How to Calculate Brand Mention Percentage vs. Competitor Citation Rate Across LLMs? A Tales from the Query Post

This article is part of a series for my blog called “Tales from the Query,” where I look at a random query in GSC or BWT and write a blog post about it.
Today’s query is: “how to calculate brand mention percentage vs competitor citation rate across llms?”
While I already have an article that ranks for this query, “Evaluating Brand Retrieval and Ranking in Large Language Model Recommendations (Hamsterdam Research)”:

That article doesn’t specifically address the question at hand.
I also think this query has some good inherent lessons to learn, especially for people who are newer to AI search optimization.
So let’s get to it.
First things first …
The query asks about “brand mention percentage” vs. “competitor citation rate.” Those are different metrics.
A mention is where a brand is listed in an AI-generated answer. A citation is where a URL from a website is both used as a source and linked within the answer.
You can be mentioned without being cited, and vice versa.
Thus, comparing your brand mention percentage against your competitors’ citation rate doesn’t necessarily make sense. Just because two AI visibility metrics are expressed as percentages doesn’t mean they’re directly comparable.
A more meaningful comparison would be your brand mention rate vs. your competitors’ brand mention rate or your citation rate vs. your competitors’ citation rate, so it’s apples to apples.
Next let’s talk about what a “brand mention percentage” is.
That’s the percentage of prompts in which your brand is mentioned. (We also tend to call it a “rate.”) But how can you measure that? One way I can think of — short of using ChatGPT’s Business Hub, which I haven’t gotten access to yet, also this counts mentions where ChatGPT references the business’s website, and simply naming the brand doesn’t count — is with prompt-tracking.
The limitations of prompt-tracking are that you’re typically measuring synthetic prompts that you’ve chosen (selection bias), not real user prompts. Also, commercial prompt trackers focus on single-turn interactions, unlike real conversations that can be multi-turn. (You could also measure prompts manually, but that would provide less representative coverage at scale than prompt tracking.)
The formula for calculating brand mention rate is responses mentioning your brand divided by total responses analyzed times 100. For example, if you analyze 200 AI-generated responses and your brand appears in 50 of them, your brand mention rate is 25%.
Finally, let’s talk about “competitor citation rate.”
A citation rate is the percentage of prompts that a particular brand’s website is cited in. The aggregated citation rate of all of your identified competitors would be the competitor citation rate. (Remember also that multiple competitors can be cited in a single answer, so in this case, you’re counting unique responses citing any tracked competitor, rather than summing the individual rates.)
For your own brand, you could use a tool like Bing Webmaster Tools AI Performance Report and look at Citations & Pages, but that’s limited to mostly Copilot and doesn’t show your competitors. For competitor citation data, you could look at Share of Authority in Microsoft Clarity, but again that’s limited to mostly Copilot and shows relative citation visibility, not the exact competitor citation rate. Bing Webmaster Tools also has a Citation Share preview in the Grounding Queries section, but it measures your site’s share of citations for a single grounding query and doesn’t reveal which competing domains received the other citations. For broader data across LLMs, you’d need to use a prompt-tracking tool, but again, it carries the same limitations mentioned earlier.
The formula for calculating competitor citation rate is responses citing a competitor’s domain divided by total responses analyzed times 100. For example, if a competitor’s website is cited in 30 out of 200 responses, its citation rate is 15%.
Also, we need to address the “across LLMs” portion of the query.
Because of their different operations behind the scenes, LLMs, or more accurately, AI assistants or search experiences, will produce different mention percentages and citation rates. This means you’ll need to set up multiple AI assistants for your prompt-tracking. The downside is this can involve additional costs, so my advice is to select the main AI assistants your audience uses, and focus on those. My big four would include ChatGPT, Gemini, AI Overviews, and AI Mode. Others to consider include Copilot, Claude, and Perplexity. (For data consistency, you’ll also want to use the same prompt set, geography, language, and measurement period across platforms.)
You’ll also need to consider internal updates to AI assistants, like when models or backend search infrastructure update and how that impacts mention and citation rates. There can also be differences between logged-in and logged-out users. In my experience, prompt-tracking doesn’t qualify these nuances, which is another limitation. That’s why I’d combine prompt-tracking data with manual runs of multi-turn prompts using different logged-in statuses and model configurations.
Key takeaways
- Mentions and citations measure different outcomes: Being recommended isn’t the same as being used as a cited source.
- Percentages depend on how you define the denominator: A metric without a clear methodology has limited value.
- Prompt tracking measures visibility within a selected test environment: It shouldn’t be mistaken for a complete picture of real user behavior.
- Cross-platform comparisons require consistent testing: Otherwise, you’re potentially comparing different things.
- Visibility doesn’t necessarily equal business impact: A brand appearing frequently in AI responses doesn’t establish that those appearances drive awareness, qualified traffic, or conversions.
Outro
Thanks for checking out this Tales from the Query post addressing “how to calculate brand mention percentage vs competitor citation rate across llms?”
Stay tuned for more Tales from the Query posts, or check out past ones below.
And if you need a hand with your SEO/GEO strategies, feel free to contact me. I’m an independent consultant helping brands and agencies.
Until next time, enjoy the vibes:
Thanks for reading. Happy optimizing!
Related posts
What Types of Media Coverage Are Most Valuable for GEO? A Tales from the Query Post
What Types of Media Coverage Are Most Valuable for GEO? A Tales from the Query Post This article is part of a series for my…
How to Write a Blog Post for AI Search Visibility and Citations: A Tales from the Query Post
How to Write a Blog Post for AI Search Visibility and Citations: A Tales from the Query Post This article is part of a series…
How Does Comparison Content Affect AEO Rankings? A Tales from the Query Post
How Does Comparison Content Affect AEO Rankings? A Tales from the Query Post This article is part of a series for my blog called “Tales…
Leave a Reply