Documentation Decay Audit (QA)

Documentation Decay Audit (QA)

1. 01

   Pull the list of pages that no client deliverable has linked to in 90 days

   Sort by last-modified date and flag anything untouched since the previous quarter close. A page nobody opens is either dead weight or a search failure, and both cost onboarding hours.

2. 02

   Cross-check each flagged page against the tool that owns the underlying process

   If a runbook describes a Slack workflow, confirm the workflow still exists in Slack. Slite and Guru both monitor connected tools for this drift, but the human check catches renames and permission changes the connectors miss.

3. 03

   Interview the two people who actually run the process the page describes

   Ask them to walk through the steps out loud while you follow along on the page. Every place they deviate is a defect, not a preference.

4. 04

   Mark each page as current, rewrite, or archive, and assign an owner with a date

   Unowned pages decay fastest. One named person per page, with a review date no more than 180 days out.

5. 05

   Rewrite the top five pages by traffic before touching the long tail

   Agency onboarding paths usually run through a handful of pages. Fixing those five cuts new-hire ramp time more than a full-library sweep.

6. 06

   Delete or archive anything that duplicates a page in another system

   Two versions of the same client onboarding checklist is the failure mode that makes teams stop trusting the knowledge base entirely.

7. 07

   Log the audit date and the pages changed in a single changelog entry

   This is the artifact you show a client when they ask how delivery consistency is maintained across a retainer.