Association vs market prominence, explained
Association is what AI ties to your brand. Market prominence is which brands own an attribute in your category. Why the two diverge and what to do about the gap.
- Association asks what the model ties to your brand; market prominence asks which brands own an attribute in the category.
- A brand can score high on association and still lose most category answers to an incumbent.
- When products look identical in an answer, the model reaches for the brand it has seen most, which is usually the incumbent.
- Pick attributes you can credibly own, work the gap between the two scores, and read the intervals rather than the point estimates.
Two questions with two different denominators
Association asks what the model ties to your brand: what is X known for. Market prominence asks which brands own an attribute in the category: who is known for this. Same attribute, opposite direction, and the gap between them is where category battles are actually lost.
Association is built from answers about your brand, so it moves when your own reputation moves. Prominence is built from answers about the category, so it moves when competitors out-shout you for an attribute. Both are scored 0 to 100 with a 95% confidence interval, which is the range the true value most likely falls in, and they are deliberately never plotted on one axis because they measure different populations.
Why a strong association still loses the category
You can be tightly linked to an attribute and still be absent from the answers that list brands for it, because being known for something and being the first name recalled for it are separate feats.
Read the pair that way and the diagnosis changes. Low association with decent prominence means the market knows you for the attribute but your own answers undersell it. High association with low prominence means the attribute is yours among insiders and invisible everywhere else, which is a distribution problem rather than a positioning problem.
Why incumbents win when products look identical
When everything in an answer looks the same, the model reaches for the brand it has seen most, and that is almost always the incumbent. Brand becomes the tiebreaker precisely when the model has no other signal left.
A 2026 study tested this directly across three commercial LLMs and found well-known brands getting recommended 100% of the time when all products shared the same specifications, with that dominance disappearing once a competitor held even a small rating advantage of less than a tenth of a star (Incumbent Advantage, arXiv:2606.17443). The tiebreak is brittle: give the answer any other signal, a rating, a spec, a dated review, and the brand advantage stops deciding on its own.
The associations themselves lean the same way. An EMNLP 2024 analysis found models disproportionately tying global brands to positive attributes across four brand categories (Global is Good, Local is Bad?, arXiv:2406.13997). Neither paper tells you how to win; together they tell you what you are competing against: a default preference for the familiar, broken only by a concrete differentiator the answer can quote.
What to do about the gap
Pick attributes you can credibly own, then work the distance between your two scores rather than chasing the highest number on the board.
- Choose attributes with proof behind them: a spec, a dated comparison, a number a source can cite. The tiebreak research above says concrete signals are what displace the incumbent, so an attribute nobody can substantiate is a bet on fame you do not have yet.
- Work the gap: where association outruns prominence, the job is getting named in category answers (coverage, comparison content, third-party sources). Where prominence outruns association, the job is making your own answers say it about you.
- Read the intervals, not the point estimates. Two scores of 61 and 58 mean nothing if their ranges overlap; a gap only becomes real when the intervals separate, and week-over-week moves only count when both weeks clear the sample minimum.
Where this lives in advisible
The Market tab holds both scores per attribute, side by side but never merged, with a biggest-gap card for the attributes where the two disagree most.
The gap card only appears when the intervals on the two scores no longer overlap, so what it surfaces is a difference you can defend rather than a rounding artefact. Alongside it you get a brand shape view, a comparison heat map against tracked competitors, and term-level drill-downs for the wording behind each attribute. Positive attributes only: anything negative the answers repeat about you lands in the Objections tab with its category and sources, where somebody can actually act on it. The wider frame for these numbers is brand perception in AI answers.
Common questions
What is association in AI visibility?
The strength of the link between your brand and a positive attribute in the answers that mention you: what the model says you are known for, scored 0 to 100.
What is market prominence?
How strongly an attribute is owned in the category as a whole: when an answer lists brands for easy setup, who gets named. Also scored 0 to 100, measured across category answers rather than your own.
Why are both scores needed?
Association alone says you have a reputation; prominence says whether that reputation wins the category. A brand can be strongly linked to an attribute and still never be the name the answer leads with.
Which score should I try to raise?
Prominence, but only on attributes you can credibly own. Prominence on an attribute your product does not deliver is a claim the first trial user will knock down, and the objections tab will show you doing it.
Brand perception in AI answers: what AI says about you
Visibility tells you whether AI answers mention you. Perception tells you what they say: the attributes, objections and factual claims behind each mention.
How ChatGPT, Perplexity and Gemini choose which brands to mention
No engine publishes how it picks brands. The inputs are observable: what the sources say, query fit, consistency and prominence. What you can influence.
How to get cited by ChatGPT and Perplexity
How engines choose sources at answer time, how to write a page worth quoting, and a 30-day procedure for earning citations in ChatGPT and Perplexity answers.
What is AI search visibility?
AI search visibility is how often your brand shows up when buyers ask ChatGPT, Perplexity or Gemini about your category. What it means and how to measure it.
Track this for your own brand
advisible runs your prompts on ChatGPT, Perplexity, Gemini, Grok and Claude every day, and every number shows how many answers it is based on.