Checklist: documentation audit
Reference
A checklist for auditing an existing documentation set - finding orphaned pages, duplicate content, broken links, and stale information architecture.
Reachability
Section titled “Reachability”- Every published page is linked from navigation, a landing page, or internal cross-links - a page with zero inbound links is effectively invisible, even if it’s technically online. Check your site’s build output or sitemap against your navigation config directly; don’t rely on memory.
- Every navigation entry points to a page that still exists and still matches its label.
Duplication
Section titled “Duplication”- No two pages cover the same TopicThe smallest autonomous, titled unit of information in DITA XML, typed as concept, task, or reference and assembled into documents via ditamaps.View in glossary → with the same intent - if they do, decide which is canonical and redirect or fold the other into it.
- Overlapping pages that serve genuinely different audiences (a quick-start vs. a full reference) are cross-linked, so a reader who lands on the wrong one can find the right one in one click.
Freshness
Section titled “Freshness”- Pages referencing a specific product version, tool version, or external service have been checked against the current reality.
- Screenshots and command output reflect the current UI/CLI, not an old one.
- Dated caveats (“as of version X…”, “this is temporary until…”) are revisited - either resolved or reconfirmed as still relevant.
Structure
Section titled “Structure”- Page titles are meaningful out of context (a reader arriving from search sees the title before anything else).
- Heading hierarchy is consistent across similar page types.
- Content-type metadata (concept/task/reference) is present and accurate, if your platform supports it.
- Internal links resolve (a broken-link checker run against a full site build catches most of these mechanically).
- External links still point where they claim to - link rot on external references is easy to miss without checking.
Output
Section titled “Output”- Findings are written down with a concrete next action per item (redirect, merge, update, delete, or “reviewed, no action needed”) - an audit that ends in a list of problems with no owner tends to repeat itself next year unchanged.
After the audit
Section titled “After the audit”- To check individual pages as they are fixed or rewritten, use the documentation quality checklist.
- If the audit shows that problems come from the toolchain itself (binary sources, manual publishing, no review trail), see the docs-as-code adoption checklist.