Schema Markup Did Not Impact AI Citations in Ahrefs Test

Schema markup is becoming more common on web pages that AI cites. However, a recent report from Ahrefs showed that adding schema markup didn’t significantly increase citations by AI tools like ChatGPT and Google AI.

What is Schema Markup?

Schema markup is a special code added to websites to help search engines understand the content better. Many pages with high AI citations use this code. The Ahrefs report tracked 1,885 web pages that added a type of schema called JSON-LD. They compared these pages to others that did not use schema.

Key Findings from the Ahrefs Report

Ahrefs analyzed over 6 million web pages and found that pages cited by AI are about three times more likely to include JSON-LD schema. However, the report aimed to see if adding schema alone would lead to more citations without considering other factors, such as better content or more links.

They matched each page that added schema with three similar pages that did not. They then measured any changes in citations 30 days before and after adding schema.

The results were as follows:

• Google AI Overviews: Decreased by 4.6% (this is a notable drop)
• Google AI Mode: Increased by 2.4% (too small to be significant)
• ChatGPT: Increased by 2.2% (also too small to be significant)

Overall, all tests showed no clear positive or negative effects from adding schema markup.

The Decline in AI Overviews

The −4.6% decline in Google AI Overviews requires some explanation. The pages that added schema were already experiencing a decline in citations. The pages with schema declined slightly faster, losing around 12 citations per day. However, this number is small compared to the several hundred citations most pages get.

Ahrefs notes that this decline could be due to a minor negative effect from schema or just random chance. They did not reach a definitive conclusion.

Limitations of the Report

Every page in their study already had over 100 AI Overview citations before adding schema. This suggests these pages were already visible to AI tools. The report mentions that for pages not yet cited, schema could help with crawling and indexing, but this is not confirmed by the data.

Other limitations include the fact that pages adding schema might change other elements at the same time, making it hard to isolate the impact of the schema itself. They also pooled all types of schema together, and the 30-day observation period might not capture slower effects.

A related study from searchVIU found that five AI systems did not use schema when retrieving pages in real time. They only looked at visible HTML, ignoring JSON-LD and other formats.

Why This Is Important

Schema markup is often recommended to improve AI visibility. However, Ahrefs’ findings suggest that just adding schema does not boost AI citations for pages that are already noticed. The correlation between schema use and AI citations seems more connected to the overall quality of the site rather than the schema markup itself.

Looking Forward

The report could not determine if schema helps pages that are not yet cited. This needs more study. If pages are already visible to AI, adding JSON-LD may not increase their citations.

This report shows that while schema markup can enhance search visibility and display rich results, its role in increasing AI citations is more complex than previously thought. As the landscape of web content continues to evolve, further research will be essential.