Trust in a site like this one should not be granted on the strength of a promise. It should be granted, if at all, on the strength of things a reader can check. This page sets out both: how the site is paid, and which parts of our work do not depend on taking our word for anything.
How the site earns
Some pages carry a link or a button that opens a registration form operated by a third party, and the site can be paid when a reader uses it. That is the commercial model, stated plainly. A short disclosure appears on every file above the text, because a disclosure buried in a footer is worth nothing to somebody who stopped reading halfway down.
Which platform a visitor is shown by that form is decided by the operator of the form according to the visitor's country and how they arrived. We do not select it, and a file on this site is never written to steer a reader toward a particular name.
Where the conflict sits
It sits exactly where you would expect. A site paid when readers register has a financial reason to make registration look attractive, and that reason is present on every page whether or not we act on it. Pretending otherwise would be the least trustworthy thing we could publish.
Our answer is structural rather than rhetorical. Every file ends at the same verdict scale, most of them at unknown, and the unknown verdicts are not softened to make the sign-up route more appealing. There is at most one call to action in a file, it never appears before the reader has seen the verdict, and it asks a reader to look at what a form collects rather than urging them to deposit.
What we do not sell
We do not sell a verdict, a place in a list, the removal of a file, a friendlier summary or a promise of coverage. No platform has been offered the chance to review a file before publication, and none has editorial influence over the sections describing what could not be verified.
We do not run a ranking table, which matters here because rankings are the standard product this industry buys. A table implies measurement, measurement implies data, and the data is exactly what these platforms do not publish.
What you can verify without us
Every check described in a file can be repeated by a reader in about twenty minutes. The registers we use are public, free and linked at the foot of every file. If we say a search under a brand name returned nothing, you can run the same search. If we say no company was named in the terms, you can open the terms and look.
That is the part of our work that does not require trust, and it is deliberately the largest part. Where we ask for trust is narrower: that we ran the checks on the date printed, and that we did not leave out something we found. Those are claims about conduct, and the only guarantee we can offer is that corrections are published with their dates and that the contact page tells you how to send one.
What would make us wrong
A file can be wrong in two directions. It can miss a company that was disclosed somewhere we did not look, and it can carry a check that has gone stale because circumstances changed after the date shown. Both happen in work of this kind, and both are fixed the same way: send the source, and the file is rewritten with a dated note of what changed.
We do not claim to be comprehensive, impartial by nature or incapable of error. We claim to publish a dated record, to name the limits of it in the file itself, and to correct it in public when it is shown to be wrong.