Tech audit items
Reference catalogue of every audit check across all three pillars. Definitions, weights, per page pass, partial, or fail status.
Executive summary
Three short paragraphs explaining the feature and value.
Tech audit items is the reference catalogue of every audit check the system runs across all three pillars. Three tabs, one per pillar (AEO, GEO, SEO), hold the full list. Each check has a name, definition, weight in scoring, and the per page pass, partial, or fail status from the most recent scan.
Read this before fixing anything. The weight column tells you what moves the score and what does not. A check at weight ten with twenty failing pages is the highest value target. A check at weight one with three failing pages is busy work disguised as remediation. The catalogue makes priority obvious at a glance.
Definitions matter. Two teams looking at the same audit failure can draw different conclusions about what is broken if the check description is ambiguous. The catalogue gives the canonical definition for every check, plus pointers to the relevant remediation surface inside the workspace. Auditors and developers read it side by side during sprint planning.
Key highlights
Five capability points teams should know about quickly.
- Three tabs: AEO, GEO, SEO catalogues
- Every check carries name, definition, weight
- Per page pass, partial, fail status
- Read before fixing to know what moves score
- Canonical definitions for auditor and developer alignment
Top FAQs
Five common questions answered for fast practical clarity.
How many checks are there per pillar?
Roughly thirty to forty per pillar, depending on category. AEO has the most because the AEO surface is the newest and the rubric is still expanding. SEO has the fewest because the SEO check set has stabilised over fifteen years. New checks land in releases and are flagged with a date so users see what has changed between scans.
What is the weight column actually scoring?
The relative importance of each check to the pillar's headline score. A check at weight ten contributes ten times more than a check at weight one when computing the score. Weights are calibrated against impact on real visibility movements observed across thousands of brand scans, updated quarterly as the impact analysis evolves consistently across the index.
Why are some checks AI judged and others rule based?
Rule based checks like meta description length are deterministic, fast, and cheap. AI judged checks like does the first paragraph answer the implied question require reading the page content in context, which a model does better than a rule. Both kinds are documented in the catalogue with their type clearly labelled for transparency.
Can I exclude checks from scoring?
Yes. Each row carries a pause toggle. Paused checks are still run and reported but excluded from the pillar score calculation. Useful for checks that do not apply to your category, like FAQ schema on a brand site that legitimately does not have FAQ pages, or shopping pack checks on a non commerce brand without misleading the score.
How do I know which checks are worth fixing first?
Sort by weight descending, then by failing page count descending. The top of that sort is your highest leverage work. The Tech audit report PDF and the pillar Score tabs auto generate this prioritised list every scan, but the catalogue lets you confirm the recommended order against your own context and sprint capacity before committing time.