Share of Model: The New Currency for Brand Visibility in the AI Era

For years, you have optimized Share of Voice, SEO rankings, and social media engagement as central KPIs for your brand's visibility. But a new force is fundamentally changing the rules of the game: Generative AI and Large Language Models (LLMs) like ChatGPT, Gemini, and the new AI Overviews in Google Search. These models are becoming the primary interface for information, recommendations, and purchasing decisions for millions of people. If you're not present here, you're losing out.

The critical question for you as a CMO is no longer just, "What position do we rank for?" but rather, "Are we part of the answer the AI provides?". To quantify and strategically manage this, we need a new metric: the Share of Model. Here at Davies Meyer, we have developed a robust methodology to do just that for our clients. I'll show you how you can apply this approach to your brand.

What is "Share of Model" and Why is it the New Benchmark?

Share of Model is your brand's percentage share of relevant and positive mentions within an LLM for a defined set of topics and questions. It is the AI-era equivalent of Share of Voice. It doesn't measure how loudly you shout (advertising spend), but how relevant and trustworthy your digital ecosystem is to be used as an answer source by an AI.

Why is this relevant for your decision-making? Because LLMs are redefining the customer journey. Instead of clicking through ten blue links, users receive a curated, synthesized answer. If your brand is missing from this synthesis—or worse, your competitor is featured prominently—you cease to exist in this critical moment of need. Share of Model is therefore a direct indicator of your future market presence and relevance.

The Methodology: How to Make Share of Model Measurable

Abstract concepts won't help you. What you need is a tangible, repeatable process. Our methodology is built on four pillars to clear the fog around the AI black box.

• Systematic Prompt Sets: We can't just ask one question and take the result as gospel. We develop hundreds of prompts that represent different user intents: from general informational queries ("What is the best skin cream for dry skin?") to comparisons ("Compare Brand A and Brand B on sustainability") and specific purchase intentions ("Which yogurt has the most protein?"). These sets are individually calibrated for each industry and brand.

• Controlled Sampling: A single answer from an LLM is a snapshot in time. The models are non-deterministic; they can provide slightly different answers to the same prompt. Therefore, we query systematically and repeatedly across different accounts and sometimes even simulated geographic regions to minimize random effects and obtain a stable picture. We test the relevant models such as ChatGPT (GPT-4), Google Gemini, and Perplexity.

• Quantitative Metrics: We extract hard KPIs from the thousands of collected answers. The most important are:
• Mention Rate: In what percentage of relevant answers is your brand mentioned at all?
• Citation Rate: In what percentage of cases is your website cited as a source for a piece of information? This is an extremely strong signal of authority.
• Sentiment Score: In what context are you mentioned? Is the mention positive, neutral, or negative? We analyze the tonality in the immediate vicinity of your brand's mention.
• Rank/Position: If the LLM generates a list ("The 5 best..."), what is your average position?

• Competitive Analysis: Your own scores are only half the story. The true strategic value comes from comparison. We conduct the same analysis for your 2-3 main competitors. This is the only way to see if you are a market leader, a challenger, or a laggard in the AI's "mindspace."

The Anatomy of a Good LLM Answer: Levers for Optimization

Once we know where you stand, the real work begins: targeted optimization. Unlike classic SEO, which was primarily about keywords and backlinks, here we need to understand how an LLM "thinks." The goal is to establish your brand and products as reliable entities in the AI's knowledge graph. The three most important levers for this are:

• Strengthening Entities: An LLM understands the world in terms of entities—clearly defined concepts like people, places, organizations, or indeed, brands and products. Your job is to enrich the "your brand" entity with clear, consistent, and interconnected information across the entire web. This starts with a complete Wikipedia entry and Wikidata profile and extends to consistent mentions in trade magazines and the way you talk about yourself on your own website.

• Leveraging Trustworthy Sources: LLMs derive their "knowledge" from a vast corpus of training data. This includes high-quality journalistic content, scientific publications, recognized industry portals, and of course, your own digital channels. Content marketing and PR take on a new, technical dimension here. It's no longer just about reaching people, but also about "feeding" the AI with facts that underpin your authority.

• Providing Structured Answers: Make it as easy as possible for the AI to understand your content and use it as an answer. Clear, structured data is worth its weight in gold. This includes detailed product pages with Schema.org markup, comprehensive FAQ sections that directly answer user questions, and how-to guides that focus on solutions rather than just product features. When an LLM has a clear question, you should provide the clearest answer on the web.

Monitoring Stack and Reporting: Keeping the Process Under Control

Share of Model is not a one-time analysis but an ongoing process. Your competitors are not sleeping, and LLMs are updated daily. That's why a lean but effective monitoring process is crucial.

Our stack typically consists of a combination of custom scripts (e.g., via Python) that automatically send the defined prompt sets to the LLM APIs, and specialized SEO tools that are beginning to integrate LLM tracking. The collected raw data is visualized in a dashboard (e.g., in Looker Studio or Power BI) that shows you and your team at a glance how your KPIs are developing over time.

We adapt the reporting cadence to your needs:

• Monthly Performance Dashboard: An overview of the core metrics (mention rate, sentiment, competitive comparison) for operational management.
• Quarterly Strategy Review: A deeper analysis of developments. Which of our measures have worked? Where have competitors caught up? What new topic areas are becoming relevant? This is where we derive the strategic adjustments for the next quarter.

Experience shows that targeted measures can significantly increase mention rates in relevant topic clusters, sometimes by 15-25%, within 6-9 months. However, this is a marathon, not a sprint.

Conclusion

The rise of generative AI is not a distant fantasy but an immediate reality that will determine the future visibility of your brand. CMOs who act now and begin to systematically measure and optimize their Share of Model will secure a decisive competitive advantage. It's about moving from a reactive SEO mindset to a proactive strategy where you actively shape the AI's information foundation.

The process is clear: define your relevant topics, measure your current status with a clean methodology, analyze your competitors, and implement targeted measures to strengthen your digital authority. Don't wait until your brand has disappeared from the relevant set of AI-generated answers. The work of becoming an integral part of the AI's knowledge base starts today.