Measuring ‘Share of Model’ (SoM): The New Executive KPI Replacing Organic SERP Tracking

CMOs and marketing directors can no longer justify SEO budgets purely through rank trackers. When an executive runs an enterprise evaluation in ChatGPT, Perplexity, or Gemini, traditional rank trackers register zero data. Share of Model measures brand presence inside the synthetic consensus.

1. Share of Voice (SoV) vs. Share of Model (SoM)

Metric Dimension	Share of Voice (Legacy SEO)	Share of Model (Generative Era)
Data Collection	Scraping position 1-10 on static Google SERPs	Multi-model automated prompt sampling across hundreds of prompt variants
Measurement Unit	Estimated click curve percentage based on rank	Brand citation frequency, sentiment polarity, and recommendation order
Model Coverage	Google and Bing desktop/mobile indices	GPT-4o, Claude 3.5 Sonnet, Gemini 1.5, Perplexity Sonar
Executive Relevance	Declining; disconnects from zero-click outcomes	Directly correlates with autonomous buyer shortlisting and brand recall

2. Algorithmic Calculation of Share of Model

# Share of Model (SoM) Scoring Engine
def calculate_share_of_model(prompt_runs, target_brand):
    """
    Evaluates brand inclusion and sentiment across synthetic answer runs.
    """
    total_runs = len(prompt_runs)
    cited_runs = [r for r in prompt_runs if target_brand.lower() in r["text"].lower()]
    
    citation_rate = len(cited_runs) / total_runs
    top_recommendation_rate = len([r for r in cited_runs if r.get("rank_in_answer") == 1]) / total_runs
    
    return {
        "share_of_model_pct": round(citation_rate * 100, 2),
        "primary_recommendation_pct": round(top_recommendation_rate * 100, 2)
    }

3. Frequently Asked Questions