Metrics glossary

Each term as the product uses it, with its operational definition. Exact parameters and normative formulas live in the methodology.

Presence rate: Your appearances divided by ALL the responses in the period, including those that mention nobody. It is the honest metric of absolute visibility, which is why we never show Share of Voice without it alongside.
Share of Voice: Your appearances divided by the appearances of all the brands we track. Each response contributes at most one appearance per brand, even if the text repeats the name eight times. If the denominator is zero, the figure is “no data”, never zero.
Citation rate: Responses where one of your domains is cited, divided by the responses from engines capable of citing. Engines that structurally cannot cite are left OUT of the denominator: averaging them in would produce a meaningless figure.
Position: The midpoint of the text where you first appear, calculated only over the responses where you appear. That is why it is never shown without the presence rate: if you lose visibility, your average position can improve simply because only the good responses remain.
Sentiment: Classification of the tone with which a response treats the brand when it appears. It is only computed over responses with an appearance, and on plans that include it.
Visibility score: A 0 to 100 summary with published weights: presence 40, share of voice 25, position 15, citation 10 and sentiment 10. The same formula in the scanner and the dashboard, with its version stored next to each report.
Denominator: The number every percentage comes from, always visible. A metric without enough sample is declared “insufficient data”; it is never painted as zero.
Appearance: A response counts as an appearance of a brand if it mentions it by canonical name, alias or domain. Each response contributes at most one appearance per brand, even if it repeats the name eight times.
Sweep: The run of the prompt catalogue on your plan’s engines, with repetitions. Its cadence depends on the plan: manual on the trial, weekly, weekly with a daily core, or daily.
Repetitions: How many times the same prompt is run on the same engine within a sweep. Engines are not deterministic: repeating stabilises the sample.
Window: The fixed period metrics are aggregated over: 30 days with daily detail or 90 days with weekly detail. Fixed on purpose, so figures stay comparable.
Wilson interval: The confidence interval used to decide whether a change is real. An alert only fires if the 90% intervals of two windows do not overlap.
Engine fidelity: How close what we measure is to what a user sees. Chat engines are queried through their official APIs with web search enabled; Google AI Overviews are observed on the search page through an external provider. Each engine's condition is declared in every report.

Every metric on this page has its formula and parameters published. Read the full methodology