The AI Visibility Gap: What I Learned Analyzing 409 E-commerce Sites

The Debate That Started It All

Last week, <a href='https://www.linkedin.com/posts/tomfgoodwin_everyone-is-talking-about-agentic-commerce-activity-7416886828286402560-WFAy' target='_blank' class='text-blue-600 hover:underline font-bold'>Tom Goodwin</a> posted a provocative take on LinkedIn that caught my attention:

Everyone is talking about Agentic commerce and they are wrong. What will be big is LLM shopping—asking for advice on what to buy, asking for differences to be explained.

He's absolutely right. But there's a critical piece missing from this conversation: <strong>most e-commerce brands aren't even visible to LLMs during this discovery phase.</strong>

So I decided to test this hypothesis. Over the past month, I analyzed <strong>409 e-commerce websites</strong> across fashion, beauty, electronics, home goods, and other major categories to understand their AI readiness.

The results were more alarming than I expected.

The AI Visibility Gap: What I Learned Analyzing 409 E-commerce Sites

The Methodology

I built an automated audit tool that scans websites for the technical signals that make them readable by AI search engines like ChatGPT, Claude, and Perplexity.

Here's what I measured:

• <strong>JSON-LD structured data</strong> (the language LLMs use to understand content)
• <strong>Schema.org markup</strong> (product details, organization info, FAQs)
• <strong>Robots.txt configuration</strong> (is the site even crawlable by AI?)
• <strong>Content extractability</strong> (can an LLM parse your product information?)
• <strong>E-E-A-T signals</strong> (expertise, authoritativeness, trustworthiness)

I ran this analysis on:

• <strong>409 e-commerce websites</strong>
• Spanning <strong>8 major categories</strong> (fashion, beauty, home, electronics, shoes, food, travel, general)
• From small DTC brands to major retailers

The Findings: A Visibility Crisis

Finding #1: The JSON-LD Illusion

At first glance, the numbers looked promising: <strong>40.3%</strong> of sites had JSON-LD structured data.

Sounds good, right?

Wrong.

Only <strong>1.7%</strong> had Product schema.

Finding #2: The Average Site Is Effectively Invisible

The average site scored <strong>25.6/100</strong> on AI readiness.

But averages hide the real story. Here's the distribution:

• <strong>56.7%</strong> scored below 20/100 (critical - essentially invisible)
• <strong>3.7%</strong> scored 20-40/100 (poor - barely detectable)
• <strong>35.2%</strong> scored 40-60/100 (fair - partial visibility)
• <strong>3.7%</strong> scored 60-80/100 (good - decent visibility)
• <strong>0.7%</strong> scored above 80/100 (excellent - only 3 sites!)

Finding #3: Category Performance Reveals Platform Effects

Not all categories perform equally:

<strong>The insight:</strong> The 3.8x difference between Shoes (31.8) and General (8.3) isn't about industry sophistication—it's about <strong>platform choice</strong>.

Finding #4: The Organization Schema Trap

Here's a pattern I noticed across hundreds of sites:

Many had Organization schema (name, logo, social profiles) but were missing:

• Product schema
• FAQ schema
• BreadcrumbList schema
• Review/Rating schema

This creates a <strong>false sense of AI-readiness</strong>. Marketing teams see 'we have structured data' in their audit reports and assume they're covered. Meanwhile, when a potential customer asks an LLM to compare products, <strong>those sites simply don't show up in the answer.</strong>

The Bottom Line

After analyzing 409 sites, one thing is clear:

The AI visibility gap isn't coming.

<strong>It's already here.</strong>

And 60% of e-commerce brands are on the wrong side of it.