Findings · 2026-09-17 · By VSNARY | Emmanuel Orta
Correction: our paywall said sites with findings were clean
For about forty minutes, a locked report told unpaid readers there was nothing to fix, and its spreadsheet export said there were no findings. Hiding findings was intended; claiming there were none was the bug.
When CrawlCheck put the list of findings behind a paywall in September 2026, the unpaid view of a report showed an empty list, and two places that read an empty list as "nothing wrong" said so. For about forty minutes the page printed "Nothing to fix" and the spreadsheet export wrote a "No findings reported" row for sites that had findings. Withheld results now say they are locked and give the real count. A clean site still reads as clean.
What readers saw #
The grade, the score, which stage fails and how many findings a site has stayed free. Which findings, their evidence and the files that fix them moved behind a licence. For about forty minutes after that change, a reader who had not paid could open a report for a site with several serious defects and read Nothing to fix. Every measure that could be scored on this page passes. The spreadsheet export said the same thing in a row of its own. It was caught before the change was announced anywhere.
What was wrong #
The locked view showed an empty list of findings. Two places, the report page and the export, already had a correct message for a genuinely clean site, and they printed it whenever the list was empty. After the paywall, empty had two meanings and neither place could tell them apart. Hiding the findings was working as intended; replacing them with a false claim was the defect.
What changed #
| Where | With an empty list, before | Now |
|---|---|---|
| Report page | Nothing to fix. Every measure passes. | The real count of findings, with locked rows |
| Spreadsheet export | No findings reported | Findings withheld, with the real count |
| Answers to AI assistants | An empty list and nothing else | The count and a note that the list is withheld |
| A genuinely clean site | Nothing to fix (correct) | Nothing to fix (unchanged) |
The reply to AI assistants mattered most. A person can see a grade of C above the words nothing to fix and distrust one of them. A model handed an empty list with nothing beside it cannot.
The lesson for your own site #
Anywhere a system shows a zero, an empty state or a clean result, ask what else could produce that output. A zero that means none and a zero that means not measured or not shown look identical, and when a second cause appears the original sentence becomes false without anyone editing it. The test that catches it is dull: render the locked view of a record you know has findings and search the output for your clean-site message. We had the same lesson from three rows that passed because nothing was measured and proofs that said pending for 26 days. The free scan reports three outcomes per section, passed, failed and could not be measured, and never collapses the third into the first; what a licence opens is on the pricing page.
The main product
Found this on your own site? We fix it for $100.
Scan free to see where you stand. The fix is one site, every finding implemented and re-measured, with a sealed before and after.
Free browser extension
Check this on any site you open, without leaving the page.
CrawlCheck for Chrome grades every site you visit A–F and shows, from your own browser, what search engines and AI assistants are actually served. Free, no account.
Questions this post answers
Is the CrawlCheck scan still free?
Yes. Running a scan, the grade, the score, which stage fails and how many findings a site has are free on any domain with no account. A licence or the fix service adds which findings, their evidence and the files that correct them.
How long did locked reports say there was nothing to fix?
About forty minutes, between the paywall shipping and the check that caught it. It was found before the change was announced anywhere.
Why not remove the nothing-to-fix message altogether?
Because it is correct for a genuinely clean site. The fix was to make a withheld list say it is withheld, so the clean message appears only when nothing was found.
How do you test for a false clean result?
Render the locked or partial view of a record you know has problems, and check the output never contains the message reserved for a clean result.
Related findings
Comments
Comments are read before they appear. Nothing is published automatically, and no account is needed.
Writing about this? Facts, live figures and marks — every number on that page is dated and traceable to a scan.