Editorial standards

From fitting and configuring these locks on real doors, and from dealing with them afterwards — the app, the account model, the failures. That is genuine first-hand experience and it is also a limitation: it tells you how a lock behaves in service, not how it performs under attack or over ten thousand cycles.

So we publish no numeric scores. A score implies a repeatable measured methodology, and dressing an informed opinion up as a measurement would be a false claim about how the judgement was reached.

Not as balance for its own sake — because a product with no listed drawbacks has not been assessed, it has been advertised. If we cannot say where something is the wrong answer, the piece is not finished.

Reviews we did not write. Testimonials we cannot attribute. Author names belonging to people who do not exist. Specifications copied from a retailer without checking. Claims of accreditation, partnership or approval we do not hold.

Disclosed on the page they apply to, above the link rather than at the foot of the article. No coverage is for sale, and no manufacturer sees a piece before it is published.

Tell us through the contact page. If we are wrong we will fix it and note that the piece was changed, rather than editing silently. Pieces revised materially carry an updated date at the top.