Share of Model (SoM): The New Metric for AI Search Success

Every era of digital marketing has a defining metric. Impressions defined the banner ad era. Click-through rate defined the pay-per-click era. Organic rankings defined the SEO era.

Share of Model defines the AI search era.

As AI platforms become primary research tools for buyers, the question isn't how your website ranks on Google—it's how often AI mentions your brand when buyers ask relevant questions. Share of Model (SoM) is the metric that answers this question.

What Is Share of Model?

Share of Model (SoM) measures the percentage of relevant AI queries where your brand is mentioned, cited, or recommended, compared to your competitive set.

The Simple Formula

SoM = (Queries where your brand is mentioned / Total relevant queries measured) x 100

For example, if you track 100 queries relevant to your category and your brand appears in 25 of the AI responses, your Share of Model is 25%.

Why "Share of Model"?

The name reflects that you're measuring your share of an AI model's "awareness" of your category. Just as Share of Voice measures your share of advertising visibility, Share of Model measures your share of AI visibility.

The term acknowledges that AI responses are shaped by the model itself—its training data, its retrieval systems, and its generation patterns. Your visibility is literally a function of what the model knows and chooses to surface.

Why Share of Model Matters

The New Competitive Battlefield

In traditional search, 10 brands share page one. In AI search, 2-7 brands share a response. This compression means that small differences in Share of Model translate to significant differences in brand exposure.

Leading Indicator of Pipeline

Research shows that AI visibility correlates with pipeline metrics:

• Brands with 20%+ SoM see 23% higher inbound inquiry rates
• Brands below 5% SoM experience measurably lower brand recall among AI-using buyers
• Changes in SoM often precede changes in branded search volume by 4-8 weeks

Competitive Intelligence

SoM reveals competitive dynamics that other metrics miss:

• Which competitor is the "default recommendation" in your category?
• Are new competitors gaining AI visibility before they gain market share?
• Is your competitive position in AI aligned with your actual market position?

How to Calculate Share of Model

Step 1: Define Your Query Set

Build a comprehensive set of queries that your target buyers would ask AI:

Category queries (15-20 queries):

• "Best [category] tools"
• "Top [category] for [segment]"
• "Leading [category] platforms in 2026"

Comparison queries (10-15 queries):

• "[Your brand] vs [Competitor]"
• "Compare [category] options"

Use case queries (10-15 queries):

• "[Category] for [specific use case]"
• "How to [problem your product solves]"

Recommendation queries (5-10 queries):

• "What [category] should I use?"
• "Recommend a [category] for [need]"

Total: 40-60 queries minimum for statistically meaningful results.

Step 2: Run Queries Across Platforms

Execute each query across all major AI platforms:

• ChatGPT (latest model)
• Perplexity AI
• Google Gemini
• Google AI Overviews
• Claude

Step 3: Document Results

For each query, record:

• Was your brand mentioned? (Yes/No)
• What type of mention? (Primary recommendation, listed option, passing reference)
• What sentiment? (Positive, neutral, negative)
• Which competitors were also mentioned?
• What language was used to describe your brand?

Step 4: Calculate Your Scores

Overall SoM:

Total mentions across all queries and platforms / (Total queries x Number of platforms) x 100

Platform-specific SoM:

Mentions on a specific platform / Total queries on that platform x 100

Category-specific SoM:

Mentions for a specific query category / Total queries in that category x 100

Share of Model Benchmarks

By Competitive Position

• Category Leader: 30-50% SoM
• Strong Competitor: 15-30% SoM
• Emerging Competitor: 5-15% SoM
• Invisible: Below 5% SoM

By Industry

Benchmarks vary by industry maturity and competition:

• Enterprise SaaS - Category leaders typically have 35-45% SoM
• Marketing Technology - Highly competitive; leaders at 25-35% SoM
• Financial Services - Lower overall AI citation rates; leaders at 20-30% SoM
• Healthcare Technology - Growing AI search usage; leaders at 25-40% SoM
• Professional Services - Location-dependent; national leaders at 20-30% SoM

By Platform

Platform-specific benchmarks differ due to different source preferences:

• ChatGPT - Tends to mention established brands more frequently (training data effect)
• Perplexity - More dynamic; favors fresh, authoritative content (real-time search)
• Gemini - Favors Google ecosystem-optimized brands
• Claude - Tends toward balanced, multi-option responses
• AI Overviews - Correlates with organic rankings but isn't identical

Advanced Share of Model Metrics

Weighted Share of Model

Not all mentions are equal. Weight mentions by quality:

• Primary recommendation = 3 points
• Listed among top options = 2 points
• Passing mention = 1 point
• Negative mention = -1 point

Weighted SoM = (Your weighted points / Total possible weighted points) x 100

Share of Model Velocity

Track the rate of change in your SoM over time:

• Improving velocity - SoM increasing month over month (positive trend)
• Stable velocity - SoM consistent month over month (maintenance mode)
• Declining velocity - SoM decreasing month over month (requires intervention)

Competitive Share of Model

Calculate SoM for all competitors in your set:

• Relative SoM = Your SoM / Leading competitor's SoM
• A relative SoM above 1.0 means you lead; below 1.0 means you trail

Platform Gap Score

Identify your weakest platform:

• Calculate SoM per platform
• Compare your platform-specific SoMs
• The platform with the lowest SoM represents your biggest opportunity

Improving Share of Model

Quick Wins (2-4 weeks)

• Schema markup - Implement Organization, Article, FAQ, and Product schema
• Content structure - Reformat key pages with answer-first structure
• Brand consistency - Standardize brand information across all web properties
• FAQ content - Create FAQ pages targeting common AI queries in your category

Medium-Term (1-3 months)

• Original research - Publish data-driven research with citable statistics
• Comparison content - Create honest comparison pages for your category
• Authority content - Publish definitive guides in your domain
• Press coverage - Secure media mentions in high-authority publications

Long-Term (3-6 months)

• Third-party validation - Analyst reports, industry awards, review platform presence
• Wikipedia presence - Create or optimize your Wikipedia page
• Thought leadership - Bylines, speaking engagements, podcast appearances
• Training data strategy - Build presence in sources likely included in AI training data

Common Share of Model Mistakes

Mistake 1: Too Few Queries

Measuring SoM with 10-15 queries produces unreliable results. Use 40-60 queries minimum for meaningful data.

Mistake 2: Ignoring Platform Variation

Your SoM may be 30% on Perplexity but 5% on ChatGPT. Platform-specific measurement is essential for targeted improvement.

Mistake 3: Measuring Only Once

SoM is a trend metric. Single measurements are snapshots. Monthly tracking reveals trajectories and enables strategic response.

Mistake 4: Not Tracking Competitors

Your SoM in isolation is less meaningful than your SoM relative to competitors. Always benchmark against your competitive set.

Mistake 5: Counting All Mentions Equally

A primary recommendation carries more brand impact than a passing mention. Use weighted SoM for more accurate performance measurement.

Key Takeaways

• Share of Model (SoM) measures the percentage of relevant AI queries where your brand appears
• SoM is the defining metric for AI search performance, analogous to Share of Voice for traditional marketing
• Calculate SoM across 40-60 queries on all major AI platforms for reliable measurement
• Category leaders typically achieve 30-50% SoM; below 5% signals invisibility
• Use weighted SoM to account for mention quality (recommendation vs. passing mention)
• Track SoM monthly to identify trends and respond to competitive changes
• Improving SoM requires a combination of content optimization, authority building, and entity consistency