Misadventures in SEO: How to self-sabotage your documentation
summary
Documentation site with ~25 historical versions loses search for core term bpmn after switching to a Google programmable search engine and accidentally applying noindex to older versions. Google crawls older copies, picks an old canonical, and the site and site-search return no results found for months. Fixes include submitting a sitemap with only the latest version, adding rel=canonical hints from old to new, and eventually removing noindex; after the release recrawl ~99% of pages index correctly and users can find content again.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 1:06About speaker
- 2:31DX engineer role
- 4:36Documentation search timeline
- 7:51C7 search built
- 9:14Versioning duplication problem
- 11:24Search engines duplicate content
- 14:14No-index mistake impact
- 17:05Failed fixes attempts
- 18:50Final fix experiment
- 21:42Multi-version docs takeaways
- 23:53C8 lessons learned
- 28:07AI search future
- 29:54Key takeaways
- 30:37Q&A
- 31:20Talk illustrations process
- 33:08Downtime during fix
- 33:50Search terms insights
- 34:30Google Search Console debugging
- 35:14SEO black magic research
- 36:16Mistake reflection
- 37:44Docs platform migration
- 38:48Non-canonical sources
- 39:32SEO and AI search
related talks
45mOptimizing SEO for Technical Documentation
31mMaking documentation discoverable in search engines
51mSEO Gone Wrong! - Learn Risk Free from Terrible SEO Mistakes!
36mWeird Hidden Things on a Website That People Don't Check
36mSeo Insights
32mSEO for technical articles
15mSEO for the New Millenium
16mConducting SEO Testing - The Live Smell