What problem does it solve? Standard keyword tools report zero volume for nearly every technical query, so volume-ranked lists bury the exact error strings and version-specific phrasings developers actually search. This Skill replaces tool volume with first-party evidence - search console exports, docs-site search logs, support tickets, issue trackers, and error telemetry - to produce a ranked, auditable keyword map for developer-facing content. ## Core Features & Use Cases - First-party demand harvesting: Pulls queries from eight tiered evidence sources in efficiency order, tagging every metric as Measured, User-provided, or Estimated. - Error-string clustering: Normalizes hosts, IPs, ports, and wrapper text into placeholders so one cause becomes one page target, not ten thin pages. - Winnability and refusal filters: Names the incumbent result per cluster, refuses queries the product cannot honestly answer, and routes regressions to engineering instead of docs. - Use Case: Given a Search Console export, 90 days of docs-search logs, and a Zendesk export for a webhook API, produce a keyword map that keeps zero-volume error strings, ranks clusters by outcome per page-hour, and ships Scope, Ranked clusters, Company-adoption set, Fix first, Do not target, and Backlog sections. ## Quick Start Ask the assistant to build a keyword map for your developer tool from your search console export, docs search logs, and support tickets.
