Skip to content
Snapshots, Mar 2026
SnapshotTwo studies, frozen 14 and 30 Mar 2026 · not updated

Content: what AI cites

of AI citations point to a blog or editorial page, the most of any type

Crawlers read product pages. Answers cite blogs.

Each page type’s share of AI crawler visits and of citations, on a log scale.

Page type
Blog and editorial14.4% → 20.2%1.4×
Homepage10.7% → 7.9%0.7×
Product page13.0% → 6.4%0.5×
Service and use case1.4% → 2.6%1.9×
About and contact0.4% → 1.7%4.6×
Resource and report0.4% → 1.4%3.8×
Review and directory0.3% → 1.2%4.8×
Category and browse1.9% → 0.9%0.5×
Documentation0.7% → 0.8%1.1×
Help and FAQ1.9% → 0.4%0.2×
Integration0.4% → 0.4%1.0×
Legal0.4% → 0.2%0.6×
Comparison0.4% → 0.2%0.6×
Blog and editorial pages get 20.2% of citations from 14.4% of crawler visits, the most cited of 13 types.

Top cited pages carry more schema and tables. Length barely differs.

The top 10% of 1,465 cited pages against the bottom half. Words: 2,521 against 2,305.

Pages withBottom → topGap
Any schema65.6% → 80.0%+14.4 pts
Article schema23.3% → 37.8%+14.5 pts
Tables28.2% → 40.0%+11.8 pts
FAQ schema5.3% → 11.1%+5.8 pts
FAQ content38.3% → 42.2%+3.9 pts
How-to content70.0% → 73.3%+3.3 pts
Person schema19.4% → 17.8%−1.6 pts
Any schema: 80.0% of the top 10% against 65.6% of the bottom half.

Common questions

What kind of pages does AI cite most?

In this Trakkr panel, citations are concentrated across a narrower set of page types than crawler requests. The crawl-share and citation-share comparison is descriptive, not a page-level conversion rate or proof that a crawl caused a citation.

Does schema markup help you get cited by AI?

In Trakkr’s data, AI-cited pages carry structured data more often than the open-web baseline, and some schema types show a measurable lift. Schema is not a magic switch, but it correlates with being cited.

What is the difference between being crawled and being cited?

Crawling means an AI bot fetched your page; citation means a model actually used it as a source in an answer. This dataset measures the gap (crawl share against citation share) by page type.

Shares are descriptive, not a conversion rate. Page types come from a classifier that covers 58% of pages; the rest are left out. Schema patterns are correlations from 1,465 pages, not proof that schema causes citations.

Snapshots, 14 and 30 Mar 2026337,362 citations from 882 brands11.4M crawler visits on 92 sitesPage types cover 58% of pages; the rest are unclassifiedTrakkr Data · CC BY 4.0