Positioning Match — Does AI Reflect Your Brand Positioning?

What Positioning Match Measures

Positioning Match measures how closely an AI model's response reflects your intended brand positioning. You define desired attributes with weights in your brand profile — and the metric checks whether those attributes are supported or contradicted in model responses.

This is a unique metric: it doesn't just measure "good or bad" but reveals the gap between desired and actual brand image as perceived by AI.

When It Applies

Positioning Match is calculated only when both conditions are met:

1. Brand is mentioned in the response
2. Brand is eligible (Eligibility ≠ Not Eligible)

If the brand isn't mentioned or doesn't fit the query, the metric is not applicable (N/A).

How Attribute Evaluation Works

In your brand profile, you set desired attributes — properties your brand wants to convey. Each attribute has a weight from 1 to 5 reflecting its importance.

For each attribute, the model response is evaluated:

Each supported attribute is backed by an evidence quote — a verbatim excerpt from the model response.

Formula

desired_supported = Sum(weight of supported attributes) / Sum(all weights)
desired_contradicted = Sum(weight of contradicted attributes) / Sum(all weights)

Positioning Match = 100 × clamp(0, 1, desired_supported − 1.2 × desired_contradicted)

Key insight: contradiction penalizes very heavily. The 1.2 coefficient means a contradicted attribute can reduce the score more than its supported weight would add — actively contradicting your positioning is worse than not mentioning it at all.

Calculation Example

Consider a brand with three attributes:

Total weight: 5 + 4 + 3 = 12.

desired_supported = 5 / 12 = 0.42
desired_contradicted = 3 / 12 = 0.25

PM = 100 × clamp(0, 1, 0.42 − 1.2 × 0.25) = 100 × clamp(0, 1, 0.42 − 0.30) = 100 × 0.12 = 12

Result: PM = 12 — a very low score. Despite "eco-friendly" being confirmed, the contradiction of "innovative" with the ×1.2 penalty cuts deeply into the score, and "premium" wasn't mentioned at all.

How to Configure Attributes

The metric quality depends directly on how you set up your attributes:

1. Choose specific, verifiable properties. "Best on the market" is vague. "Free delivery within 30 minutes" is specific and verifiable in model responses.

2. Set weights deliberately. Weight 5 = a core brand attribute you stand behind firmly. Weight 1 = desirable but not critical.

3. Limit the count. 3–7 attributes is optimal. Too many dilute the score — no single attribute carries meaningful weight.

4. Frame as properties, not actions. "Eco-friendly" is good. "Uses recycled materials" is even better.

Learn more about brand profile configuration in Brand Profile.

Interpreting Results

Related Metrics

• Relevance — mention appropriateness
• Usefulness — whether the mention helps decision-making
• Visibility — brand prominence
• Metrics System Overview — all metrics and Overall Score

Configure brand attributes and start evaluation