3 entries
Rand Fishkin, SparkToro
The documentation references Chrome-derived clickstream data (e.g. a chrome_trans_clicks field tied to sitelink prominence) and host-level NavBoost evaluation. Fishkin argued this sits awkwardly with Google's repeated public denials that Chrome data influences rankings.
Mike King, iPullRank
The docs describe NavBoost storing click-based signals (goodClicks, badClicks, lastLongestClicks, unsquashedClicks) with click data normalised to limit manipulation, and a rolling click window. King read this as evidence Google captures and processes click behaviour as a ranking input.
Mike King, iPullRank
The documentation stores a compressed siteAuthority quality signal (used within a Q* ranking system) and a hostAge attribute whose stated purpose is to sandbox fresh spam at serving time. King set these against Google's public statements of 'we don't have website authority score' and 'there is no sandbox'.