How to Read GEO Metrics: Visibility Score, Mention Rate and Share of Voice

The first time you open a GEO dashboard you find four numbers: Visibility Score, mention rate, Share of Voice and average rank. Each answers a different question about how ChatGPT and Gemini talk about your brand. This guide explains what each one measures, how it's calculated, how to read them together and what to do when they aren't where you want them.

The four metrics at a glance

• Mention rate: do I show up? The percentage of monitored prompts in which your brand appears in the answer.
• Average rank: in what position? Your brand's average position, calculated only over the answers where it appears.
• Visibility Score: how well do I show up overall? It combines the two above into a number from 0 to 100.
• Share of Voice: how much space do I take against competitors? Your mentions over all brand mentions in those answers.

What is Visibility Score and how is it calculated?

For each AI answer, Visibility Score is 100 ÷ the position where your brand appears: first, 100; second, 50; third, 33; fifth, 20. If you aren't mentioned, it's 0. That isn't a penalty, it's a signal: a non-mention is the worst outcome, because the buyer gets a competitor's answer and never considers yours. Scoring 0 makes gaps visible on the dashboard instead of hiding them inside an average.

The overall score you see on the dashboard is the average of all those values, with non-mentions counting as 0. Put differently: mention rate × (100 ÷ average rank). A brand mentioned in 70% of prompts with an average rank of 2 scores 35, not 50, because in the other 30% it doesn't appear. That's why a brand that ranks 5th in half its prompts and is absent from the other half averages 10, not 20.

How to read a Visibility Score

There are no universal benchmarks: they vary by category, how many brands compete and how long yours has been in the market. What you can read is what each range means, straight from the formula:

• 0: you weren't mentioned in any monitored prompt. For those queries, you're invisible.
• 1 to 20: you appeared rarely, or always near the bottom of the list. Marginal visibility.
• 21 to 49: you appear regularly, typically in 3rd or 4th place. Competitive, but not the first choice.
• 50 to 79: you appear often and near the top. Solid presence.
• 80 to 100: engines recommend you first in nearly all tracked prompts.

Mention rate: the first metric worth moving

It's the percentage of prompts you appeared in, regardless of position. If you monitor 20 prompts and appear in 14, your mention rate is 70%. It's usually the most important one at the start: a brand that shows up third in 18 of 20 queries is in a much better spot than one that shows up first in 4. Before fighting for first place, close the prompts where you don't appear at all. It's also usually the fastest to move, because it responds to the citable content and third-party mentions you keep adding.

Average rank: where you appear when you are mentioned

An average rank of 1.3 means you're almost always the first recommendation; 3.7 means you sit near the end of the list. Because it only counts prompts where you appeared, it doesn't penalize non-mentions (Visibility Score does that). Start working on it once your mention rate is high: if you're mentioned often but always low, the problem is the depth of your authority, not your coverage.

Share of Voice: your place against competitors

Share of Voice (SoV) in AI is the percentage of all brand mentions, yours and your competitors', that belong to your brand. If four brands add up to 100 mentions across 25 queries and yours has 35, your SoV is 35%. Unlike advertising SoV, you can't buy it with media spend: it measures how often an engine chooses to associate your brand with a category or a problem. It can rise even when your Visibility Score hasn't moved, if a competitor loses mentions, so look at both.

How to read your dashboard in 5 steps

1. 01
   Start with mention rate: Filter by prompt category. Every category with a low rate is a content gap: that's where competitors take all the AI-generated referrals.
2. 02
   Sort prompts by Visibility Score: The lowest ones, at 0, are your priorities. Those between 33 and 50 are position-improvement targets: they need authority signals such as third-party citations or comparison content.
3. 03
   Track Share of Voice over time: A rising SoV with the same number of mentions means your position is improving. A falling one with a stable position means competitors are showing up in new queries you don't cover.
4. 04
   Cross the data by engine: ChatGPT and Gemini don't weigh sources the same way. A high score in one and a low one in the other shows where coverage is missing. Measure them separately.
5. 05
   Set a drop alert: A sharp fall on a high-intent prompt (pricing, comparisons, "best for") deserves an immediate review: engine answers change when models update or when a competitor publishes new content.

From metrics to action: three moves

1. 01
   Publish pages that answer directly: For each prompt with a 0% mention rate, a page that answers exactly that question in the first lines, with data and sources, and names your brand as the solution.
2. 02
   Earn mentions in the sources the engine already cites: Third-party mentions weigh more than what you publish on your own domain. The Authority Sources panel shows which domains each engine cites in your category.
3. 03
   Publish your own data: Original numbers get cited. If your brand publishes research, even from your own platform, other sites reference it with attribution and every citation strengthens your authority.