Skip to main content
Average Position sample card

​
What is it?

The Average Position score tells you where your brand tends to appear in AI answers when it is mentioned. While AI Visibility answers “Do we show up at all?”, Average Position answers:
“When we do show up, how high or low are we in the answer?”
Lower is better:
  • 1 means you’re usually mentioned first.
  • 3 means you’re usually around third.
  • Higher numbers mean you appear later in the answer.

​
How Average Position is calculated (per prompt)

For each tracked prompt, Mudra:
  1. Runs the prompt across multiple AI providers (for example: ChatGPT, Claude, Gemini, Perplexity).
  2. Keeps only responses where your brand is mentioned and a position has been assigned (1st, 2nd, 3rd, etc.).
  3. Collects the list of positions, like: positions = [2, 4, 3]
  4. Computes the average of those positions:
Average Position (for this prompt)=sum of positionsnumber of positions\text{Average Position (for this prompt)} = \frac{\text{sum of positions}}{\text{number of positions}} If your brand is not mentioned in any provider for that prompt, the Average Position for that prompt is not available.

​
Example: Average Position for one prompt

Prompt:
“Best AI platforms for marketing automation?”
You run this prompt on 4 providers and get:
ProviderBrand mentioned?Position
ChatGPTYes2
ClaudeYes4
GeminiNo—
PerplexityYes3
Step by step:
  • Keep only providers where your brand was mentioned: positions = [2, 4, 3]
  • Sum: 2 + 4 + 3 = 9
  • Count: 3
  • Average Position:
Average Position=93=3.0\text{Average Position} = \frac{9}{3} = 3.0 Result: For this prompt, your Average Position is 3.0. Whenever you are mentioned, you appear around third on average. If none of the providers had mentioned you, the Average Position for this prompt would be not available.

​
Overall Average Position (across prompts)

Mudra also shows an overall Average Position for the project. To compute it, Mudra:
  1. First computes Average Position per prompt (as above).
  2. Builds a list of valid averages (ignoring prompts with empty referencing): [avg_pos1,avg_pos2,…,avg_posn][\text{avg\_pos}_1, \text{avg\_pos}_2, \dots, \text{avg\_pos}_n]
  3. Computes the average of those values:
Overall Average Position=sum of valid prompt averagesnumber of prompts with a valid average\text{Overall Average Position} = \frac{\text{sum of valid prompt averages}}{\text{number of prompts with a valid average}} If all prompts are null (your brand is never mentioned), the Overall Average Position is also null.