Metrics Glossary

A composite of how often your brand is mentioned across tracked prompts, where it appears in each answer, and how consistently that holds across models.

Read it as a trend and against competitors. The absolute value is not comparable between categories, because how many brands a model names varies by topic. See Reading Your Score.

The count of tracked answers in which your brand name appears. Binary per answer — either the model named you or it did not.

If your brand name collides with a common word or another company, check the raw answers to confirm the mentions are really yours.

Where your brand sits inside the answer. Being named first reads as a recommendation; appearing near the end of a long list rarely survives the reader's attention.

Position explains cases where mentions are high and outcomes are not.

Your portion of total brand attention across tracked prompts, relative to the competitors in your project. Combines frequency, position, and sentiment. See Share of Voice.

How the answer characterises the brand — favourable, neutral, or qualified. "The most complete option" and "a cheaper alternative with fewer integrations" are both mentions with very different sales impact.

How often a given source is cited across your tracked answers. The ranking metric for prioritising source work: high-citation sources carry the category. See Citations and Sources.

The category of a cited source — community thread, review site, comparison page, official documentation, or owned content. Reveals which kinds of sources your category runs on and which ones you are missing from.

How much of the buyer journey your tracked prompts span. Thin coverage produces a confident score built on one narrow slice of intent. See Prompts.

The stored record of each response: model, full answer text, brand mentions, citations, and prompt context. The audit trail behind every other number here.