GEO Fundamentals

A GEO Audit Checklist You Can Act On and Retest

Turn broad advice into verifiable actions across crawlability, brand facts, content evidence, and observation records.

An audit checklist connecting facts, page checks, and retesting
On This Page

A GEO audit should produce actions and evidence. Use this checklist for an initial review, recording the page, time, and result for each item.

Access and Discovery

  • Important pages return the correct HTTP status.
  • Titles and body content are readable without browser JavaScript.
  • Robots rules, noindex, and canonicals match the environment's publication goal.
  • The sitemap contains published, indexable pages.
  • Internal links expose useful URLs without requiring a button action.

Keep test and production separate. Checking production indexability should not make test pages indexable.

Brand Facts and Content

  • Brand names, aliases, product names, and ownership are consistent.
  • Capability, pricing, and limitation claims link to maintained official information.
  • Important claims include sources, periods, and conditions.
  • Images, tables, and code examples serve the question being answered.
  • Editorial notes, placeholders, and unsupported outcome promises are removed.

Review individual pages with the content guide, rather than labeling an entire website as optimized.

Prompts and Observations

  • The prompt set covers discovery, comparison, and brand verification.
  • Market, language, and model scope are defined.
  • Original prompts, answers, and sources are preserved.
  • Failed collections are recorded separately from valid answers.
  • Mentions, recommendations, and citations have explicit definitions.

Actions and Retesting

  • Each action has a target page, owner, and completion condition.
  • Changes and publication times are recorded.
  • Retests use comparable prompts and scope.
  • Completed edits, crawling evidence, and observed answer changes remain distinct.

Fix directly verifiable issues first, then observe subsequent behavior. A single metric increase does not establish that one content edit caused it.