Mention

USP metric

Buyers stopped Googling.
They started asking.

Every “best X for Y” question now lands inside a chat window you’ll never see. Recommendation Rate is the number for that window — the share of relevant buyer prompts where AI actually recommends your brand. Not mentions. Recommends.

Buyer questionWhen buyers ask AI who to use — are we on the shortlist?
All featuresPublished methodology

Overview

Visibility ScorecompositeMention · position · citation
  • Recommendation RateUSP
  • Mention RateKPI
  • First-choice RateKPI
  • Mention → Recommend gapKPI
  • Share of VoiceKPI
  • Average PositionKPI
  • Citation ShareKPI
Product layout of Overview — Recommendation Rate leads the strip. Values fill from stored engine runs on your scan, never invented here.

The metric

There’s a number for this.

We call it Recommendation Rate. In MentionX that is the share of engine runs where your brand appears in a ranked shortlist. A name-drop in a paragraph does not count. A numbered or listed option does.

Scores are computed in application code from stored rankings — not invented inside an LLM prompt. Pair it with First-choice Rate (#1) and the Mention → Recommend gap (named, but not chosen).

Mention Rate

Named

Share of engine runs where your brand is said at all — including as an aside, a caveat, or “I’ve heard of them.”

Passing mention

“Some teams also look at Your Brand, though most start with the tools below.”

Recommendation Rate

Chosen

Share of runs where you land on the ranked list the buyer is meant to pick from.

Shortlist

1. Competitor A · 2. Your Brand · 3. Competitor B

First-choice Rate

Position #1 on that shortlist — the engine’s opening pick.

Mention → Recommend gap

Known but not chosen. The work lives in this hole.

The snub

You can be discussed and still lose the list.

A typical buyer prompt does not ask “have you heard of us?” It asks who to use. If your brand is missing from that answer, there is no deal to enter the funnel — mentions elsewhere do not rescue it.

best project management tool for a 10-person startup

For a small, fast-moving team, here are the strongest options:

  • Linear
  • ClickUp
  • Asana
  • Monday.com

✕ Not mentioned: your brand

Illustration of a shortlist-style answer — the shape of the snub, not a live MentionX scan.

How it’s built

The formula is simple. The evidence is not optional.

  1. 01

    Any company URL

    Discovery from the public site — no hardcoded industry.

  2. 02

    Buyer prompts

    Category questions engines actually get asked.

  3. 03

    Stored answers

    Raw engine text kept before any score is calculated.

  4. 04

    Scored in code

    Recommendation Rate from ranked-list position — not an LLM vibe.

times you were recommended

times the question was asked

× 100 = your rate

What counts

A run counts when the stored ranking has a position for your brand — a ranked list, numbered options, or an explicit shortlist. That is recommendation_rate = shortlisted runs ÷ total engine runs.

What does not

Any name-drop still raises Mention Rate. First-choice Rate uses position #1 only. The gap is max(0, mention rate − recommendation rate). Every scan records a scorer version so reruns stay comparable.

Read the published methodology →

Same category

Same buyers. Wildly different fate.

Two brands can sit in the same category, face the same buyer questions, and still live in different answers. One is built for AI discovery. The other did not know it was being asked about at all.

Illustration of how far two brands can diverge on the same prompt set — not a published average, and not a specific customer result.

The leak

Not a metric. A leak.

A low Recommendation Rate is a pipeline leak with no dashboard. Every unanswered prompt is a deal that never entered your funnel. Mentions can look busy. The shortlist is where demand actually moves.

Prompts asked
Mentioned
Recommended
First choice

The drop from mentioned → recommended is the Mention → Recommend gap — the leak.

Conceptual funnel of the four rates. Widths are diagrammatic, not a customer result.

That is why MentionX leads Overview with Recommendation Rate — then the gap, then First-choice. Visibility without a shortlist is a story you cannot take to the board.

Around the rate

Three things decide how you show up when you do get named.

  1. Sentiment

    How AI talks about you when it does bring you up — strengths, caveats, and the frame it repeats.

    Brand sentiment →

  2. Sources

    What it cites. Citation gaps are often why you are known and still not chosen.

    Citation intel →

  3. Positioning

    How engines pair you with alternatives — and which story they tell instead of yours.

    Positioning →

Weekly monitoring reruns the same prompt library so you can watch Recommendation Rate move. MentionX does not invent query volumes. It re-asks the questions you actually care about. Weekly monitoring →

Why it matters

  • Recommendation Rate on every plan Overview
  • First-choice Rate when you win #1
  • Mention → Recommend gap for named-but-not-chosen

Outcomes

  • Stop celebrating vanity mentions that never convert to recommendations
  • Prioritize prompts where you are mentioned but not shortlisted
  • Report a board-ready recommendation story — not a vague visibility vibe

Platform

How MentionX delivers

Shortlist detection

Scored from stored rankings in engine answers — computed in application code.

Gap narrative

Mention Rate minus Recommendation Rate surfaces opportunities GEO Content Agent can draft against.

Cross-engine view

See Recommendation Rate by engine — strongest and weakest chips on Overview.

Close the gap with GEO Content Agent (Nova+) — briefs and AEO / GEO / SEO drafts from your Actions board, grounded in the same stored evidence. Pair with Mentionie and Boardroom when the story has to leave the dashboard.

We just shipped this

Recommendation Rate is live inside MentionX

See where you stand across the AI answers that matter to your category. Any legitimate company URL. Stored engine answers before scores.

How MentionX works

← features