I Audited 100 Pages Cited by ChatGPT — Here's What They All Have in Common

What I personally did

• Designed the research question and the sampling method.
• Collected and graded the data personally — no outsourced or scraped third-party dataset.
• Wrote every word of the published piece and maintain it after publication.

How the numbers were measured

1. 600 buyer-intent prompts across 12 niches, each run three times in fresh sessions to control for response variance.
2. Cited URLs logged, deduplicated to 100 unique pages, then crawled and graded against 14 fixed variables.
3. Wikipedia, Reddit, and YouTube excluded so the findings stay actionable for commercial and editorial sites.

Evidence you can check

• Prompt set and scoring rubric

  The 14 scored variables and the niche breakdown are published inside the article itself.

  Verification: Method is described in full in the article — the study is reproducible with your own prompt set.

  Open source

• Anonymised dataset

  URL-level schema profile and citation counts.

  Verification: Released to newsletter subscribers; available on request for researchers who cite the source.