Structured data is markup that tells search engines and AI systems what a page is. The authoritative rules for it come from a small set of primary sources, not from third-party studies. Adding it to pages that AI already cites produces little to no measured gain, but it remains a precondition for eligibility.
Definitions
- Structured data: machine-readable markup that labels page content by type and property, drawn from the Schema.org vocabulary.
- JSON-LD: the structured data format Google recommends across its documentation.
- Rich result: an enhanced listing (Article, Product, FAQ, and others) that requires eligible structured data.
- AI Overviews / AI Mode: Google's AI-generated search summaries, which run on Google's normal indexing and eligibility rules.
Key Facts
- AI Overview click impact: search-result clicks fall from 15% to 8% when an AI summary is present. Source: Pew Research Center, March 2025.
- Cited-source clicks: 1% of visits that include an AI summary end in a click on a cited source. Source: Pew Research Center, March 2025.
- Schema effect on AI citations: adding JSON-LD produced no meaningful citation uplift across Google AI Overviews, AI Mode, or ChatGPT. Source: Ahrefs, May 2026.
- Zero-click searches: rose from 56% to 69% between May 2024 and May 2025. Source: Similarweb.
- Source hierarchy: Google instructs publishers to treat its own documentation as definitive for Google Search behavior. Source: Google Search Central.
The five primary sources
Measured effect
- Ahrefs method: 1,885 pages that added JSON-LD between August 2025 and March 2026, matched against 4,000 control pages, measured by difference-in-differences 30 days before and after. Every page already held 100+ AI Overview citations in February 2025. The AI Overviews change was a small, statistically significant decline; AI Mode and ChatGPT changes registered as statistical noise.
- Pew context: about 18% of Google searches triggered an AI Overview in March 2025, and 88% of summaries cited more than three sources.
- Untested case: for pages not yet cited by AI, schema may still aid initial crawling, parsing, and indexing. The Ahrefs study could not test this directly.
Platform statements
- Google Search team (April 2025): structured data provides an advantage in search results.
- Microsoft, Fabrice Canel (March 2025): schema markup helps Microsoft's models understand content for Copilot.
- Neither statement specifies the exact mechanism or commits to keeping the behavior fixed.
The operative implication: treat structured data as an eligibility precondition, not a growth lever. The primary sources converge on three requirements: use a supported format (JSON-LD), match markup to visible content, and follow the documented rules. The largest controlled study to date shows no citation gain from adding schema to pages already cited, so resourcing should favor correctness and content quality over schema volume.
Contexts
