Why do we publish the claims we got wrong?
Every article below began as a planned piece with a claim already in the title. Before writing, we tried to source the claim. Three of them failed, and the article became the correction rather than the original plan. Each one is published here with the article it became.
That is worth publishing rather than quietly binning, for two reasons. Bad claims in this industry propagate because nobody checks them, and an agency that shows its working is easier to verify than one that shows its conclusions.
What did we check, and what did we find?
Six checks are published so far. Two are platform changes, each carrying the date the platform changed. Four are claim checks, where we went to the primary source before writing. The table lists every one, the kind of check it was, and what the check found.
| What we checked | Kind of check | What we found |
|---|---|---|
| FAQ rich results are gone | NEWS · MAY 2026 | Google removed them on 7 May 2026 with no carve-out. What changes, and what does not. |
| The LSA Money Back Guarantee is gone | NEWS · OCTOBER 2025 | Discontinued 20 October 2025. The badge was replaced; the protection was not. |
| Does llms.txt actually do anything? | CLAIM CHECK | We checked what Google published. The answer is no, and the file has been sold to a lot of people. |
| Does review recency beat review count? | CLAIM CHECK | We went looking for the three-month rule and could not find a source for it. |
| What AI Overviews actually did to clicks | CLAIM CHECK | The widely quoted 58% figure is real, and narrower than the people repeating it think. |
| Do AI search visitors convert better? | CLAIM CHECK | One dataset says yes by a wide margin. It is a software company measuring its own website. |
How often does this page update?
This page updates when something changes or when a claim needs checking, rather than on a schedule. We would rather publish six checks a year that mattered than one a month that did not.
Where a platform changed something, the entry carries the date it changed. Where a claim failed verification, the entry says what we looked for and where we looked.
What is missing from this page, and why?
Nothing on this page is estimated. Where a claim could not be traced to a primary source, the failure is published instead of a softened version of the claim. The standing disclaimer below runs on every Mindflow Marketing page, and the full method is linked with it.
Rankings are never guaranteed. Anything we could not trace to a primary source is absent from this page, not estimated. The audit runs the same six layers described on the pricing page, and Share of Answer is scored quarterly.
Fair questions about this page
Why publish claims you got wrong?
Because they were about to be published as right. Showing the check is more useful than showing only the conclusions, and unchecked claims in this industry spread precisely because nobody publishes the correction.
How often does this update?
When something changes or a claim needs checking, rather than on a fixed schedule. A cadence promise we could not keep would be the wrong thing to make on a page about verification.
Do you test things on client sites?
No. Nothing here is an experiment run on a client's site. These are verification passes against primary sources, and where we run measurements they follow the published protocol.
Start with a measurement, not a claim
A free Visibility Check runs twelve real buying questions across Google and AI. You keep the findings.
Get my free Visibility Check