Skip to main content

GSR - A New Metric for Measuring Brand Visibility in AI-Generated Results

· 4 min read

The Generative Search Ratio (GSR) is a new metric that developer Todd Gray has proposed to assess how discoverable a brand is within environments driven by large language models (LLMs).

As businesses contend with the shift from traditional search methods to those powered by generative AI, they must now ensure visibility not just in Google rankings but also in the synthesized responses provided by LLM tools like ChatGPT and others.

What is GSR?​

The GSR serves as a way to quantify this new kind of digital presence, helping brands understand their positioning in this emerging search landscape. It basically assesses the relationship between your visibility in large language model (LLM) responses and your performance in traditional SEO.

The basic formula is: GSR = (GEO Score × Visibility Weight) / SEO Score

Here's how the components break down:​

  • GEO Score: The percentage of key prompts where your brand appears in an AI generated response.

  • Visibility Weight: A modifier that adjusts for how prominent the mention is, based on prompt type.Core prompts (e.g., "best CRM for startups") receive more weight than niche ones (e.g.,"top CRM that works with an internal Postgres database").

  • SEO Score: How well your brand performs in the keyword rankings you care about—usually pulled from search analytics or third party tools. These two metrics combined give a rough sense of whether your brand is keeping up, falling behind, or pulling ahead in this new era of AI driven search results

Together, these factors produce a single number that helps assess whether your brand’s AI presence is keeping up with—or surpassing—your SEO presence.

Interpreting GSR​

  • GSR ≥ 1.0 → Your brand has a strong presence in AI generated content. It appears as though your brand ranks higher in AI answers than it does in traditional organic search results.

  • GSR 0.5 – 1.0 → You are moderately visible to LLMs. Your content is discoverable, but it is not prominent enough in AI responses to ensure that you are being "seen" by these models.

  • GSR < 0.5 → You have good SEO and poor visibility to LLMs. You are not being picked up by the LLMs that generate AI responses, which may necessitate changes in your content structure, clarity, or authoritative signals. (most brands are here?)

  • GSR > 1.0 with low SEO → You are doing well in AI generated responses but underperforming in traditional search results. It is time to delve into the basics of SEO.

Low GSR and low SEO → You're essentially invisible on both fronts. It is time for a comprehensive strategy reset to ensure full visibility.

Not All Prompts Are Equal​

One of the most insightful aspects of the GSR metric is how it differentiates between types of prompts. As Todd Gray points out, not all citations generated by LLMs are equally valuable:

  • Core Prompts (broad, high-intent questions) acarry the most weight.

  • Conditional Prompts (requiring specific parameters) provide real value and are worth tracking, but they are not as impactful as the more general prompts. They're context-dependent.

  • Niche/Web Prompts are less impactful but still worth tracking for completeness.

This granularity helps the GSR account for meaningful discoverability in the real world, not just raw mention counts.

Limitations and Next Steps​

GSR is still an emerging concept. It requires:

  • A reliable prompt library across core/niche/conditional types

  • Access to AI-generated responses, from APIs or scraping)

  • Standardized SEO performance benchmarks

LLM Visibility methodology is still being refined, but as we move from links to language, we'll need new metrics that follow. So GSR might be the first of many.