How to track Share of Voice (SoV) in AI Search
A step-by-step guide to measuring your brand's visibility in generative AI engines like ChatGPT, Perplexity, and Gemini.
What is AI Share of Voice?
AI Share of Voice (SoV) is a marketing metric that measures the percentage of times a brand is mentioned or recommended by an Artificial Intelligence model (such as ChatGPT, Perplexity, or Claude) in response to relevant category queries, compared to its competitors.
Unlike traditional SEO, which tracks ranking positions (e.g., "Rank #1"), AI SoV tracks citation frequency and sentiment.
Why Metric AI SoV Matters
As users shift from traditional search engines (Google) to Answer Engines (LLMs), visibility in AI responses becomes critical. If an AI tool does not cite your brand, it effectively does not exist for that user.
Steps to Measure AI Share of Voice
1. Define Your Keywords
Identify the transactional queries users ask when looking for your product.
- Example: "Best CRM for small business", "Top open source alternatives to Salesforce".
2. Query the Models
Ask the same set of questions to major LLMs:
- OpenAI (GPT-4)
- Perplexity (Sonar)
- Anthropic (Claude 3)
- Google (Gemini Pro)
3. Analyze the Responses
For each response, record:
- Mention: Was your brand named? (Yes/No)
- Rank: What order was it listed in?
- Sentiment: Was the description Positive, Neutral, or Negative?
- Competitors: Who else was listed?
4. Calculate the Score
Use the following formula:
SoV % = (Total Brand Mentions / Total Queries Run) × 100
- Example: If you run 50 queries and your brand is mentioned in 10 of them, your SoV is 20%.
How to Automate SoV Tracking
Manual tracking is time-consuming and prone to bias. Tools like ViaMetric automate this process by:
- Running scheduled queries via API.
- Using a secondary AI agent to objectively analyze sentiment.
- Generating a historical trend line of your visibility.
Improving Your Score (GEO)
To improve your AI Share of Voice, focus on Generative Engine Optimization (GEO):
- Citation Authority: Ensure your brand is mentioned in high-authority sources (Wikipedia, G2, TechCrunch) that LLMs use for training.
- Structured Data: Use JSON-LD schema on your website so bots clearly understand your value proposition.
- Brand Clarity: Ensure your homepage explicitly states what you do in simple, unambiguous language.
