The AI Marketing Metric
Have you considered evaluating your brand's AI visibility? Imagine your customer goes onto ChatGPT, Anthropic, or even Google and asks, "Who's the best brand for product A or B?" or "Is brand B reliable for these products?"
I will be exploring the concept of the "share of model." Let's get started.
Marketing Week introduced the “share of model” concept, describing it as the new marketing measure for the AI era. Basically, it counts the number of times a keyword or brand is mentioned by large language models (LLMs) like ChatGPT, Meta's Llama, Microsoft's Copilot, Google's Gemini, or Anthropic’s Claude, compared to other keywords or brands in the same category.
Think of it like search engine optimization (SEO). Will new services emerge to optimize content for these LLMs, which answer our consumers’ questions? Possibly. Some call it "model ops," where services might help optimize content for LLMs.
Other articles compare this new metric to share of voice or market share metrics. What would this look like in practice? According to an Adweek article, it could resemble brand recall or share of voice metrics, focusing on how each model perceives your brand, the sentiment, and how it prioritizes your brand in response to questions.
Marketing Week defines "share of model" as the number of brand mentions by LLMs relative to total mentions in the same category. Calculating this might be tricky, but it's one of those complex marketing metrics.
Some suggest measuring your brand's "footprint" by conducting a bespoke study of your brand and competitors, looking at perception, total mentions, and performance against specific brand-centric queries.
I believe this new metric will gain traction, not just in measurement but in the optimization strategies it inspires. SEO agencies should start thinking about this.
Standards will be crucial as LLMs become the new search index. It'll be fascinating to see where your brand appears and where it doesn't. Managing this long-term will be a significant effort, likely leading to new technology companies offering cross-LLM analysis for quantitative research.
