How we work

Reviews

We buy what we review. When a publisher provides a code, we say so in the first paragraph — not in a footnote, and not in grey text at the bottom. A free code buys us nothing except time, and publishers know that before they send one.

We never accept conditions in exchange for access. No agreed talking points, no embargo on the score alone, no early look in return for a tone. If a publisher offers a deal like that, the offer becomes part of the story we publish.

Minimum fifteen hours before a score. Every review lists the hours played and the exact hardware, because a review from a flagship phone on fibre is a different review and you deserve to know which one you're reading.

Scores

Ten-point scale, and we use all of it. Scores below six happen regularly here. If a publication's scores all land between seven and nine, the scale is decoration.

A score is one person's judgement, and it's signed. We don't average opinions into a house verdict — you're reading a named human being, and you can decide over time whether you agree with them.

Sponsorship

Sponsored items say sponsored at the top, in the same size as everything else on the page. Sponsors don't see copy before it runs, don't get approval over anything, and never receive the subscriber list.

We won't take sponsorship from a company whose product we're actively reviewing that week. If a sponsor's product comes up in editorial, the sponsorship gets disclosed again, in the piece.

Affiliate links

Where a link earns us a commission, it's disclosed on that page. Commission never decides a score, never decides whether we cover something, and never decides which product goes first in a list. If we recommend a shop that pays us nothing over one that does, that's the job working correctly.

Corrections

When we get something wrong, we fix it and add a note at the bottom of the piece: what was wrong, what it says now, and the date. The note stays permanently. We don't silently edit a live piece and pretend the first version didn't exist.

Serious errors also get named in the next newsletter. If we were wrong in front of everyone, we'd rather be right in front of everyone too.

Sources

We say where numbers come from. When we test something ourselves, we describe the method well enough for you to repeat it. When we can't name a source — someone talking about their employer, usually — we say why they're anonymous and what they had to gain or lose by talking.

AI

No article, review or guide on this site is generated by an AI model. We use spellcheck and we use search. The judgement, the testing and the sentences are human, and if that ever changes we'll say so here first.