Metrics Explained
Understanding how XAUDIENT measures your brand's AI visibility
Overview
XAUDIENT tracks multiple dimensions of brand visibility in AI responses. Each metric captures a different aspect of how AI systems perceive, position, and recommend your brand. Together, they form a comprehensive picture of your Prediction Landscape.
1. Mention Rate
Definition: The percentage of relevant queries where your brand appears in AI responses.
Mention Rate = (Queries with brand mention / Total queries) × 100%
What It Tells You
Mention rate indicates your brand's coverage in the AI landscape. A high mention rate means AI systems frequently include your brand when users ask relevant questions.
Pro tip: Compare mention rates across different query categories. You may have strong coverage for "best of" queries but weak coverage for "how to" queries.
2. Position Ranking
Definition: The ordinal position (1st, 2nd, 3rd, etc.) where your brand appears when mentioned in AI responses.
Average Position = Σ(position in each response) / Number of mentions
Why Position Matters
When AI lists multiple brands, earlier positions receive more attention. Being mentioned first carries significantly more weight than being mentioned fifth.
Position Score
We convert raw positions into a normalized score (0-100):
Position Score = (1 - (position - 1) / total_brands) × 100
A score of 100 means first position; lower scores indicate later positions.
Interpretation
- • Position 1: AI recommends you first — strongest positioning
- • Position 2-3: Strong alternative — you're a top consideration
- • Position 4+: Listed option — mentioned but not prioritized
3. Context Classification
Not all mentions are equal. XAUDIENT classifies how AI frames your brand in each response:
Leader (Weight: 1.0)
AI positions your brand as the primary recommendation or market leader.
"The best option is [Brand] because..."
Strong Alternative (Weight: 0.8)
AI presents your brand as a top alternative with specific advantages.
"If you prefer X feature, [Brand] is an excellent choice..."
Listed Option (Weight: 0.6)
AI includes your brand in a list of options without special emphasis.
"Options include [Brand A], [Brand B], and [Brand C]..."
Example (Weight: 0.5)
AI uses your brand as an example to illustrate a concept.
"Companies like [Brand] have implemented this approach..."
Mentioned (Weight: 0.3)
AI references your brand without specific context or recommendation.
Cautionary (Weight: -0.3)
AI mentions your brand with warnings or negative framing.
"Be aware that [Brand] has issues with..."
4. Tone
Definition: How AI replies describe your brand, as an average from -1 (negative) to +1 (positive). English runs use a keyword indicator. Italian runs are AI-assessed and shown as Provisional.
Tone Scale
Note: Neutral Tone isn't bad. Many factual queries produce neutral responses. Watch for Tone trends over time and compare against competitors. “Not measured for this run” is different from neutral.
5. Share of AI Voice (SAV)
Definition: SAV is your brand's mentions as a share of all mentions across your confirmed competitive set.
The SAV Formula
SAV = brand mentions ÷ (brand mentions + confirmed competitors' mentions) × 100%
WSAV
WSAV (weighted share of AI voice) is the same share weighted by position in the answer and by tone.
WSAV = Σ brand (position weight × Tone weight) ÷ Σ all (position weight × Tone weight) × 100%
Organic SAV
Organic SAV covers queries that name nobody: it is a true share (every company's share of all mentions in that bucket; brand plus competitors sum to ~100%).
Against Competitors
Your confirmed competitors are counted the same way as your brand and are shown by their number of mentions. In a market of four companies (you and three competitors), 25% is an even share; above it, AI assistants name you more often than your competitors on average. WSAV is not measured where Tone is not measured, and trend deltas never cross language boundaries.
For the complete mathematical specification, see Share of AI Voice documentation.
6. Provider Breakdown
Each AI provider may perceive your brand differently. XAUDIENT tracks metrics separately for:
OpenAI (ChatGPT)
Largest user base. Training data cutoff varies by model version.
Anthropic (Claude)
Known for nuanced, thoughtful responses. Growing enterprise adoption.
Google (Gemini)
Integrated with Google Search. May have more recent information.
Perplexity (Beta)
Search-focused AI. Includes citations and sources.
Strategy tip: If you see significant variance across providers, consider provider-specific optimization strategies. A brand may lead on Claude but lag on ChatGPT.