Share of Model: How Brands Are Transacting on Generative Visibility as Web Clicks Evaporate

The relentless rise of zero-click generative search interfaces has triggered an existential crisis for digital marketing attribution. With Google AI Overviews, Perplexity, and ChatGPT satisfying user intent directly within synthetic chat responses, organic referral click-through rates across B2B and consumer categories have dropped by an estimated 38% to 54% year-over-year.

In response to this evaporating click stream, forward-looking media agencies and enterprise brands are coalescing around a new North Star metric: Share of Model (SoM). Championed in an influential September 2026 whitepaper by AI research consultancy Smalk, Share of Model measures the percentage of generative AI responses—evaluated across a statistically rigorous, category-specific prompt set—that cite, recommend, or favorably mention a specific brand. However, as Smalk cautioned, transitioning Share of Model from an internal analytics dashboard into an industry-wide transactable media currency presents severe methodological challenges.

+-----------------------------------------------------------------------------+
|               SHARE OF MODEL (SoM) CALCULATION & TRANSACTION FLOW           |
+-----------------------------------------------------------------------------+
|                                                                             |
|   FORMULA:                                                                  |
|   SoM (%) = [ Total Generative Answers Recommending Brand X ]               |
|             -------------------------------------------------  * 100        |
|             [ Total Evaluated Category Prompts across Engines ]             |
|                                                                             |
|   -----------------------------------------------------------------------   |
|                                                                             |
|   ATTRIBUTION PARADIGM SHIFT:                                               |
|                                                                             |
|   Legacy Media Model (Pre-2025):                                            |
|   Ad Spend / SEO Content  --->  SERP Impression  --->  Website Click  --->  Lead |
|                                                                             |
|   Generative Media Model (2026+):                                           |
|   Brand Authority Work  --->  LLM Ingestion  --->  Share of Model  --->  Direct Deal |
|                                                                             |
|   -----------------------------------------------------------------------   |
|   CRITICAL BARRIER: "A metric nobody transacts on is a dashboard, not a     |
|   currency." — Standardization required across prompt sets and sampling.    |
+-----------------------------------------------------------------------------+

Technical & Strategic Deep Dive

For twenty-five years, digital marketing relied on the click as the fundamental atom of value exchange. CPC (cost-per-click), CTR (click-through-rate), and multi-touch attribution models were all anchored to HTTP request logs. When a user asks an AI assistant for buying recommendations, that click trail vanishes.

1. Defining the Share of Model Equation

To move beyond crude vanity metrics, Share of Model requires rigorous mathematical boundaries. Smalk defines SoM through three weighted dimensions:

1. Mention Share: Does the brand appear in the generated response?
2. Recommendation Rank: In what position is the brand presented (e.g., #1 Recommended vs. “Other alternatives to consider”)?
3. Sentiment & Qualification: Is the recommendation positive, neutral, or accompanied by critical caveats (e.g., “High enterprise capability, but expensive and complex to deploy”)?

Formally, unweighted Share of Model across engine set $E$ and prompt battery $P$ is expressed as:

$$ ext{SoM}_{ ext{brand}} = rac{1}{|E| \cdot |P|} \sum_{e \in E} \sum_{p \in P} \mathbb{I}( ext{brand cited in } ext{Response}(e, p))$$

To account for probabilistic variance, each prompt must be sampled across multiple randomized user agents and network exit nodes to eliminate geographic and user-profile bias.

2. The “Dashboard vs. Currency” Dilemma

As the Smalk report bluntly notes, *”A metric nobody transacts on is a dashboard, not a currency.”* For Share of Model to function as a true media currency—where brands pay agencies or publishers based on guaranteed SoM gains—the industry requires three structural pillars:

1. Standardized Prompt Taxonomies: Who defines the questions? If an agency writes the test prompts, they can easily game the battery with leading queries that bias toward their client’s known strengths.
2. Third-Party Verification: Independent measurement auditors must run the synthetic evaluations from air-gapped infrastructure to prevent vendor tampering.
3. Fraud Resistance: As SEO shifts to GEO, black-hat reputation firms are flooding Reddit, GitHub, and review hubs with synthetic persona consensus designed to manipulate model weights during pre-training and RAG scraping passes.

3. Impact on Content Licensing and Digital PR

Because LLM retrieval pipelines heavily favor trusted third-party consensus, the economic value of brand PR has skyrocketed. Digital PR is no longer about acquiring PageRank backlinks; it is about injecting semantic entity associations into the authoritative sources that LLM crawlers query during live synthesis.

Publishers with established editorial authority are capitalizing on this by licensing structured product databases directly to AI search providers, creating a closed ecosystem where unlisted brands are rendered invisible.

Real-World Utility & Limitations

Practical Value for Enterprise

• Executive Clarity: Gives CMOs a single, quantifiable metric to demonstrate brand presence inside generative search environments to boards and investors.
• Competitive Intelligence: Enables instant gap analysis showing exactly which competitor brands are winning recommendation share across specific product categories.

Methodological Weaknesses

• High Volatility: An algorithmic update to Google AI Overviews or a parameter tweak in Perplexity can cause a brand’s SoM to swing by 25% overnight.
• Attribution Disconnect: Proving direct pipeline revenue originating from an unclicked AI recommendation remains notoriously difficult, requiring post-purchase onboarding surveys.