AI Market Discovery Metric Definitions
Overview
This page provides the authoritative terminology, denominator rules, formulas, and interpretation guidance used across LLM Authority Index AI Market Discovery industry benchmarks.
Industry pages can define a metric briefly in context, but the stable technical definition should resolve here.
For the collection and qualification process, see AI Market Discovery Methodology.
Core Research Units
Source Query
The category question selected for testing.
Search-demand data may be used to prioritize source queries, but that demand signal should not be interpreted as measured AI-platform prompt volume.
Prompt-Surface Observation
One source query tested on one AI/search surface together with the resulting answer and associated collection metadata.
Example: The same source query tested in ChatGPT and Gemini produces two prompt-surface observations.
Brand-Relevant Observation
A prompt-surface observation that satisfies the benchmark's tracked-brand relevance rules and can proceed to commercial buyer-intent classification.
Incidental citations or references do not automatically qualify.
Qualified Benchmark Observation
A brand-relevant observation that also fits one of the benchmark's defined commercial buyer-intent classes and enters the applicable public analysis denominator.
The current framework includes:
- Brand Recommendation;
- Pricing & Value; and
- Multi-Brand Comparison.
Qualified Surface Breadth
The number of tested AI/search surfaces that contribute at least one qualified benchmark observation during a measurement period.
Important: Qualified surface breadth is a post-qualification result. It is not the number of AI/search surfaces tested.
Valid Recommendation Coverage
Definition
The percentage of qualified benchmark observations in which a tracked brand appears in a recommendation context that satisfies the benchmark's recommendation-validity rules.
Formula
Valid recommendation observations for the brand ÷ applicable qualified benchmark observations × 100
Interpretation
This is a recommendation-coverage metric within the qualified benchmark universe.
It should not be interpreted as:
- the percentage of all raw prompt-surface observations in which the brand is recommended;
- market share;
- sales share; or
- total AI prompt share.
Example
If a brand receives valid recommendations in 41 of 87 qualified observations:
41 ÷ 87 × 100 = 47.1%
Top-Three Recommendation Rate
Definition
The percentage of the applicable qualified benchmark denominator in which a tracked brand appears among the first three valid recommendations.
Formula
Qualified observations where the brand ranks 1–3 ÷ applicable qualified benchmark observations × 100
Interpretation
Top-three rate is a placement metric. It distinguishes prominent recommendation inclusion from lower-position appearances.
Rank-One Recommendation Rate
Definition
The percentage of the applicable qualified benchmark denominator in which a tracked brand is the first valid recommendation.
Formula
Qualified observations where the brand ranks first ÷ applicable qualified benchmark observations × 100
Interpretation
Rank-one rate is a stronger first-choice signal than simple recommendation coverage.
A brand can have high recommendation coverage but a lower rank-one rate if it appears frequently in shortlists without being selected first.
Raw Mention Presence Rate
Definition
The percentage of the applicable observation set in which a tracked brand is explicitly present, regardless of whether the mention qualifies as a valid recommendation.
Interpretation
Presence and recommendation coverage are not interchangeable.
A brand may be:
- mentioned factually;
- included in a comparison;
- cited incidentally;
- discussed neutrally; or
- recommended.
Only applicable recommendation contexts receive valid recommendation credit.
Valid Recommendation Shortlist Share
Definition
The share of qualified benchmark observations that contain a recommendation shortlist satisfying the benchmark's validity criteria.
Interpretation
This is a benchmark-level metric, not a brand-specific coverage metric.
It describes how frequently the qualified observation universe produced a valid recommendation shortlist.
Net Sentiment Score
Definition
A normalized summary of coded brand sentiment in the applicable benchmark observations.
The active public benchmark interprets:
- 1.0 as entirely positive coded sentiment;
- 0.0 as neutral coded sentiment; and
- negative coded mentions as reducing the score.
Interpretation
Net sentiment should be read alongside:
- sample size;
- brand presence;
- valid recommendation coverage; and
- recommendation placement.
A high sentiment score supported by very few observations can be less informative than a similar score supported by broad recommendation visibility.
Net sentiment is not a probability that a consumer likes the brand.
Modeled AI Authority Value
Definition
A modeled comparative measure intended to represent relative AI discovery opportunity inside the LLM Authority Index framework.
It is not:
- measured revenue;
- attributable sales;
- measured website traffic;
- advertising-equivalent value; or
- a guaranteed commercial outcome.
Because modeled value uses demand inputs and benchmark weighting, it is documented separately under Modeled AI Authority Value.
Recommendation-Shaped Answer Share
Definition
The share of the applicable qualified benchmark observations whose answer format is recommendation-oriented under the benchmark's classification rules.
Interpretation
This is a category-level structural metric. A decline does not necessarily imply that brand recommendation coverage fell; it describes the form of the qualified answer set.
Percentage-Point Movement
Changes between two rates are measured in percentage points.
Example:
- July coverage: 27.3%
- August coverage: 44.8%
- Difference: 17.5 percentage points
Publication Convention
Use:
- Prose: "increased by 17.5 percentage points"
- Tables/charts: "Up 17.5 points"
- Declines: "Down 8.0 points"
Do not write "+17.5%" when the intended meaning is a 17.5-percentage-point increase.
A move from 27.3% to 44.8% is:
- 17.5 percentage points, but
- approximately 64% relative growth.
Those are different calculations.
Counts and Denominators
Pair Rates With Counts Where Practical
A rate is easier to interpret when the underlying numerator and denominator are visible.
Example:
47.1% — 41 of 87 qualified observations
This is particularly important for lower-visibility brands.
Name the Denominator
A percentage should not imply a broader observation universe than the one used in its calculation.
If valid recommendation coverage uses qualified benchmark observations, the report should not imply that the rate applies to all raw source observations.
Use Stored Counts
Production reports should calculate metrics from exact stored numerators and denominators rather than reconstructing counts from rounded public percentages.
Metric Interpretation Framework
| Research question | Primary metric(s) |
|---|---|
| Is the brand appearing at all? | Raw mention presence rate |
| Is the brand receiving valid recommendations? | Valid recommendation coverage |
| Is the brand appearing prominently? | Top-three recommendation rate |
| Is the brand the first choice? | Rank-one recommendation rate |
| Is coded sentiment favorable? | Net sentiment score |
| How broadly are qualified observations distributed across AI/search environments? | Qualified surface breadth |
| How frequently does the benchmark produce valid recommendation shortlists? | Valid recommendation shortlist share |
| What comparative modeled opportunity does the Index assign? | Modeled AI Authority Value |
The strongest interpretation usually comes from the relationship between several metrics rather than one number in isolation.
Metric Governance
Metric names, denominator rules, and formulas should remain stable across categories and measurement periods.
If a definition changes:
- update the authoritative definition on this page;
- record the methodology/version change;
- determine whether historical values can be recalculated;
- annotate affected longitudinal reports when needed; and
- avoid presenting pre-change and post-change values as directly comparable without disclosure.
See AI Market Discovery Research Standards.