How to measure and analyze Share of Model compared to competitors?
Share of Model (SoM) is the frequency and favorability of a brand’s appearance in Large Language Model (LLM) responses compared to competitors. Unlike Share of Search, which tracks user query volume, SoM measures the AI's "preference" for a brand when synthesizing answers. According to Gravity Global, measuring SoM is critical as 50% of search queries now end without a click, making the AI's direct answer the final destination for many users. This guide covers the specific metrics, manual audit workflows, and analysis frameworks needed to benchmark your brand's AI visibility against competitors in the Generative Engine Optimization (GEO) era.
What is Share of Model (SoM)?
Share of Model (SoM) is a metric that quantifies a brand’s visibility, sentiment, and citation frequency within AI-generated responses. According to Marketing Association NZ, SoM serves as the "AI equivalent of share of search," directly influencing brand perception in zero-click scenarios. For marketing teams, this shifts the KPI focus from "ranking position" to "mention frequency" and "sentiment quality" across platforms like ChatGPT, Claude, and Perplexity.
Why SoM Matters in B2B SaaS
In B2B SaaS, buyers often use LLMs to generate "Best of" lists or compare software solutions (e.g., "Compare DECA vs. Jasper"). If an LLM fails to mention a brand in these synthesized answers, the brand effectively does not exist for that prospect. High SoM ensures that a brand is not only visible but also framed as a top recommendation during the evaluation phase.
How does SoM differ from Share of Search?
Share of Search tracks human curiosity (queries), while Share of Model tracks AI probability (outputs). Search Engine Land notes that Share of Search has historically been a leading indicator of market share, but SoM is now the predictor of future market share in an AI-first world. The following table contrasts the two metrics to clarify their distinct roles in a modern strategy.
Primary Data Source
Search Volume (Google Trends, Keyword Planner)
AI Generated Outputs (ChatGPT, Perplexity, Gemini)
Measurement Unit
Query Volume (Input)
Mention Frequency & Context (Output)
Key Driver
Brand Awareness / Ad Spend
Entity Authority / Structured Data / GEO
Outcome Goal
Click-Through Rate (CTR)
Citation & Recommendation
How can we measure SoM manually?
A manual SoM audit involves systematically querying LLMs with high-intent prompts and recording the presence of brand entities. Hallam Agency suggests using a "incognito" or "fresh chat" mode to minimize personalization bias during this process. This 3-step workflow provides a baseline assessment of how major AI models currently perceive your brand versus competitors.
Step 1: Define Target Prompts
Create a list of 20-30 prompts that your target persona (e.g., "In-house Marketer") would ask.
Category Prompts: "What are the best GEO tools for B2B marketing?"
Comparison Prompts: "Compare DECA and Jasper for content writing."
Use-Case Prompts: "How to optimize blog posts for AI search?"
Step 2: Execute & Record
Run these prompts across 3 major engines (ChatGPT-4, Claude 3, Perplexity) and record the data.
Mention: Did the AI mention your brand? (Yes/No)
Rank: In a list of 10, where did your brand appear? (e.g., #3)
Sentiment: Was the description Positive, Neutral, or Negative?
Citation: Did it link to your site? (Yes/No)
Step 3: Calculate the Score
Calculate your "Mention Rate" (Total Mentions / Total Prompts). If you appeared in 15 out of 30 prompts, your raw visibility score is 50%. Compare this against the same calculation for your top 3 competitors to determine your relative Share of Model.
What are the key metrics to analyze?
The three core pillars of SoM analysis are Mention Frequency, Sentiment Score, and Citation Authority. HubSpot emphasizes that measuring frequency alone is dangerous; negative mentions can damage brand equity faster than invisibility. Tracking these specific KPIs ensures a holistic view of your brand's health in the AI ecosystem.
Inclusion Rate (Visibility): The percentage of times your brand appears in relevant category searches.
Formula: (Brand Mentions / Total Category Prompts) * 100
Share of Synthesized Voice: Your brand's dominance in comparison queries (e.g., being the first recommended solution).
Metric: % of times listed as "#1" or "Best Overall."
Sentiment Quality: The qualitative framing of your brand.
Scale: -1 (Negative), 0 (Neutral), +1 (Positive).
Citation Density: How often the AI provides a clickable link (Source) to your owned properties.
Goal: High citation density drives referral traffic even in zero-click environments.
How to analyze competitors' Share of Model?
Competitor analysis requires identifying the "Gap Prompts" where rivals are mentioned, but your brand is excluded. Superlines recommends analyzing the specific sources cited in competitor mentions to reverse-engineer their authority strategy. By understanding why an AI prefers a competitor, you can adjust your content strategy to close the gap.
Source Gap Analysis: If ChatGPT cites TechCrunch for a competitor but not you, that media outlet is a high-priority PR target.
Feature Gap Analysis: If AI consistently praises a competitor's "ease of use," your content must explicitly highlight your own usability features to counter this narrative.
Co-occurrence Analysis: Identify which entities (tools, experts, concepts) frequently appear alongside competitors. aligning your brand with these entities can help "borrow" their relevance.
Translating SoM Metrics into Actionable GEO Strategy
Measuring Share of Model is the first step toward reclaiming control of your digital narrative in the Age of AI. By shifting focus from keyword rankings to entity visibility, brands can ensure they remain discoverable in the zero-click future. The next strategic move is to implement Offensive GEO tactics to actively influence these metrics by optimizing content for machine readability and citation.
FAQs
What is the difference between Share of Model and Share of Voice?
Share of Model tracks brand presence in AI-generated answers, whereas Share of Voice typically measures advertising reach or social media mentions. According to Marketing Association NZ, SoM is the modern evolution of Share of Voice for the generative AI era.
Which tools can track Share of Model automatically?
Tools like Profound, Brandlight, and custom scripts using OpenAI's API can automate the tracking of brand mentions across LLMs. Wix Studio lists these as emerging "AI Visibility" platforms that replace manual checking.
How can I improve my brand's Share of Model?
Improving SoM requires Generative Engine Optimization (GEO), which involves structuring content for AI readability and securing mentions in authoritative sources. Unusual VC states that "grounding" your brand in high-trust data sources is the most effective way to influence model outputs.
Is Share of Model a ranking factor for Google?
No, SoM is a metric for LLMs (like ChatGPT), while Google uses its own algorithms; however, Google's AI Overviews are blurring this line. Search Engine Land predicts that as search becomes more AI-driven, the two metrics will increasingly converge.
Can Share of Model be negative?
Yes, if an AI consistently associates your brand with complaints, bugs, or controversies, your SoM has a negative sentiment value. HubSpot advises that sentiment analysis is crucial to ensure visibility doesn't become a liability.
Reference
Gravity Global | Measuring AI and Zero-Click Impact | https://www.gravityglobal.com/blog/measuring-ai-and-zero-click-impact-on-brands
Marketing Association NZ | Share of Model: A New Metric | https://marketing.org.nz/resource-hub/share-of-model-a-new-metric-for-marketing-strategies
Search Engine Land | Guide to Share of Search | https://searchengineland.com/guide/share-of-search
Hallam Agency | Share of Model: A Key Metric | https://hallam.agency/blog/share-of-model-a-key-metric-for-ai-powered-search/
HubSpot | Brand Sentiment Analysis | https://www.hubspot.com/aeo-grader/brand-sentiment-analysis
Superlines | How to Track Brand Mentions in AI | https://www.superlines.io/articles/how-to-track-brand-mentions-in-ai-search-results
Unusual VC | How LLMs Evaluate Brands | https://www.unusual.ai/blog/how-llms-evaluate-brands
Last updated