »).
- Wired.com, sciencedirect.com, harvard.edu: same model.
Consumer brands with a strong public identity are mentioned in the results at almost 100%. The AI doesn’t feel the need to quote. Instead, he mentions mainstream brands outright. It knows that brand data comes from somewhere, but doesn’t feel the need to explicitly tell users that. For publishers whose value proposition is news authority, this is a structural problem.
*A mention rate above 100% means that the brand is named in the answer text even when it is not cited as a source link – the engine references the brand by name without linking to it. For values in this dataset greater than 100%, consider being cited 10x and mentioned 10x as = 100%. If a brand is mentioned 12 times and cited 10 times, that’s 120%.
4. LLMs disagree on the same brand 22% of the time
454 invite+domain combinations were tested on several engines. In 22% of these results (100 total), LLMs disagreed on whether to mention the brand:
- Instagram.com was mentioned by ChatGPT and Gemini but only cited (not named) by Google.
- Facebook.com was mentioned by Gemini in 3 out of 3 appearances.
- Google’s AI mentioned Facebook 9 out of 9 times, but only named it 1 time.
Same brand, same query, but different engines and different results. This is important for measurement: a brand can appear “visible” in one engine’s data while being completely anonymous in another. Overall AI visibility metrics mask this divergence.
5. In-text brand mention rates vary by geography
Taking into account the LLM, the country-level differences in mention rates are significant:
- India and Sweden have the highest mention rates (50%), suggesting more conversational or brand-focused query patterns in these markets.
- Italy, Brazil and the Netherlands have the lowest mention rates (18-22%), with very high citation rates (82-94%).
- The UK and Canada are average but above the global average.
*Note: The dataset uses localized prompts confirmed by SEMrush, so the language is not a confusion.
Being cited and being named are not the same thing and require a different approach
From this analysis, four takeaways stood out to me the most for brands and their content strategies:
1. Being cited means that an AI is based on your content. To be mentioned means to name yourself. We don’t yet know enough about the implications of mentions and citations, but we can say with certainty that there is a system that decides when you are cited or mentioned.
2. Your strategy should be LLM specific. A Gemini-first strategy is different from a ChatGPT-first strategy. Any AI visibility report that lumps all LLMs together is misleading.
3. Comparative content helps name brands. The informative content feeds the machine anonymously. If the goal is brand mentions, not just quotes, focus your content strategy on rating, comparison, and recommendation.
4. Prompt format is important. Brands need to map not only the topics they want to appear in, but specifically the wording patterns that produce mentions versus ghost quotes. Short conversational queries and long structured queries behave like different products.
Methodology
Data source: Semrush AI Toolkit: 3,981 domain appearances in 115 prompts, 14 countries and four AI search engines (ChatGPT, Google AI Overviews, Gemini, Google).
Each row in the dataset represents a domain that appeared in an AI response. Each appearance is labeled as “cited” (the domain appears as the source link) and/or “mentioned” (the brand name appears in the response text). The gap between these two states is what this analysis calls a ghost quote: The AI used your content but did not say your name.
Featured image: Roman Samborskyi/Shutterstock; Paulo Bobita/Search Engine Journal
