Author: Kita Yohei Published: June 9, 2026
SOM (Share of Model) is a metric indicating the percentage of AI-generated responses that mention or cite a brand for specific queries. It is positioned as the AI equivalent of "Share of Voice (SOV)" in traditional marketing. SOM is one of the representative metrics for quantitatively measuring GEO strategy outcomes, functioning as a KPI for understanding the progress and effectiveness of GEO measures. Note that the term "Share of Model" is not yet standardized in the industry, with multiple expressions including AI Visibility, AI Mention Rate, and Recommendation Rate used in parallel.
What You'll Learn on This Page
- The meaning, definition, and calculation method of SOM
- The difference from SOV and SOS
- The full picture of SOM and related metrics
- Its role in GEO strategy
- Points to note when measuring SOM
- Common misconceptions
What Is SOM?
Share of Model (SOM) is a metric expressing as a percentage how often a brand appears when AI responds to specific topics or questions. For example, if "Genview" was mentioned in 30 out of 100 AI responses to the question "What GEO strategy tools do you recommend?", the SOM would be 30%.
SOM has different values per model and per query. Measuring and managing ChatGPT SOM and Perplexity SOM separately is the practical standard.
Differences Between SOM, SOV, and SOS
SOM differs from traditional metrics in the following ways.
| Metric |
Measurement target |
Primary measurement method |
| SOM (Share of Model) |
Brand mention rate in AI responses |
Multiple query submissions to AI, mention count aggregation |
| SOV (Share of Voice) |
Brand exposure rate in advertising/media |
Impressions, ad spend, mention counts |
| SOS (Share of Search) |
Brand search rate in search engines |
Competitive comparison of search volume |
SOM is a metric newly needed in the AI era, quantifying brand visibility in AI responses that SOV and SOS cannot measure.
How to Calculate SOM
[SOM Formula]
SOM (%) = Number of responses mentioning brand ÷ Total responses × 100
[Measurement Steps]
1. Define target queries
Example: "Recommended GEO strategy tools," "How to get cited by AI," etc.
2. Select target AI models
Example: ChatGPT, Gemini, Perplexity, Claude, etc.
3. Submit each query multiple times and collect responses
4. Check whether each response contains a mention of the brand
5. Mention count ÷ Total responses × 100 = SOM (%)
Note that SOM has multiple measurement methods, and values may differ depending on whether the basis is mention-based, citation-based, or recommendation-based. It is important to clearly define the measurement approach based on the measurement objective before calculating.
The Full Picture of SOM and Related Metrics
To accurately understand brand visibility in AI responses, it is recommended to observe multiple metrics in combination rather than SOM alone. SOM measures "whether the brand appeared," but how it appeared needs to be understood through separate metrics.
| Metric |
What it measures |
Key measurement point |
| SOM (Share of Model) |
Whether the brand was mentioned |
Percentage of responses where the brand name appears |
| Citation Rate |
Whether the brand was cited |
Percentage cited with source URL |
| Sentiment |
How it was described |
Ratio of positive/negative/neutral context |
| Recommendation Rate |
Whether the brand was recommended |
Percentage actively presented as a "recommendation" |
For example, even if SOM is high, negative Sentiment may have a negative impact on the brand. And if Citation Rate is low, it becomes a signal that content structure improvement is needed. Combining all four metrics provides a comprehensive picture of GEO strategy.
Its Role in GEO Strategy
In GEO strategy, SOM is positioned as "one of the representative metrics for quantifying strategy outcomes."
① KPI for measuring GEO strategy effectiveness
After implementing GEO measures such as content improvement, Entity maintenance, and citation acquisition, SOM changes can confirm strategy effectiveness. SOM changes become objective evidence for judging GEO strategy success or failure.
② Understanding relative competitive position
Measuring not only your own SOM but also competitors' SOM allows you to understand the competitive landscape for AI response share. SOM is also an effective metric from a competitive analysis perspective.
③ Setting priorities by model and query
Analyzing which AI models have lower SOM and which query categories have lower SOM allows you to set GEO strategy priorities.
Points to Note When Measuring SOM
① Measure per AI model
ChatGPT, Gemini, Perplexity, and Claude each have different response tendencies. Even if SOM is high on one model, it may be low on others. Prioritize measurement on the AI models most commonly used in your target market.
② Submit the same query multiple times
AI responses are not completely identical every time. Submitting the same query multiple times and taking an average improves measurement precision.
③ Set multiple query types
Measuring SOM across multiple query types — "recommendation queries," "comparison queries," "definition queries" — provides a more accurate picture of brand AI visibility.
④ Track regularly
AI models are updated regularly, so SOM fluctuates. Understanding trends through regular monthly or quarterly measurement is important.
Genview's Definition
In the context of GEO strategy, SOM is defined as "a metric quantifying the percentage of AI-generated responses that mention or cite a brand for a specific query set — one of the representative KPIs for measuring GEO strategy outcomes."
Genview positions continuous SOM tracking as a means of measuring GEO strategy effectiveness. By standardizing query design, model selection, and measurement frequency, SOM functions as a reproducible GEO strategy evaluation metric.
This definition reflects Genview's perspective and is not an industry consensus.
Related Terms
- GEO (Generative Engine Optimization): The overall initiative to optimize brand visibility in AI-generated responses. SOM is one of the representative metrics for quantifying GEO strategy outcomes.
- Citation: Mentions and citations of a brand's content in AI responses. SOM quantifies the percentage at which citations occur.
- Citation Marketing: The practice of strategically designing and acquiring brand mentions in external media. Accumulated citation acquisition leads to SOM improvement.
- Authority: The degree to which AI judges a brand as a trustworthy source on a specific topic. Higher authority is considered to correlate with higher SOM.
- Entity: The mechanism by which AI recognizes a brand as a distinct concept. Being recognized as an Entity by AI is a prerequisite for SOM acquisition.
- Grounding: The mechanism by which AI generates responses based on specific information sources. Being grounded leads to the state measured as SOM.
- Zero Click Search: The state where users complete a search without clicking any links. SOM functions as a metric for measuring brand visibility in a Zero Click environment.
Common Misconceptions
Misconception 1: "High SOM leads to business results"
SOM is a metric indicating brand visibility, but does not itself guarantee business outcomes (sales, lead acquisition). The correlation between SOM and actual business results varies by industry, product, and query type. While setting SOM as a KPI, it is important to also observe its correlation with business outcome metrics.
Misconception 2: "SOM is a fixed number"
AI models are updated regularly, so SOM fluctuates. SOM at a given point in time is a snapshot, and the correct way to use it is to understand it as a trend through continuous tracking.
Misconception 3: "0% SOM means insufficient strategy"
SOM changes significantly depending on the queries measured. When SOM is low, there may be insufficient strategy, but the possibility that query selection doesn't reflect reality should also be considered. Revising query design and improving measures should proceed in parallel.
Frequently Asked Questions
- Q: How should I set SOM targets?
- A: This varies by industry, competitive landscape, and current brand status. First measure the current SOM as a baseline, then compare with competitors' SOM before setting targets. Depending on the industry, with major players narrowed to 3–5 companies, setting a target of making the Top 3 may be realistic.
- Q: Which AI model's SOM should I measure first?
- A: In the Japanese market, starting with ChatGPT and Gemini is recommended. Include AI Overviews as a target when you prioritize impact on search traffic. Copilot is also important for B2B companies.