Recommendation rate

Recommendation rate is the percentage of usable sampled answers in which an AI system endorses your brand as a choice for the buyer's need, counted once per answer.

Source: https://contentcucumber.com/aeo-glossary/recommendation-rate/

Topic: Measuring visibility


Why it matters

An answer can name a brand as an option to avoid. A count of mentions overstates how often the brand is recommended.

Example (illustrative, not a client result)

A brand is named in 12 of 20 answers and recommended in 5 of them, which gives a 25% recommendation rate.

What to check

Label each usable answer as recommended, named without a judgment, or named with a caution, then divide the recommended answers by the usable answers and multiply by 100.

What this does not tell you

We count an answer as a recommendation when the answer presents the brand as a suitable choice for the buyer's need. Because a second reviewer can label a borderline answer differently, keep the labeling rules with the report.

Sources

Evidence type: Content Cucumber reporting definition. Sources checked on 2026-11-21

Related terms