Editorial Policy

A review is only worth reading if you know how it was made. This page is the standard every piece on Keen Reviewis held to — how a product gets picked, what “tested” means here, how a score is arrived at, when a review gets revisited, and what happens when we get something wrong.

If a published review ever departs from what's described here, that's a bug. Tell us at corrections@keenreview.comand we'll fix the review or fix this page.

We start from what people are actually trying to decide between, not from what a brand would like reviewed. In practice a product enters the queue because:

• it's a category leader people keep asking about;
• it's a new release that plausibly beats the current pick;
• it's cheap enough to be the value option and we need to know if it holds up;
• readers asked for it, or a current pick has started drawing complaints.

A brand cannot buy a place in the queue, and pitching us doesn't improve a product's chances. Where a category has an obvious market leader, we test it even when there's no affiliate program behind it.

Four rules apply to every category:

• We buy everything. No samples from brands. Every product is bought at retail, like you would.
• Real-world testing. Weeks of daily use, not a lab afternoon. We note how things hold up.
• Consistent scoring. Every product is judged on the same criteria for its category.
• We update picks. Reviews are revisited as prices change and new models launch.

“Real-world testing” means daily use over weeks, in the conditions the thing is actually for — a travel bag goes travelling, a blender makes breakfast every morning, a mattress gets slept on. Where a measurement is meaningful (battery runtime, noise, brightness, heat retention) we measure it and repeat it, rather than quoting the manufacturer's number. Where the honest answer is subjective — comfort, fit, how a fabric ages — we say so and describe who it suited and who it didn't.

We buy every unit at retail, at the same price you'd pay, from the same shops. That rules out the two most common ways review units get flattering: a cherry-picked sample, and the unspoken pressure of a loan you have to return.

Scores are relative to a category, not to the whole site: an 8.4 for a £40 kettle and an 8.4 for a £900 laptop each mean “strong against its own rivals at its own price”. Within a category every product is judged against the same written criteria, weighted the same way, so two reviews can be compared line by line.

Price is part of the score, which is why a cheaper product regularly beats a better one. A score is set from testing notes before commission rates are looked at, and it isn't adjusted afterwards.

We publish the bad ones. A product we bought, tested and wouldn't recommend gets a review saying so — those are the pages that make the recommendations mean anything.

Specifications, prices and claims are verified against the manufacturer's own documentation or a retailer listing before publication, and we link out so you can check them yourself. Where we cite someone else's testing, standard or teardown, we name and link it rather than absorbing it as our own finding.

Prices shown are what we saw at the time of writing, in the currency stated. They move constantly — always confirm at the retailer before buying.

We use software — including AI tools — for routine production work: transcribing testing notes, tidying grammar, generating draft structure, and formatting specification tables. We do not publish AI-generated verdicts, scores or testing observations. Every judgement on this site comes from someone who used the product, and a human editor is accountable for every published page.

Affiliate commission funds this site and never influences it — the mechanics are set out in our affiliate disclosure. Beyond that:

• Nobody who writes or edits a review is compensated on the commission that review earns.
• Anyone with a personal or financial connection to a brand steps off that brand's coverage, and we disclose the connection if the coverage runs anyway.
• Brands get no advance sight of a review, no approval over its content, and no opportunity to change a score before publication.
• We don't remove or soften a negative review because a brand, a retailer or an affiliate network asked us to. Losing a program is cheaper than losing the point of the site.

A review is a snapshot, and the market moves. We revisit a page when a successor model launches, when a price shift changes the value calculation, when long-term use turns up a durability problem, or when enough readers report a fault we didn't see.

Updated pages carry a visible date. Where a verdict changes materially — a pick demoted, a score revised — we say what changed and why rather than quietly editing it.

We get things wrong. When we do, the fix is fast and visible rather than silent:

• Send it to corrections@keenreview.comwith the page URL and what's wrong. Anyone can — reader, brand or competitor.
• We verify it against the source, and re-test where the claim came from testing.
• Factual errors are corrected promptly, typically within two business days of confirming them.
• Substantive corrections are noted on the page, so you can see what changed. Typos and formatting fixes are made without a note.

We don't unpublish a review to make a problem go away. If a page is withdrawn entirely, the reason stays at that URL.

Corrections: corrections@keenreview.com. Editorial questions and story ideas: hello@keenreview.com. Commercial enquiries — which are handled entirely separately from editorial — partners@keenreview.com. Or use the contact page.