Where review content comes from
Every review is built on the same data snapshot: the weekly crawl, the scripted browser test and the verified feature checks described in our methodology. Measured claims — resolutions, popup counts, load times, prices, free-content share — may only come from that snapshot. If our pipeline has not verified a feature, the review says so instead of guessing.
What we never publish
We do not publish paid reviews, sponsored rankings or advertorials. No site has ever been moved up or down a ranking by hand, and there is no mechanism in our pipeline to do it. We do not copy competitor reviews, and we do not publish claims about a site's catalog that we cannot attribute either to our crawl or to well-established public knowledge about that site.
Scores and updates
Scores are recomputed from fresh crawl data on a weekly cycle. When a site changes — quality drops, ads get heavier, a feature disappears — the score follows the next crawl, and the written review is updated on a slower editorial cycle. Every review page shows the date of the test run its numbers come from.
Estimated versus measured data
Not everything can be measured directly. Where a number is estimated rather than observed — for example freshness on a site that hides upload dates — the review and the methodology page say so in plain words. We would rather flag lower confidence than present a guess as a measurement.
Corrections
If a review states something factually wrong about your site, use the contact page. We check disputed claims against the next crawl and correct the text if the claim does not hold. Corrections change the text; they never change the score.