How I Review Mattresses
Every review here is built from four things: the published specification, the full warranty documentation, current UK pricing, and owner reports read and counted. Here is exactly how that works, including the limitations.
Every review on this site is built the same way, and I would rather tell you exactly how than leave you guessing. What follows is the full method, including the parts that are limitations rather than strengths.
The four things every review is built from
A mattress review is only as good as what sits behind it. Mine draws on four sources, in this order of weight.
The manufacturer's own specification
Depth, layer count, layer composition, spring count and gauge, cover material, firmness rating, weight tolerance. I take these from the brand's own product page and technical documentation, not from a press release or a retailer listing, because those two are where numbers get rounded in the brand's favour.
Where a brand refuses to publish something, I say so. Foam density is the clearest example. It is the single number that best predicts how long a foam mattress lasts before it starts to dip, and almost no bed-in-a-box brand in the UK publishes it. When a brand tells you a mattress has "premium memory foam" but will not tell you whether that foam is 30 kg per cubic metre or 60, the omission is the answer.
The warranty and trial documentation, read in full
This is the part most review sites skip, and it is where the real differences between brands live. I read the actual guarantee terms, not the marketing summary. What counts as a defect. What depth of sagging qualifies. Whether you need a specific base for the cover to remain valid. Who pays for collection if you return within the trial. Whether the trial excludes certain sizes or clearance stock.
A ten year guarantee that only covers sagging over 25 mm, measured unweighted, on an approved slatted base, with proof of a mattress protector, is a far weaker promise than a five year guarantee with none of those conditions. Both get advertised as a guarantee. Only one of them is worth much.
Owner reports, read and counted
For every model I cover, I read UK owner reviews across Trustpilot, Reddit, Mumsnet and the retailer listings, and I count what people actually complain about rather than summarising a general mood. Where I report a figure, it carries the sample size and the period it was drawn from, so you can judge how much weight to give it.
This is not a scientific survey and I will not pretend otherwise. Review platforms skew towards people who are delighted and people who are furious, with the quiet middle under-represented. What the counting does reliably show is the shape of a problem: whether complaints about a particular model cluster around heat, or edge collapse, or firmness being different from what the brand described. That pattern is useful even when the absolute percentage is not precise.
Ten years in the trade
I have spent over a decade in the UK mattress and bedding industry, across retail, distribution and brand partnerships. That gives me context a spec sheet does not: which construction choices are genuinely expensive and which are cheap to make and expensive to market, what a brand's move into high street retail usually signals, why spring counts climbed across the industry when the cost of adding springs fell, and which complaints turn into warranty claims rather than staying complaints.
What I do not do
I do not run laboratory tests. I do not have a pressure mapping rig, a climate chamber or a rolling durability machine, and I am not going to describe measurements I have not taken. Sites that publish surface temperature figures and edge deflection numbers without owning the equipment are inventing them, and that is worse than having no numbers at all.
What I do instead is judge construction against outcome. If a mattress puts a 25 mm comfort layer over a firm polyfoam core and owners consistently report shoulder pressure on their side, those two facts explain each other. That is analysis, not measurement, and I label it as such throughout.
How the scores work
Every mattress gets six scores out of ten plus an overall. Each one carries a sentence saying why, because a number without a reason is decoration.
| Criterion | What it measures | What pushes it down |
|---|---|---|
| Value | What you get for the money against everything else at that price in the UK | Thin comfort layers, undisclosed foam density, a price that only looks reasonable during a permanent sale |
| Support | Whether the construction should hold the spine in line for the sleeper type it is sold to | Too little depth for the firmness claimed, no zoning where zoning is needed, a support core that owners report bottoming out |
| Temperature | How likely it is to run warm, based on materials, airflow and cover | Deep memory foam with a sealed cover, "cooling" claims resting on a marketing name rather than a material |
| Edge support | Whether you can sit or sleep near the edge without it collapsing | All-foam construction with no perimeter reinforcement, which is most of the category |
| Motion isolation | Whether a partner moving wakes you | Continuous coils, tall unpocketed springs, a bouncy latex-heavy build |
| Trial and guarantee | The quality of the terms, not the headline number | Sagging thresholds, base requirements, protector conditions, charges for collection |
The bands mean something specific:
- 9 to 10. Best in class in the UK right now. Rare, and I expect to give few of these.
- 7 to 8. Genuinely good. A sensible buy for the right sleeper.
- 5 to 6. Compromised. Works for some people, has a real weakness you should know about.
- 3 to 4. Poor for the money. There is something better at the same price.
- 1 to 2. Avoid.
If every mattress on a site scores above eight, the scale is decoration rather than judgement. I use the full range. Several mattresses I cover score five on edge support, because most all-foam mattresses genuinely do collapse at the edge, and saying so is more useful than pretending the category has no weak point.
Why everything here is UK only
Sizes, prices, availability, trial terms and your legal rights all differ by country, and a mattress site that blurs the UK and US markets is useless in both. A king in the UK is 150 by 200 cm. In the US it is 193 by 203 cm. A "queen" is not a UK size at all. Every price on this site is in pounds, every size is a UK size, and every product is one you can actually buy here.
That last point matters more than it sounds. Casper withdrew from the UK in July 2020 and Leesa no longer sells here, yet both still appear in current UK best-of lists. If a page recommends something you cannot buy, nothing else on it is trustworthy either.
Your rights matter more than the guarantee
This comes up in almost every review, so it is worth stating once, clearly.
Under the Consumer Rights Act 2015, goods must be of satisfactory quality, fit for purpose and as described, and they must last a reasonable length of time. Those rights sit with the retailer who sold you the mattress, not the manufacturer, and in England, Wales and Northern Ireland you have up to six years from purchase to bring a claim through the courts. In Scotland it is five. A manufacturer's guarantee is a separate, voluntary promise on top of that, and it cannot reduce your statutory rights however the terms are worded.
In practice that means a mattress that dips badly at three years may well be a valid claim against the retailer even if the guarantee has conditions you cannot meet. After the first six months the burden of proof shifts to you, so an independent report helps. This is general information rather than legal advice, and if you need advice specific to your situation, Citizens Advice is the right starting point.
How often pages change
Prices are checked at least monthly and carry the date they were last verified. Model ranges are checked when a brand announces a change, and reviews of discontinued models stay live with a notice at the top rather than being deleted, because people still search for them and deserve an answer.
The date on every page is the date the content was genuinely last reviewed, not a date changed automatically to look fresh.
When I get something wrong
I will, at some point. When it happens, the page gets corrected and the correction is logged publicly on the corrections page with the date, what was wrong and what it now says. Nothing gets quietly edited.
If you spot an error, tell me. Brands are welcome to do the same, and several have. A brand pointing out a factual mistake gets it fixed. A brand asking for a score to be raised does not.
Questions people ask
Do you test mattresses in a laboratory?
No, and I never claim to. I do not own pressure mapping or durability equipment, so I do not publish measurements I have not taken. Reviews are built from published specifications, full warranty documentation, UK pricing and counted owner reports, weighed against ten years working in the mattress trade.
How are the scores calculated?
Every mattress gets six scores out of ten: value, support, temperature, edge support, motion isolation, and trial and guarantee. Each carries a sentence explaining it. The bands are defined, the full range gets used, and several well-known mattresses score five on edge support because that is what all-foam construction gives you.
Where does the owner data come from?
UK owner reviews on Trustpilot, Reddit, Mumsnet and retailer listings. I count what people complain about rather than summarising a mood, and every figure carries its sample size and date range. It is not a scientific survey, but the pattern of complaints is reliable even when the exact percentage is not.