How we assess products
Most sites that compare products claim to have “tested for two weeks” every model they write about. When a catalogue holds dozens of them, that’s arithmetically impossible. We’d rather tell you what we actually do.
What we do
We verify specs at the source. Every figure in our spec sheets comes from the manufacturer’s official documentation, recorded and dated. Not copied from a press release, and not from another site that copied it in turn.
We apply the same grid to every model. The same standard spec sheet, the same criteria, from entry level to flagship. That’s what makes the comparison honest: without a shared grid, comparing two products means comparing two marketing pitches.
We track prices regularly. The list price is a commercial fiction: what matters is the real price, and above all the right price — the one below which you shouldn’t expect to go.
We cross-check against real-world feedback. Published independent tests, user reports on recurring flaws, reliability problems flagged after months of use. That’s where the gaps between the spec sheet and real life show up.
What we don’t do
We don’t make up test results. When we haven’t measured something ourselves, we don’t pretend otherwise. A site that publishes “test results” for dozens of machines it has never touched is lying to you — and Google eventually catches on.
We don’t fill data gaps. When a manufacturer doesn’t disclose a value, our sheet says “Not disclosed”. That’s information in itself: a manufacturer who won’t publish a figure rarely has good reasons to withhold it.
We don’t score what we can’t assess. Our scoring covers what is objectively measurable and comparable. It doesn’t claim to capture what only shows up over time: three-year reliability, the quality of after-sales support.
Why this method beats a fake test
A genuine test of a single product doesn’t tell you which one to buy among all the others. Our value isn’t in unboxing a machine: it’s in making sense of a lineup that’s deliberately hard to read, where the names all look alike and price doesn’t order quality.
That’s the work that reveals a model costing twice as much performing worse than another, or last year’s discounted model still being the best buy today. No two-week test reveals that. Only a grid applied across the whole lineup does.
Our independence
The most obvious commercial interest, for any retail-facing site, is to steer you toward the most expensive model. Our conclusions do the opposite whenever the figures demand it.
That’s the only independence test that matters: check whether a site’s conclusions run against its most immediate interest.
Frequently asked questions
Do you physically test every product?
We say so when we do, and we say so when we don’t. Our work is first and foremost a documented comparative analysis: specs verified at the source, prices tracked regularly, cross-checked against user feedback and published independent tests. When a manufacturer doesn’t disclose a figure, we say so instead of making it up.
Do your scores favour the most expensive models?
No. Our scores come from a public grid applied to the same criteria for every model. It regularly produces conclusions that are unfavourable to the high end — that is, to the products that would earn us the most.
How do you correct a mistake?
Every page carries its update date. A wrong figure is corrected in the catalogue, which updates the product review and the comparison table in one go — they read the same source and cannot contradict each other.
The best picks right now
Roborock Q10 S5+
Roborock Qrevo Curv 2 Pro
