What is brand visibility — and where is it measured now?

The definition has not moved: brand visibility is how often, and how prominently, your brand turns up where the people who might buy from you are paying attention. Search results, social feeds, press, review sites, shelf space. What has moved is where that attention goes — and one of the new destinations does not produce the numbers the old definition was built on.

If you are here because someone in a meeting used "visibility" and "awareness" as though they were the same thing, they are not — the difference decides which of the two you can measure at all.

The units stopped working

Every classic brand visibility metric is a count of exposures. Impressions. Rankings. Reach. Mentions. They work because the surface being measured is a list of things, and you can count your position on a list.

An AI answer is not a list. Ask an assistant "where should I buy X in the UK" and you get a paragraph naming three or four businesses. There is no position five. There is no second page. You are named or you are not, and there is nothing in between to optimise toward.

That is the first thing to absorb: on this surface, visibility is closer to binary per answer, and only becomes a rate when you ask the same question many times. Which is also why a single check tells you almost nothing — assistants are non-deterministic, so asking once and seeing your name is the same evidence as one coin landing heads.

The measurement blind spot

The second thing is worse for anyone running marketing off a dashboard: this visibility is largely invisible to your analytics.

When an assistant recommends you in prose without a link, and the reader types your name into a browser a day later, that arrives in your reporting as direct traffic — indistinguishable from someone who already knew you. When it does link, referral attribution from AI surfaces is patchy and inconsistent between providers. So the honest position is that you cannot infer your AI visibility from your traffic. You have to go and measure the answers themselves.

We see the size of that gap in our own readings. A brand can be named in around one in ten answers about its category — a genuinely meaningful presence — while nothing in its analytics reflects that at all.

The three questions it now has to answer

Any serious attempt to measure brand visibility today has to answer three separate questions, and most tools blur them into one figure:

Am I present? Do assistants name you when asked a buyer's question — not a question about you. "Tell me about Acme Cigars" always produces a description; that is not visibility, that is the model doing as it is told. "Best online cigar shop UK" is the test.

How often? Presence is a rate, and a rate needs a sample. One answer is one sample of a random process. Hundreds of answers, across several assistants, give you a percentage — and that percentage still carries a range around it.

Against whom? A rate on its own cannot tell you whether you are doing well. Being named in 8% of answers is a failure or a category lead depending entirely on what everyone else scored. That comparison is share of voice, and it needs a field somebody has actually measured.

What has not changed

The lever. Brand visibility has always been won mostly off your own property — by being covered, reviewed, listed and discussed by other people — and that is still true here, more starkly than before. An assistant's sense of who belongs in an answer comes from the web's discussion of your category, not from your homepage. A site rebuild makes you readable. Other people's pages make you recommended.

What that means in practice is the improvement guide. What it means for the number is the scoring guide.

You can see the readable half of your own position in a free scan, and the measured half — how often assistants actually name you, against your category — is what the index publishes.