Every claim about a licence number, a bonus percentage, a wagering requirement, a payout timeframe or a payment rail is checked against the operator's own terms page on the day of publication. The fact-checker opens the operator site in an incognito window and screenshots the clause that supports the sentence; the screenshot sits with the review draft in our shared document. When a page is republished during a quarterly refresh, the fact-checker repeats the exercise rather than copying over the previous screenshot. If a figure in a draft is not supported by a screenshot of a primary source, the fact-checker returns the draft to the author and the sentence is removed until a source can be found.
We prefer primary sources: the operator's own terms, the regulator's own register, the game provider's own press release, the payment-rail operator's own documentation. Where we rely on a secondary source — an industry publication, a trade conference announcement, an interview — we name it in the body of the review and link to the page we read. We do not cite anonymous forum posts, we do not reproduce rumours as fact and we do not treat press releases as reviews. For historical facts about an operator we accept the Wayback Machine as a stand-in for a page that has since been edited, and we link the archived version directly rather than paraphrasing it from memory.
Editorial independence sits at the top of the list: no operator buys a positive review or a higher position in a table, and no commercial team gets to approve editorial copy before publication. Readability is the second: an operator review that no one can read because it hides the key facts behind two thousand words of copy is not a review, it is marketing. Honesty about limitations is the third: we tell you when we have not been able to test a payment rail or a game category rather than inventing a result. Timeliness is the fourth: a review that lists a defunct welcome offer is worse than no review at all, which is why the quarterly refresh exists.