Brand Mention Rate

A prompt counts once if your brand is named anywhere in the answer, no matter how many times it appears. That keeps the metric a measure of presence rather than of repetition, so a single clear mention and five passing ones score the same.

Because presence comes first. Sentiment, position and citations only describe answers you are already in — a low mention rate exposes the more fundamental problem, which is that the model does not reach for you at all.

It has to reflect the questions buyers in your category actually ask, in their words, and the identical set has to run against every model you track. A set written in your own product language will flatter you and measure nothing useful.

Group the prompts where you are never named. Each cluster points at either missing content or missing third-party sources, and closing one cluster at a time is what moves the number.

Not on its own. Being named in every answer at the very end is weaker than being named in half of them first, which is why mention rate is read alongside position rather than instead of it.

There is no published price. The value of one mention is consideration lift multiplied by your close rate multiplied by what a customer is worth, and all three come from your own funnel. Any single dollar figure quoted publicly was modelled from someone else's business.

Nothing, because there is nothing to buy. No model sells placement inside a generated answer, so there is no auction and no rate card. What has a price is the work that earns the mention, and dividing that spend by the mentions it produced gives a cost per mention for your brand.

An AI answer is generated for one request and never published, so there is no impression record to price against. No marketplace sells the position, so no clearing price forms. And a mention carrying no link produces no referrer, so the visit it causes is recorded as branded or direct traffic instead.

Rarely. Being named first in a sentence of its own differs from ninth in a list, a cited mention can be attributed while a bare one cannot, and a mention inside a buying prompt sits closer to the decision than one inside a definition. Mention rate counts mentions; it does not weigh them.