Metrics

Rollups are per workspace, prompt, engine and day. The dashboard averages per-prompt visibility so a prompt with more engines does not dominate.

Visibility

|{r ∈ R : brand mentioned}| / |R| × 100

R is successful runs (status = ok) for the prompt, engine and day window. Shown as 42% · 8 of 20. A bare percentage is not used.

Average position

mean(position) over mentioned runs

1 is the first brand in the answer. Runs with no mention are excluded, not treated as last.

Share of voice

brand mentions / mentions of you and tracked competitors in R

Untracked brands in the answer do not enter the denominator.

Citation share

runs citing any URL on your domains / |R|

A mention without a URL on your domain does not count as a citation.

Times cited

count of runs in the window where the item appears

Used on Sources and Top recommendations.

Citations (access)

verified user-fetch hits, unless you expand the purpose filter

Default is live answers only. Search index, training and unverified hits are opt-in.

Percent change

current range vs the equal range before it

7d compares to the previous 7 days. The period is always named.