Experiment
Real SEO™ Experiment No. 001 ... The Discovery Channel Hypothesis
Testing whether a domain built to Real SEO™ infrastructure standards earns AI citations before it earns meaningful Google traffic ... and what that result reveals about the discovery channel split.
Google's role in 2026 has shifted: it captures high-intent, transactional queries while conversational AI now handles exploratory, research-phase discovery. If true, a new domain (NaturalHerbLibrary.com) built specifically for AI citability ... structured schema, machine-readable endpoints, deep reference content, tight entity signals ... should earn organic AI citations before it earns meaningful Google rankings. This is a testable claim. The six branches below each isolate one of the assumptions this thesis depends on.
What is being tracked.
How the test is structured.
NaturalHerbLibrary.com launched Q1 2026 as a clean-slate test property in the supplements and longevity niche. Zero backlink equity at launch. No paid promotion. No outreach or link-building of any kind. Architecture built to Real SEO™ standards: machine-readable content layer at /ai/catalog.json, FAQPage and Article schema markup throughout, deep reference articles on longevity and supplement topics, consistent author entity attribution across every page.
This experiment runs six active branches ... each one drawn from the six questions the seed article posed. Each branch is a standalone article under active investigation, updating independently as data comes in. The main experiment closes when a pattern is clear across branches ... not at an arbitrary calendar date.
What has happened so far.
[April 2026 ... Baseline] Launched. Zero AI citations recorded across ChatGPT, Perplexity, Claude, and Gemini. Structured data validated. Content indexing underway.
[May 2026 ... Update] Six branch articles published and indexed. No organic AI citations recorded.
[July 2026 ... Update] Raw logs showed heavy multi-vendor crawler attention. Search Console was already showing clicks and impressions. A crawler request was treated correctly as a precondition for possible use, not proof of a citation.
[September 11, 2026 ... Result] Google visibility arrived first. From July 29 through September 8, NaturalHerbLibrary.com earned 15,562 Search Console impressions, 11 clicks, visibility across 1,046 pages, and a weighted average position of 21.4. The latest 28-day window alone recorded 6,949 impressions across 742 pages. No confirmed AI citation was logged during the experiment.
The warehouse also records 23,954 successful crawler content requests since the July update, but that does not change the order of arrival. Crawling is not citation.
What would support or challenge it.
Status: hypothesis challenged.
NaturalHerbLibrary.com did not give me the result I predicted. It earned measurable Google discovery before it earned a tracked AI citation.
The experiment still proved that crawler attention can run far ahead of clicks. It did not prove that AI discovery arrives before Google discovery. On this property, Google won that race.
That is enough to close Experiment No. 001 and carry the better question forward: not which channel always comes first, but which combination of crawler, impression, query, human, and citation signals shows exposure building before clicks become obvious.
Keep Following the Tests
Move from this open thread back into the full experiment library.