We audited 200 pages consistently cited in Google AI Overview panels. Here are the patterns.
Methodology
400 informational queries across five verticals over twelve weeks. For each, we recorded cited sources and audited them against non-cited pages ranking in positions one through five.
What cited pages share
87% use descriptive headings that function as standalone labels (vs 34% non-cited). They lead with direct answers in the first 100 words. They use lists, tables, and definitions more. 71% have visible author credentials (vs 29% non-cited).
What does not matter
Word count: no correlation. Domain authority: weak correlation. Schema markup: not distinguishing. Publication date: less important than information freshness.
The takeaway
The strongest signal is extraction readiness: how cleanly content can be pulled into a summary. Write as if a machine will quote you, because one is.