RankEcho reports
RankEcho's research hub links a dated 48-prompt, two-run citation study; an aggregate one-request screening of 5,542 responding homepages; and methodology or aggregate views for product audit and proof-loop data. Each page states its population, data availability, and limits. The readiness report has no public row-level dataset, the Proof Ledger does not publish individual prompts, and none of these reports identifies why an engine selected a source.
State of AI Citations, Q3 2026
RankEcho's dated Q3 2026 study contains 384 engine-prompt-run observations: 48 US-English commercial prompts, four engines, and two runs 24 hours apart. In run 0, conditional mean cited-domain overlap was 19.5% when both engines returned sources; pooled overlap was 13.8% when comparisons with an absent source answer were assigned zero.
The two adjacent runs also measured within-engine cited-domain overlap and Google AI Overview availability for this battery. These figures describe only the declared prompts, engines, locale, dates, and calculation rules. They do not predict another engine or run, identify source-selection causes, or characterize AI citation behavior generally. The full prompt battery and per-citation dataset are available with the report.
Website AI Readiness, July 2026
The readiness report screens 5,542 responding business homepages. In the aggregate counts, 33.7% had no detected structured data and 10.9% fell below RankEcho's initial-HTML visible-text threshold.
These are technical indicators, not observations of indexing, impressions, citations, or causal effects. The initial pool was 7,126 domains from public business listings across US regions; 1,584 that did not return usable content in one request were excluded. RankEcho has not published the row-level sample.
Product-data aggregate views
The State of AI Citations page is a descriptive aggregate of completed RankEcho audit runs. Those runs can differ in prompt panels, engines, dates, repeat domains, and sampling depth, so cross-audit rates are not presented as a representative industry benchmark.
The Proof Ledger explains fixed-prompt outcome measures and shows aggregate rates, medians, and timing-band counts only when configured display minimums are met. Those minimums limit very small aggregates; they do not establish representativeness or statistical sufficiency. Neither page publishes domains, customer identities, or individual prompts.
Each page states whether an aggregate outcome table is available for the data it receives. When a table is unavailable, the methodology and display policy remain visible without presenting an outcome rate.
How these reports are made
The citation study and readiness screening use separate designs; the product-data pages use their own aggregation rules. Read each report against its stated unit, denominator, exclusions, collection dates, and availability. Sharing a hub does not make their populations directly comparable.
Aggregate Proof Ledger win rates use measured prompts as the denominator, so prompts without a reportable win affect the rate even though prompt-level records are not public.
Dataset availability is stated per report. The citation study provides a downloadable artifact. The Website AI Readiness page provides aggregate counts and method but no row-level sample, so that screening is not independently reproducible from a public artifact. The product-data pages provide aggregate views only.
Citing this research
The reports are free to cite with attribution to RankEcho and a link to the report page. Where a dataset is published, it carries its own licence terms on the report page.
For questions about method, or to request a cut of the data that is not published, the contact page reaches us directly.
Frequently asked questions
The hub includes a dated 48-prompt, two-run citation study and an aggregate one-request Website AI Readiness screening of 5,542 responding homepages. It also links descriptive aggregate or methodology views for RankEcho audit and proof-loop product data. Availability and population limits are stated on each page.
For run 0 of this 48-prompt, four-engine study, conditional mean cited-domain overlap was 19.5% when both engines returned sources. Pooled overlap was 13.8% when comparisons with an absent source answer were assigned zero. These are descriptive estimates for the declared run and do not predict another engine, prompt, or date.
Yes, with attribution to RankEcho and a link to the report page. Where a dataset is published it carries its own licence terms on that page. For method questions or unpublished cuts of the data, use the contact page.
Its aggregate win rate uses measured prompts as the denominator, so prompts without a reportable win affect the rate. The Proof Ledger does not publish individual prompt records, whether wins or misses. If its display minimums are not met, it shows methodology and availability rather than an aggregate outcome table.