Three kinds of judgement, kept apart on purpose: what users say, what we say, and what the public record shows. Every number on Benchmarked can be traced to one of them.
Calculated by Benchmarked from eligible, published reviews. It is never typed in by a person and cannot be bought.
Source: eligible published reviewsOur own structured assessment. It sits beside the user score, and the two are never averaged into one number.
Source: Benchmarked editorial reviewFactual and regulatory claims, each with the issuing body, reference, source, evidence and the date it was last checked.
Source: regulators and official registersEvery rating is a whole number from 1 to 5, always shown with its word so it never depends on colour or stars alone.
A product needs a minimum number of eligible, published reviews before any user score appears. Until then we say so plainly, and show how close it is. The threshold is set by Benchmarked and currently starts at 10.
We never show a bare “Regulated ✓”. A verification record names who issued it, where we checked it and when. We recheck it over time.
Sponsored placements are always labelled “Sponsored” and never change a score or ranking.
Where we may earn a commission we say so on the page. It never affects ratings.
Representatives can report reviews for moderation, like anyone else. They can never write, edit, hide, delete or reply to them.
AI helps flag spam and summarise themes. It never writes reviews, chooses ratings or changes scores.
No. Scores are calculated only from eligible reviews. Commercial relationships cannot change a user score, an editorial score, a verification status or a ranking.
It must be published after moderation, be one active review per person per product, and not be flagged as spam, duplicate or manipulation. Verified use may carry more weight in future.
No. Signing in with Google or Facebook confirms an account, not that you used the product. Product-use verification is a separate check.
From eligible review data across the company’s products, with care not to average unlike categories as if they measured the same thing.
