Research · 22 July 2026

The Fabrication Study: Graders That Score a Page They Never Fetched.

A dated, reproducible look at how the industry's free SEO graders return a "grade" for pages they demonstrably could not have read. The method is here; the anonymized evidence is attached to the published study.

The method — reproducible

Give a Grader a Page It Cannot Have Read. See If It Grades Anyway.

Step 1
A page they can't reach

Submit a URL behind a block, or one freshly created — a page no crawler could have fetched in time to analyse it.

Step 2
Watch for a score

Record whether the tool returns a numeric grade, letter, or checklist for content it had no way to read.

Step 3
Compare to an honest read

Run the same URL through an instrument that reports the block honestly. The gap between the two is the finding.

The receipts

Anonymized Captures — Naming Withheld Pending Approval.

By default the evidence is anonymized: the reproducible method stands on its own, and we render no legal conclusion about any named grader. Naming is a deliberate decision, not a default.

Figure — reserved

Anonymized capture: a grader returning a score for an unreadable page. Attached with the published, dated study.

Figure — reserved

Side-by-side: the same URL, read honestly — the block reported, no score invented. Attached with the published, dated study.

Two options are on record for the published version: A — anonymized (default; lower exposure, stands on method) and B — named (only on explicit approval; higher link value, higher exposure). Either way, every claim is observe → method → limits, and we render no legal verdict.

Also in Research

Studies & Measured Data

What Separates Page-One Local-Service Sites

Our flagship data study — measured on real sites, not guessed. The on-page signals top-ranked pages share more than the bottom (a privacy-policy link leads), the exact sample size, and the honest curiosities. Read the study

What Is a Page-One Ranking Worth?

A dated, sourced study of what page one is worth — click-through rate by position, where the clicks concentrate, and what one position is worth in visits and leads. Read the study

Is Your Website Visible to AI?

A technical guide to the reach → read → cite chain — the crawlers to know, what earns a citation, and how to check your own site today. Read the guide

Why Doesn’t ChatGPT Recommend My Business?

An ordered diagnosis of a silent absence — does an assistant know you, can it reach and read you, is there anything to extract, and does the wider web vouch for you — with free checks you can run today. Read the diagnosis

What AI Overviews Did to the Click Curve

Three named, dated studies on how AI Overviews change the click — laid side by side by what each actually measured, with how to read any CTR claim and a free way to check your own exposure. Read the study