TL;DR
The problem: A Confluence knowledge base that only grows ends up with outdated, unread and duplicated pages sitting next to the ones people rely on.
The fix: A data-driven Confluence content audit combines page metadata (topic, content type, owner, review cadence) with real usage data, so you review by evidence instead of gut feeling.
Category view: Tagging every page with the same metadata fields lets you spot whole topics that nobody reads, not just single pages.
The routine: A monthly review of content and space reports turns cleanup into a habit instead of a once-a-year project.
The payoff: Every archive, merge or promote decision is backed by data you can point to.
A Confluence content audit works when you can answer two questions for every page: what it’s about, and whether anyone reads it. Tag each knowledge base page with shared metadata fields in the page template, then compare those tags with view and contribution data each month. Ignored pages and topics become visible, so you act on evidence. Content performance monitoring is this recurring review of usage data against content structure, so maintenance effort goes where readers are.
Why does a Confluence knowledge base fill up with pages nobody reads?
Knowledge bases grow by addition. Teams create pages for every process, policy and how-to, but rarely remove anything, because nobody knows for sure which pages still help someone. After a few years, hundreds of pages compete for attention in search and navigation, and outdated pages look exactly as authoritative as current ones.
The real gap is visibility, not effort. Without usage data, a knowledge manager can only spread maintenance time evenly across every page or guess what to prioritize. Both options waste time, and readers who land on an outdated page start to distrust the whole space.
| Without a data-driven audit | With a data-driven audit |
|---|---|
| Maintenance time spread evenly across all pages | Time goes to the pages people actually read |
| Archive decisions based on opinion | Archive decisions backed by view data |
| Outdated pages stay until someone complains | Ignored pages are caught in the monthly review |
| Clutter erodes trust in the space | What remains has been validated by real usage |
What does a content audit look like in practice?
Meet Jamie, a knowledge manager responsible for a knowledge base of 812 pages. Jamie’s question is simple: which of these pages still earn their place? Before the audit routine, there was no reliable way to tell a well-used guide from a page nobody had opened in a year.
Here’s Jamie’s monthly review in 46 seconds:
Jamie’s monthly routine now runs in five steps:
- Tag at creation: Every knowledge base page carries metadata fields for topic, content type, owner and review cadence, built into the standard page template.
- Let usage happen: Employees use the knowledge base day to day, reading some pages often and others rarely or never.
- Review the reports: Once a month, Jamie opens the content and space reports for the knowledge base space.
- Cross-check by category: Jamie compares low-view pages against their metadata and finds that an entire topic category is being ignored.
- Act and repeat: Jamie archives or merges outdated pages in that category, promotes high-value pages in navigation, and repeats the review next month.
How do you set up metadata for a Confluence content audit?
Metadata gives every page the same set of labeled fields, so you can group and report on pages by category rather than one at a time. The key is to embed the fields in the page template. When metadata is part of the template, every new page is audit-ready as soon as it is published.
Four fields cover most knowledge base audits:
| Metadata field | What it answers | Example values |
|---|---|---|
| Topic | Which subject area does this page belong to? | Process, Legal, IT |
| Content type | What kind of page is it? | Policy, How-to, Reference |
| Content owner | Who is responsible for keeping it accurate? | The named page owner |
| Review cadence | How often should it be checked? | Annually, Bi-annually |
A metadata overview table then lists all articles with their topic, owner, last review date and status in one place. Filtering that table by a single field, such as one topic or one owner, shows a whole slice of the knowledge base at once.
How do you find unused pages in Confluence?
Usage analytics show how often each page is viewed, by how many people, and how many contributions it gets over a chosen period. Sorting pages by views quickly separates the pages driving value from the ones nobody opens. Pages with consistently high views are candidates to promote; pages with few or no visits are candidates to review.
Useful signals for each page in a monthly review:
- Views and unique viewers: Is anyone reading this page, and how many different people?
- Contributions and edits: Is the page still being maintained, or has it been untouched for months?
- Last viewed and last modified: When did someone last open or update it?
- Trend over time: Is interest growing, stable or fading?
Low views alone don’t prove a page is useless. A rarely needed policy may still be essential. That is why the metadata cross-check matters: it tells you whether a quiet page belongs to a topic that should be quiet, or to a topic people have stopped trusting.
How do you turn audit findings into archive, merge or promote decisions?
Once usage and metadata sit side by side, each page falls into a clear action. Exporting usage data as CSV, or pulling it through an API, lets you join it with metadata tags in a spreadsheet and spot ignored categories rather than reviewing pages one by one.
| Usage pattern | Metadata context | Action |
|---|---|---|
| High views, recent review | Core topic | Promote in navigation and search |
| Low views, overdue review | Topic still relevant | Update and assign to the owner |
| Low views, duplicate content | Same topic as a better page | Merge into the stronger page |
| Low views across a whole category | Topic no longer in use | Archive the category |
Repeating this review every month keeps the knowledge base “well-used” instead of letting it slowly go stale. Each cycle is smaller than the last, because the clutter doesn’t have time to build up again.
Key takeaway
A healthy knowledge base is not the one with the most pages, but the one where every page has earned its place. A Confluence content audit built on metadata and usage data turns maintenance from guesswork into a short, repeatable monthly routine.
The Meridian Knowledge Structure solution brings both halves together: Metadata for Confluence for consistent page fields and Viewtracker – Analytics for Confluence for content and space reports.
Want to see it in your own Confluence? Book a demo.
Frequently asked questions
Couldn’t find what you need?
A monthly review works well for most knowledge bases. It keeps each review small enough to finish in one session and catches ignored pages before they mislead readers. Pages with a stricter review cadence, such as policies, can be checked on their own schedule using the review cadence metadata field.
Yes. Viewtracker’s space reports list every page in a space with views, viewers, contributors and last viewed date for a chosen period. Sorting by views shows the least-visited pages first. Combining that list with metadata tags reveals whether low traffic affects a single page or an entire topic category.
Not automatically. Some pages are rarely needed but still essential, such as an emergency procedure or an annual policy. Usage data tells you where to look; the page owner decides whether the content is still accurate and needed. Archiving, merging or updating are often better outcomes than deletion.
Labels are free-text tags, so spelling and naming drift between teams over time. Metadata fields use defined values in a page template, such as a fixed list of topics or content types. That consistency lets you group pages reliably by category and report on whole topics, not just individual pages.