Deferred 2026-05-09 09:42 ET Founder deferred all GEO changes pending raydata.co migration to Cloudflare. Re-audit on clean infra post-relaunch. Decision design preserved below for reference.

/decisions · 2026-05-09 · GEO foundation + bet

Should we unblock raydata.co for AI bots and run a 60-day GEO test?

Source: Brief 1 (foundation) + Brief 2 (evidence), both filed 2026-05-09 01:15 ET by overnight /deep-research.

Brief recap (skip if you read them at breakfast)

Brief 1 — foundation broken. raydata.co is double-blocked: Cloudflare's managed Disallow:/ robots.txt covers 10 AI bots AND the WAF ai_bots_protection: "block" 403s ClaudeBot. GPTBot / PerplexityBot / Google-Extended got 0 hits in 7 days. ChatGPT-User (live retrievals) hit 13 times but 5 were 404s on dead /p/ paths from a missing redirect after URL migration.

Brief 2 — bet unproven. Zero rigorous public studies trace LLM citation → revenue. Vendor case studies (Profound / Peec.ai / AthenaHQ) measure share-of-answer only. Base rate is brutal: 93% zero-click on AI search responses. Recommendation: don't invert the X-first content engine. Augment with a 60-day reversible measured test instead.

Combined call: fix the foundation reversibly + run a 60-day measured test before committing engine resources.

Pick a direction

Test design parameters

Defaults below are my recommended cuts based on the briefs. Adjust whichever feel off.

Top 5 default-on are the ones the briefs measured at 0 hits and want unblocked first. Bottom two are stretch-test additions.

30d120d

Rationale or context (optional)

Lock it in → iMessage me
Preview of the iMessage body (live) [pick a direction above to populate]