Original measurements of SEO evidence, search-result repetition and content outcomes. Every published finding is tied to an observable sample, method and update date.
These reports update from the underlying evidence database. Studies without enough classified observations are labeled in progress, not presented as industry-wide statistics.
A longitudinal audit of quoted numerical claims on sampled SEO statistics pages. An inline link is not proof that the statistic is true or that the destination is the primary source.
We compare normalized numerical statements across a documented Spider search cohort. Exact-text normalization provides a lower-bound proxy for repetition, not a model of Google's information-gain signal.
A planned first-party observational cohort. No public performance findings are published until sites explicitly opt in, provenance and change logs are available, and the sample is adequate.
Spider search and page collection provide a historical evidence sample. Bounded TypeSafe JEV calls identify quantitative statements; Rust and SQLite calculate aggregates with explicit denominators. Claims are not validated merely because AI classified them as numeric. Public research must be separately checked for sampling bias and source quality.
See a measured real-world SEO case study →RefreshAgent connects your Search Console evidence to practical SEO decisions and improvements.
Explore RefreshAgent