Relevance — How Well Brand Mentions Match User Intent
The Relevance metric in BrandWise: measuring how well a brand mention matches the user's query. Formula, components: eligibility, fit to intent, fit to constraints, justification quality.
What Relevance Measures
Relevance measures how well a brand mention matches the user's query and dialog context. A high score means the model mentions the brand appropriately — in response to a fitting query, accounting for constraints, and with quality justification.
This is a critically important metric: even if the brand is frequently mentioned (Visibility is high), irrelevant mentions bring no value and can harm brand perception.
Components of Relevance
Eligibility — Qualification Check
Determines whether the brand fits the query at all:
| Value | Description |
|---|---|
| Eligible | Brand fits the query — category, geography, and features match |
| Unclear | Ambiguous — can't be determined definitively |
| Not Eligible | Brand doesn't fit (e.g., query is about a different category) |
If the brand is Not Eligible, the Relevance metric is marked as not applicable (N/A) — it is not calculated or included in the Overall Score.
Fit to Intent (0–2)
How well the brand mention answers the user's question:
| Score | Description |
|---|---|
| 2 | Full match — brand recommended specifically for the task in the query |
| 1 | Partial — brand is related to the topic but doesn't precisely answer the question |
| 0 | No match — mention is unrelated to the user's intent |
Fit to Constraints (0–2)
Whether constraints specified in the query are respected (budget, region, technical requirements):
| Score | Description |
|---|---|
| 2 | All constraints met — brand matches budget, geography, and requirements |
| 1 | Some constraints met |
| 0 | Constraints ignored or brand doesn't meet them |
Justification Quality (0–1)
How convincingly the model explains why the brand fits:
| Score | Description |
|---|---|
| 1 | Quality justification — specific reasons, arguments, facts |
| 0 | No justification — brand simply named without explanation |
Formula
Relevance = 40 × (Fit to Intent / 2) + 40 × (Fit to Constraints / 2) + 20 × (Justification Quality / 1)Maximum score: 40 + 40 + 20 = 100. Each component contributes a weighted percentage: intent match and constraints are the most important (40% each), while justification quality adds 20%.
Note: if the brand is Not Eligible, the metric is marked as not applicable — Relevance is not calculated.
Examples
Relevance = 100 — Perfect Relevance
Brand fits (Eligible), precisely matches query (Fit to Intent = 2), all constraints met (Fit to Constraints = 2), quality justification (Justification Quality = 1):
Relevance = 40 × (2/2) + 40 × (2/2) + 20 × (1/1) = 40 + 40 + 20 = 100Relevance = 60 — Moderate Relevance
Brand fits (Eligible), partial query match (Fit to Intent = 1), constraints met (Fit to Constraints = 2), no justification (Justification Quality = 0):
Relevance = 40 × (1/2) + 40 × (2/2) + 20 × (0/1) = 20 + 40 + 0 = 60Relevance = 0 — Brand Doesn't Fit
Brand doesn't fit the query (Not Eligible) — Relevance is not applicable (N/A).
Why Relevance Matters More Than Visibility
High visibility without relevance is noise. If the model mentions your brand in response to an irrelevant query, the user will either ignore the recommendation or have a negative experience.
The ideal combination is high scores in both Visibility and Relevance: the brand is prominent and mentioned appropriately.
Use Positioning Match to evaluate what exactly the model says about the brand, and Usefulness to assess the practical value of the recommendation.
Related Metrics
- Visibility — brand prominence in the response
- Positioning Match — positioning alignment
- Usefulness — recommendation usefulness
- Metrics System Overview — all metrics and Overall Score
Visibility — Brand Visibility in AI Model Responses
The Visibility metric in BrandWise: how brand prominence in ChatGPT, Claude, and Gemini responses is measured. Formula, components, interpretation examples.
Positioning Match — Does AI Reflect Your Brand Positioning?
The Positioning Match metric in BrandWise: verify whether ChatGPT and Claude reflect your intended brand positioning. Formula, attributes, weights, examples.