Organizing a Confluence hoard, or, does this page spark joy?
summary
Reorganizes a 33‑space Confluence estate of ~20,000 pages into one tall, narrow space of ~11,000 pages, deleting 5 spaces, stripping 13 and archiving the rest; implements team landing pages with scoped search and excerpted summaries to surface content; builds a joy detector using a weighted view count (avg daily views ÷ days since last viewed + 1, scaled) and produces heat‑maps to prioritize migrations; documents constraints (unique page names, Confluence API) and shows the system improves findability and maintenance while leaving some cruft behind.
AI-written summary · about
chapters (26)
- 0:00Abi Sutherland
- 0:42The Confluence problem
- 1:39Marie Kondo Organizing consultant and author
- 2:13Commit yourself to tidying up.
- 3:12Imagine your ideal lifestyle.
- 3:57User problems as desire paths
- 6:10Facilitating user/content meet cutes
- 8:03Users, needs and knowledge
- 9:34The biggest constraint
- 11:02The other big constraint
- 13:11Tradeoff: Hierarchies and menus
- 15:46Tradeoff: Taking back control
- 18:24Control overlap: team pages
- 20:03Limited search scope
- 21:04Local updates + excerpts
- 22:18Writing browsable page titles
- 23:37Finish discarding first.
- 24:18Tidy by category, not by location. 5. Follow the right order.
- 24:56Ask yourself if it sparks joy.
- 25:40Creating a joy detector 1. Identify examples of pages to keep
- 27:02A series of experiments
- 27:57Weighted view count
- 29:07Heat joy maps in Excel
- 31:13Where am I now?
- 32:28The new space
- 33:24Where do I go from here?
related talks
40mAll together now: Turning technologists into documentarians
38mAdventures in setting up a knowledge system for a research group
37mDrop the Docs, Get Back to What’s Important!
17mVisualization to the rescue
15mDeveloper portal
39mUnblocking Backlog Jams
31mHow We Deprecated 500+ Articles: Cleaning Up Documentation at Scale
37mTaking a step back: Analysing your documentation