Brand Mention Share
The percentage of AI answers that mention your brand by name in the response text — not as a linked source. An NER-based AEO KPI.
What is Brand Mention Share?
Brand Mention Share is the percentage of AI answer responses that mention your brand directly in the response text — by name, not via a linked source card. It is one of the four core AEO KPIs and captures the "authority exposure" users absorb just by reading an answer, regardless of whether they ever click through.
While Citation Share measures share at the source-card level, Brand Mention Share measures share at the in-body text level. It is especially important on platforms like ChatGPT, where source cards are rarely shown — making source-based metrics blind to that surface.
How it is measured
Brand Mention Share is typically computed via NER (Named Entity Recognition) extraction over response bodies.
| Step | What to do |
|---|---|
| 1. Run query bundle | 10–20 core queries × 5–10 repeats |
| 2. NER on each response | Extract ORGANIZATION / BRAND entities from response text |
| 3. Normalize aliases | Unify "Anthropic" / "Anthropic Inc." / non-English variants |
| 4. Compute share | (mentions of our brand ÷ all brand mentions) × 100 |
Spelling consistency drives reliability. Use a unified brand spelling — local-language form, English form, and domain — across every page so NER reliably attributes mentions to your entity.
Versus Answer Inclusion Rate
| KPI | What it captures | Caveat |
|---|---|---|
| Answer Inclusion Rate | Does the response include our domain (linked or text)? | Depends on source-card surfaces — under-counts on ChatGPT |
| Brand Mention Share | Does the response body name our brand in text? | Depends on NER accuracy and alias normalization |
Reading both together isolates the case where a brand appears in body text but never as a source card — an authority signal coexisting with poor extraction structure. That pattern means rebuilding the page for direct answer extraction, not building more authority.