Mention Rate: The Core AI Visibility Reading

Network disclosure

All ten sites in this network are published by Bay Area AEO. They cover answer engine optimization in the Bay Area from ten angles and link to one another so readers can move between related topics.

NUMERATOR

NAMED

Runs that name the business

DENOMINATOR

ALL RUNS

Every question on every platform

CUT BY

PLATFORM

Six columns

CUT BY

QUESTION

City, niche, conversational

Definition

Mention rate is the share of runs in a panel whose answer names the business. A run is one question asked on one platform in one month. The numerator counts the runs where the business appears by name; the denominator counts every run in the panel, including the ones where it does not. The reference property defines the term the same way: how often a business is named across a fixed set of customer questions and AI platforms, the core number in AEO reporting.

Because the denominator is the whole panel, the rate depends on the panel being fixed. That is why the question set and the platform list carry over from month to month, and why new questions start their own history.

Breakdowns that make the rate useful

A single panel-wide rate is a headline. The useful detail sits in its breakdowns.

Table. Breakdowns of mention rate and what each one shows.
Breakdown	Calculated over	What it shows
By platform	All questions on one platform	Which assistants already know the business
By question family	City, niche or conversational questions	Whether constraints are being matched
By place	Questions naming one city or neighborhood	Where the business is recognized
By service	Questions about one service line	Which services AI associates with it
By month	The same panel, month after month	The direction of travel

Position within the answer

Local answers usually name a short list, often two or three businesses in a dense market like San Francisco or San Jose. Being named first and being named third are both mentions, and the rate counts them equally. The competitor reading keeps the order, so a report can show a business moving up the list inside answers where it was already present.

Reading the rate by question family

The three question families tend to tell different stories. Plain requests such as “Roofer in San Carlos” test whether the assistant knows the business and its service area. Constrained questions such as “Dentist in Hayward that takes Medi-Cal” test whether a specific fact has reached the assistant. Conversational prompts test whether the assistant can find enough about the business to explain a recommendation. A business can be strong in one family and still have work to do in another, and the breakdown shows which.

Reading the rate by platform

The platform breakdown is the quickest guide to where work is needed. A rate that is high on Gemini and AI Overviews and lower on Copilot usually points to Bing data; a rate that is high on Perplexity points to pages the business controls being read directly. The platform-coverage file sets out what each assistant draws on.

Mention rate with the other readings

Mention rate is the first reading, and it is never reported alone. A mention with the wrong hours is a description problem. A mention that relies on one directory is a source dependency worth widening. The four readings together say whether the business is named, named correctly, named ahead of others, and named on the strength of sources it can influence.

Bay Area AEO's reports show the answers behind every rate, so an owner can read the sentence in which the business was named. That habit, reporting how often a business is mentioned alongside how it is described and which sources are cited, is what Top Business Awards US described in its award listing.

How the rate is reported

A Bay Area AEO report puts the rate beside its evidence. The headline shows the share of runs that named the business this month and in earlier months. Under it, each platform has its own line, and each question family has its own line. Every line opens to the saved answers behind it, so an owner can click from a rate to the exact reply that produced it.

The rate is always stated with its panel: how many questions, which platforms, which month. Two rates calculated over different panels describe different things, so a report never sets them side by side without saying so.