How We Research and Review
Recommendations are research-based assessments unless a page explicitly records genuine hands-on evidence.
A repeatable source and claim process keeps product selection, cooking guidance and corrections reviewable.
Source priority
Official UK safety authorities come first for safety. Manufacturer specifications and support documents support product facts and appliance-specific guidance.
Product selection
We compare capacity, external dimensions, controls, cleaning, basket format and suitability for the page intent. We do not invent scores, prices or ownership experience.
Conflicts and uncertainty
Conflicting figures remain visible in the source ledger. We separate models, cuts, sizes and starting states and use qualified ranges instead of false precision.
Testing and images
Hands-on language is prohibited without recorded evidence. Generated editorial images are labelled in the media registry and require human approval before live use.
Updates and corrections
Material claims have access and review dates. Safety concerns and specification corrections receive priority through the corrections route.
How a product fact enters a guide
A candidate begins with an exact UK model identifier. Capacity, external dimensions, power, cooking functions, included parts and cleaning instructions are recorded separately so that a figure from one variant is not silently applied to another. The record includes the source URL, access date and the wording or field that supports the fact.
When sources disagree, the conflict is not solved by choosing the most convenient number. The guide checks whether the values describe packaging, usable basket space, a different colour variant or an earlier model revision. If the difference remains unresolved, the public wording stays qualified or the claim is withheld.
How recommendations are formed
A research-based assessment matches verified characteristics to a clearly defined user need. For example, a narrow appliance may be relevant to limited counter width, while a dual-zone layout may be relevant to cooking foods with different settings. That is an assessment of fit, not evidence that one machine cooks better.
Editorial labels must describe the decision criterion. Phrases such as best overall, cheapest, quietest or most efficient need suitable comparative evidence and cannot be generated from specifications alone. The absence of a live price, rating or performance score is deliberate where the evidence is not stable.
Review and update cycle
Safety claims are checked first when official guidance changes. Product pages are rechecked when a model is replaced, a specification changes, a source disappears or a reader reports a credible discrepancy. Review dates show when the page evidence was examined; dates are not refreshed merely to make content appear new.
Automation may assemble records, detect broken links and flag unsupported language, but it does not approve claims. A page reaches the publishing gate only after its facts, wording, links, layout and evidence boundaries pass review.
Last reviewed: 19 July 2026.