PlatformPosition
where you land in the answer.
Visibility tells you whether AI names your brand. Position Frequency tells you the exact order you and your competitors appear in, across ChatGPT, Claude, Gemini, Perplexity, Grok, and Google AI Mode.
every spot you land, and how often.
Each row is a brand and each column is a spot in the answer, shaded by how often that brand lands there. Your brand is pinned and highlighted. Switch categories to see how the picture changes, and toggle between median and average.
Position Frequency
| Brand | |||||||||
|---|---|---|---|---|---|---|---|---|---|
Lululemon | 100% | 1.0 | 81% | 15% | 2% | 0 | 0 | 1% | 0 |
Vuori | 93% | 4.0 | 5% | 18% | 16% | 8% | 7% | 35% | 4% |
 Demo Athletic Co | 89% | 5.0 | 2% | 16% | 15% | 8% | 7% | 36% | 5% |
Alo Yoga | 87% | 4.0 | 1% | 18% | 14% | 15% | 18% | 20% | 1% |
Nike | 86% | 2.0 | 9% | 35% | 13% | 7% | 4% | 12% | 5% |
Adidas | 73% | 4.0 | 0 | 2% | 32% | 12% | 5% | 14% | 8% |
Athleta | 69% | 5.0 | 0 | 9% | 11% | 8% | 12% | 22% | 7% |
Outdoor Voices | 66% | 7.0 | 0 | 0 | 0 | 11% | 9% | 30% | 16% |
Gymshark | 49% | 6.0 | 0 | 1% | 5% | 4% | 12% | 22% | 7% |
Under Armour | 40% | 8.0 | 0 | 0 | 1% | 9% | 5% | 10% | 14% |
Fabletics | 40% | 9.0 | 0 | 0 | 0 | 0 | 4% | 23% | 13% |
Beyond Yoga | 28% | 7.5 | 0 | 0 | 0 | 1% | 4% | 11% | 13% |
Live with sample data from the public demo brand. Every Centium dashboard renders this from your own answers.
how position is measured.
Position is built from the answers you already collect. No extra prompts, no separate methodology, no added cost.
at the category level.
Centium asks ChatGPT, Claude, Gemini, Perplexity, Grok, and Google AI Mode about your category, never about your brand by name. Asking about the brand biases the answer toward it.
the order named.
Every answer is parsed into the ordered list of brands it names. The spot each brand occupies is logged with the model, the prompt, and the category that produced it.
across every slot.
Those spots are counted into a distribution per brand per category, so you see the full spread of where you land rather than one number standing in for all of it.
against the field.
Competitors are laid into the same grid from the same answers, so your distribution is read against the brands actually sharing the answer with you, not against a benchmark.
the drift you would otherwise miss.
One run gives you a reading. A run every cycle shows you whether you are moving into the set models name first or drifting toward the end of the answer. This tracks where the demo brand has landed over time, run over run, with lower being better, against the dashed line marking how long a typical answer runs in that category. Existing customers already have their full history filled in.
Position Over Time
most tools report a number. this reports the picture.
A single average is easy to put on a dashboard and it throws away the part that matters. These two brands have the same average position, and nothing else about them is the same.
First in four answers out of ten, then past the typical cutoff in most of the rest. Genuine upside, and a genuine risk of being invisible any time the answer runs short.
Reliably around position five, just inside where a typical answer stops. Almost always in the set, almost never the first name a buyer reads, and a completely different problem to solve.
inside the set, or outside it.
A typical answer names five to seven brands. If you only turn up past that line, you are likely only appearing in the overflow they add when the answer runs long. Position is how you tell those two apart, and Centium marks the line for every category you track.
rate and position move together.
Across the categories Centium tracks, the brands named most often are also the brands named earliest. Your recommendation rate and your position line up, which is what makes position a straight read on whether you lead your category. You can see it in the heatmap above.
read it for the shape.
Are you clustered near the front, spread evenly, or stacked at the tail? Which categories put you ahead of a rival and which bury you? Insights into how you appear can help guide your strategy.
questions, answered.
Position is where your brand lands inside the list an AI model gives, not whether you appear at all. When ChatGPT answers "best running shoes for flat feet" it names a handful of brands in order. Position is the spot you occupy in that order. Centium reports it as Position Frequency: how often you land in each spot across every answer, rather than collapsing it to a single average.
Centium prompts ChatGPT, Claude, Gemini, Perplexity, Grok, and Google AI Mode at the category level, records the ordered list of brands each answer names, and reports where you and your competitors land. The Position tab turns that into a heatmap per category, so you can see your typical spot at a glance and compare it against the brands you compete with.
Because an average hides the shape. A brand that lands first half the time and tenth the other half averages out to the same number as a brand that sits fifth in every single answer, and those are completely different situations. Position Frequency shows every spot you land in and how often, so you can see whether you are consistent or streaky. Average and median are both there as a toggle, but the distribution is the read.
No. An AI answer is not a ranked results page you climb, and there is no number one spot to win. What position tells you instead is whether you are inside the set of brands a model names by default. A typical answer names five to seven, so a brand that only appears past that line is being added when the answer runs long rather than recommended up front. That is a read on category leadership, not a search ranking.
Being named is the first question and your recommendation rate answers it. Position answers the second one: whether you are named inside the set models reach for first, or added at the end when the answer runs long. The two move together in practice, since the brands named most often are also the brands named earliest, so position is the clearest read on whether you actually lead your category. See Visibility for the primary metric.
No. It is built entirely from the AI answers you already collect, with no new prompts and no added cost. Current customers already have Position populated with their past data, and new runs keep it current on your normal schedule. See pricing for plans.
Use the toggle to see both. Median is the spot you land at most typically and is steadier against one unusual answer. Average is the mean across every mention. Median is the better default read, and the distribution underneath both is what actually tells the story.
Centium tracks four predictive pillars and one actual pillar. The predictive side maps what AI would say about your brand at scale: indexing (whether AI models have learned about you), visibility (whether they mention you, and where you land in the answer, covered on this page), sourcing (which websites they cite), and perception (the themes and sentiment they use). The actual side is attribution: real sessions and revenue from AI-referred traffic. See Visibility, Brand Perception, Sources, and Attribution.
see where you land
in the answer.
The full distribution, every category, every model, with your competitors in the same grid. Built from the answers you already collect, so it costs nothing extra and adds no prompts.
Starting at $99/month