How many websites pass the Fact Attribution check?
Based on 65 audited sites · Updated 2026-08-14
At a Glance
95% confidence interval: 81.3%–95.7%
Current cross-sectional baseline. Historical snapshots remain stored but are not presented as performance change.
About This Check
Only activates when explicit citations, sources, footnotes, or attributed quotations are present, then checks that their reference remains usable and attached to the relevant content unit.
What To Do Next
9.2% of audited sites need attention on fact attribution (6 of 65 sites are warning or failing).
- Prioritize fact attribution on templates that create many URLs, because one template fix can improve a large share of content answerability.
- After changing the page, re-run the audit and compare the observed value for this check instead of relying on visual inspection alone.
- Use the 90.8% pass rate as a benchmark: if your site fails here, passing the check moves it closer to the majority of crawlable, AI-readable pages.
Industry Breakdown
How different industries perform on this check.
| Industry | Pass rate | Avg score | Evaluated |
|---|---|---|---|
| Education | 100% | 75.6 | n=9/199limited |
| Media & Entertainment | 100% | 71.2 | n=6/518limited |
| Healthcare | 100% | 74.5 | n=3/46limited |
| Finance & Banking | 100% | 73 | n=3/97limited |
| Marketing & Advertising | 100% | 74.8 | n=3/93limited |
| SaaS & Cloud | 88.9% | 77 | n=9/323limited |
| Technology | 85% | 70.7 | n=20/880limited |
n is sites evaluated for this check / audited sites in that industry. Informational or inapplicable results are not counted.
Related Checks
Run a free audit to see whether your site passes this check, or browse the leaderboard and the audited-sites directory.
Key Insight
The largest gaps are content metadata and trust signals: only 7% to 11% of sites include visible or structured dates, testimonials or case studies, and cleanly organized rendered DOMs. In this baseline, core security and server configuration checks are universally met, while page-level content quality and machine-readable publishing signals are much less consistent.