Do you have any models to share?
We don’t mean supermodels or stock.
We’re talking about your brand’s share of answers in large learning models for search and artificial intelligence. Imagine a potential customer prompts ChatGPT, Anthropic, or even Google to find the best brand offering your product or service. Perhaps, they will ask if your organization is a reputable brand for these products.
Well, you should know your brand’s model share. But before you jump into the prompts, watch this video or read on for the perspective of CMI Chief Strategist Robert Rose:
The impact of model share on artificial intelligence
Recently, industry media published articles about “model share”, which is a new indicator worthy of attention.
Marketing Weekly Define model share as the ratio of the number of times a brand is mentioned by one or more large language models (such as ChatGPT, Meta’s Llama, Microsoft’s CoPilot, Google’s Gemini, Anthropic’s Claude) to the total number of mentions of brands in the same category.
(Just like the sudden emergence of SEO services, I have to ask whether there will be a new model of operating services. More on that later.)
as adweek explained that this new metric might look like brand recall or share of voice metrics and is designed to understand how each model views your brand. It might evaluate it through sentiment or how it prioritizes your brand as an answer to a prompt question.
I’m a little lost on how to calculate it, but esoteric marketing metrics already exist.
Some suggest that this metric evaluates your brand footprint and requires a more customized study of your brand as well as five competitors, looking at perception, total mentions, and performance for brand-focused queries. Now, it’s a project I can think about.
This is key to me.
New metric isn’t really a measurement
Model share may become a fact, but its biggest evolution lies not in how it is measured, but in the optimization strategies that come from analytics.
Prompt for new services. SEO agencies will start thinking about things like domains. (If you’re looking for ModelOptimization.AI, I found it this morning.)
If LL.M.s became the new search index, which ones would carry your brand and which ones wouldn’t? If this sounds like a lot of compilation work, you’re not wrong. This is a lot of work. I predict that new tech companies will offer cross-LLM analytics for quantitative research.
what do you think?
Please let us know via email or by commenting on CMI’s social channels.
Featured related content:
Cover image by Joseph Kalinowski/Content Marketing Institute
