Research
SEO research and data studies
Original studies built on live scans of pages that actually rank, published quarterly, aggregates only.
Websites caught lying about their last updated date: does Google care? A study of 370 page-one results.
46% of page-one results claiming a fresh update changed nothing but the date. We diffed 89 update claims against their own archived snapshots, and we have the screenshots. Health pages faked updates the most, and Google's AI Overview cited date-bumps at the same rate as real updates.
Read the study →Data study · July 2026The best and worst entity models for SEO
We ran five entity extractors over 166 ranking pages and scored them on accuracy and false positives. One found the words Google uses nearly half the time; the rest fed 47-65% false positives with no footprint in the SERP.
Read the study →Data study · July 2026What earns a citation in Google's AI Overviews?
We tested ten signals SEOs optimize, including domain rating, backlinks, traffic, brand search volume, page age, and exact-match domains, against which page-one results the AI Overview actually cites. None predicted citation.
Read the study →Data study · July 2026Do AI assistants cite original content?
We captured 793 sources cited by Google AI Overviews and ChatGPT across 50 keywords, then scored 312 ranking pages on information gain. Cited pages were no more original than uncited ones, and three in four AI Overview citations point beyond page one.
Read the study →Data study · June 2026Information Gain Index: first edition
We scored 150 pages ranking in Google's top 3 across 50 keywords and 10 verticals. The median page scored 52/100; 21% were highly original and 24% added little beyond the rest of page one.
Read the study →