GEO GLOSSARY
The percentage of sampled AI answers that actively suggest a brand as a good option, not merely mention it in passing.
A model can name a brand while comparing it unfavorably, listing it neutrally alongside alternatives, or actively steering the reader toward it. Only the last of these counts as a recommendation. Two brands can share the same mention rate and have very different recommendation rates.
Clear differentiation for a specific use case, credible third-party validation (reviews, case studies), and unambiguous factual claims the model can restate with confidence all correlate with higher recommendation rates. Vague, generic positioning gives a model little to confidently endorse.
See how these terms are used in practice in MagRank AI’s methodology.
