Share of Model
Understand your Share of Model score - what it measures and how to improve it.
Last updated August 31, 2026
Share of Model (SoM) is AEO Goal’s headline metric. It measures how often your brand appears in AI engine answers relative to all possible answers - your “market share” of AI responses.
How SoM is calculated
SoM (%) = brand_appearances ÷ total_answers × 100
- brand_appearances - number of prompt runs where your brand was mentioned
- total_answers - total number of prompt runs across all tracked prompts in the period
Example: If AEO Goal ran 200 prompt checks this month and your brand was mentioned in 50 of them, your SoM is 25%.
SoM vs. Citation Rate
- Citation rate - mentions per total runs (all prompts combined)
- SoM - same calculation, often used as the headline for comparisons
In practice they’re the same formula. The distinction matters when benchmarking: SoM is often expressed relative to competitor SoM scores to show relative share of the AI answer space.
Share of Model vs. competitors
The Competitors view shows a side-by-side SoM chart. If your SoM is 25% and the category leader is at 40%, there is 15 percentage points of gap to close.
SoM scores across all tracked competitors don’t add up to 100% - AI engines often mention multiple brands in one answer, and some answers mention no tracked brand at all.
Interpreting your SoM
| SoM | Interpretation |
|---|---|
| 0-10% | Low - AI engines rarely associate your brand with tracked topics |
| 10-30% | Moderate - present but not dominant |
| 30-60% | Strong - frequently mentioned across tracked prompts |
| 60%+ | Dominant - most AI answers about your topic mention you |
Improving SoM
- Expand tracked prompts - add more decision-stage prompts where your brand could appear
- Publish authoritative content - target the exact questions your prompts cover
- Increase structured data - FAQ schema, How-To schema, and product schema help AI engines reference your content
- Build topical authority - depth of coverage on core topics signals expertise to AI engines
- Monitor losses - use the Prompts view to see which specific prompts you’re losing and why