Certainty Without Evidence: A Confusion Audit of This Site

LinkedInXEmail

I preach certainty for a living. So I ran a Confusion Audit on my own site.

The thesis that came back was uncomfortable and accurate: claims and proof lived on different pages. Hard percentages sat in the hero of case studies. Underneath, Challenge / Solution / Results headings were empty—or repeated the same Slack/Railway sentence three times. Agents were invited into /case-studies/ by robots.txt. Humans who clicked a number found nothing to believe.

For a practice called Architect of Certainty, that is the most expensive failure mode. And it was fixable, because the work behind the numbers already existed—it just was not attached to the claim.

What the audit found

  • Empty sections under metrics — Lexus, Norton, and TeamOne stories published figures without method paragraphs.
  • Catalog drift — sitemap and index disagreed (duplicate Jira / Playwright URLs; an orphaned backend brief).
  • Substance unlinked — the Certainty Ladder deep dive existed, but About pointed at a one-line teaser.
  • Workshop pages indexed/design-system and a stub /experiments sat in the public sitemap.
  • Noun confusion — "Impact" named a nav pill, a homepage section, and a case-studies H1 pointing at two different URLs.

What we changed

  1. Proof first — empty Challenge / Solution / Results never render; CMS empty headings fall back to filled static copy; metrics require a method paragraph (what was measured, window, baseline).
  2. One catalog — duplicate TeamOne URLs redirect to a single canonical; published slug count matches the index.
  3. Workshop off the crawl — design-system and experiments are noindex; design-system sits behind the same auth as CRM/portal.
  4. Routing — Case Studies means the page; Impact means the homepage section; About's Certainty Ladder card lands on the deep dive; /mcp is in the footer.
  5. One tool list — handler, well-known manifest, and /mcp share MCP_TOOL_NAMES.

The rule I am keeping

No metric ships without a short method paragraph. If I cannot say how the number was derived, the number comes down. That is the Confusion Audit applied to my own marketing—before I ask anyone else to trust the frameworks.

Related: The Certainty Ladder deep dive · Case studies · Confusion Audit.

Get the next note

Same list as essays — no third inbox.

One field. Exportable list I own. Unsubscribe anytime.

Back to My Notes