Glossary · Measurement
Response Consistency
The degree to which repeated runs of the same prompt, on the same engine and within the same time window, return the same brand mentions and positioning.
How it works
Response consistency is tested by re-running an identical prompt several times in a short window and checking whether the same brands appear, in roughly the same order and tone, each time. High consistency means a single run is a reliable stand-in for the pattern; low consistency means it isn't.
Why it matters
Generative models produce different wording — and sometimes different brand selections — on every run, even with identical inputs and no change in the underlying data. Without checking consistency first, a team can easily mistake ordinary run-to-run noise for a real shift in visibility.
See more terms in the full glossary, or read the AI Visibility 101 guide for the full picture.