A mirror metric is a brand-visibility number reported without the signals that produced it, for example an AI visibility score with no view of the owned, rented, or borrowed real estate beneath it. It reflects the current state but offers no lever to change it. If you cannot tell which surface to move to shift the number, the number is a mirror, not insight, and mirrors do not grow brands. Citable uses the term to separate scores that merely describe from diagnostics that prescribe: a usable visibility reading must arrive attached to the underlying signals a team can actually act on.
Why it matters for AI search
AI visibility scores are everywhere now, and most are mirror metrics: they tell you where you stand without telling you how to move. The distinction matters because a founder cannot act on a reflection. A useful reading arrives attached to the owned, rented, and borrowed signals that produced it, so the number points at a specific next action rather than just handing down a verdict.