Most mattress review sites have never slept on a mattress

Search for a review of almost any UK mattress and you will get results. Scores out of ten, verdicts, star ratings, tables of pros and cons. Almost none of it comes from anyone who has slept on the mattress.

This is not a conspiracy, and for the most part it is not dishonest either. It is a business model with a particular set of incentives, and once you can see them the search results stop being confusing and start being predictable.

Some affiliate publishers do excellent work. A few genuinely buy and test products, say so, and show their working. This piece is not about them.

How the model works

An affiliate site earns money when a reader clicks a link and buys something. The commission is a percentage of the sale, typically a few per cent, paid by the retailer or brand.

Nothing about that is improper. It funds a great deal of useful publishing, and a site that discloses it properly and writes honestly is doing legitimate work.

The problem is what the model rewards. Commission is paid on purchases, not on accuracy. A page that persuades you to buy earns; a page that tells you to keep your current mattress for another two years does not. And since publishing a page costs almost nothing while buying a mattress to test costs several hundred pounds, the economics push hard in one direction: more pages, less testing.

The result is a large volume of content that has the shape of a review — a score, a verdict, a recommendation — without the substance of one.

In fairness, testing mattresses properly is genuinely hard. They cost several hundred pounds each, take weeks to assess honestly, are enormous to store and near-impossible to return once slept on. Even well-funded organisations manage only a few dozen a year. Nobody should pretend this is easy, or that a small publisher declining to do it is behaving badly. The objection is to presenting the result as though it were testing.

How to tell what you are reading

You do not need to guess. There are reliable tells, and they take about thirty seconds to check.

Look for evidence of the mattress existing in a room

A genuine hands-on review contains things that can only come from having the object: photographs of it in an actual bedroom rather than the manufacturer’s press shots, a stated period of use, a named person who slept on it, measurements taken, a description of unboxing it or getting it up the stairs.

Google’s own guidance for review content asks for exactly this — evidence of first-hand experience, quantitative measurements, and visuals documenting the work involved.¹ The absence of all of it is informative.

Read the disclaimer, not the verdict

Affiliate sites generally do disclose — and the better ones disclose clearly. What varies is placement and wording: often a line at the foot of the page, or on a separate “about” page, phrased in terms most readers skim past. Go and find it, and read it properly.

The most revealing ones say something close to: ratings are based on aggregated public review content, including reviews submitted elsewhere and evaluated by our team. Translated, that means the site did not review the mattress. It processed other people’s reviews and produced a number.

Ask where the customer reviews came from

This is the tell that most people miss, and it is the sharpest one.

A review site will often cite a large number of customer ratings — four hundred, five hundred — alongside its own score. If you go and look at the retailer’s product page, you will frequently find a very similar number there.

They are the same reviews. Someone searching for opinions “other than the shop’s own” ends up reading the shop’s own reviews, collected on a different website, with a score attached.

Count how many products the site has “tested”

Genuine testing is slow and expensive. An organisation that lab-tests mattresses gets through a few dozen a year and says so. A site with a confident verdict on every mattress in Britain has not tested them, and the sheer breadth of coverage is itself the evidence.

What happens when you check a single mattress

We took one popular UK mattress — a retailer’s own-brand model, well reviewed, around the £1,000 mark — and looked at everything that came up when searching for reviews of it, excluding the shop that sells it.

Four sites carried what looked like reviews. None of them showed any sign of having slept on it. One scored it 94 out of 100. Another gave it 4.2 out of 5 across ten measured categories with a named writer. A third stated in its own disclaimer that its metrics come from aggregated public reviews and that it takes commissions from the manufacturers it covers. The fourth was a shop.

Two of the four cited customer review counts within twenty of each other. Both figures matched the retailer’s own review count.

None of that means anyone was being dishonest. Spec sheets get copied, ranges get updated, configurations differ by size, prices move, and a page written six months ago may simply be out of date. But a reader trying to establish how many springs are in a mattress they are about to spend a thousand pounds on cannot get a straight answer from the pages that exist to review it — whatever the reason.

Where we stand, since you should ask

It would be comfortable to write all of the above and leave our own position out of it. So, plainly:

This site carries advertising. That is how it is funded. Advertisers have no say over what appears here and no advance sight of it, but money changes hands and you are entitled to know that before you weigh anything we say.

We have not slept on these mattresses either. We do not have a testing budget, and we are not going to imply that we do. What we publish is not a verdict on how a mattress feels, because we have not earned one.

What we do instead is narrower and, we think, more useful. We read the published terms — guarantees, trials, exclusions, standards, regulation — and report what they say, with the date we checked and a link to the source. Where a widely repeated figure cannot be traced to a first-party page, we say so and leave it out. There are several examples of exactly that across this site.

That is documentation rather than reviewing. It is a smaller claim, and it is one we can actually support.

The second problem: nothing lines up

Even a diligent reviewer runs into a structural obstacle in this market. Most mattresses are exclusive to one retailer.

The big chains sell substantial own-brand ranges under names that appear nowhere else. Those models are not stocked by competitors, do not appear on price comparison sites, and cannot be checked against an identical product elsewhere.

There are ordinary commercial reasons for this. Retailers negotiate specifications with manufacturers, and an exclusive range is a normal way to differentiate a shop. But whatever the intention, one consequence follows automatically: a price becomes difficult to check. If a mattress is sold in one shop under one name, there is no second price for it. A saving is presented against a reference price that someone outside the business has no straightforward way to verify. The ASA upheld a complaint against a mattress brand in August 2026 for precisely this — a savings claim benchmarked against competitors’ prices rather than the brand’s own usual selling price.³

The American trade has a name for the wider version of this — the “name game”, where near-identical mattresses carry different model names in different shops specifically so that price matching becomes impossible.⁷ We have not found UK evidence establishing that as deliberate practice here, so we are not claiming it is. The observable UK effect is simply that exclusive ranges exist and are extensive.

What that does to reviewing

Three things, all of them bad for the reader.

• Independent testers cannot cover the range. An organisation that buys and lab-tests mattresses can get through a limited number. Faced with retailers that each carry dozens of exclusive models, it will test some and not others. We looked for a lab test of the mattress above and found that the tester had covered a different model from the same range, but not that one. That is not a failure on the tester’s part; it is arithmetic.
• You cannot triangulate. With a phone or a car, half a dozen outlets review the same model and you can read across them. With a retailer-exclusive mattress there is only ever one source of primary information — the shop selling it.
• It creates the vacuum the affiliate content fills. There is real demand for a review and almost no genuine supply. Something grows into that gap, and what grows is pages assembled from spec sheets, because those are free.

And the brands are fewer than they look

The apparent variety thins out too, once you follow the ownership.

Two of the best-known bed-in-a-box names sold in Britain are the same company, trading under one UK entity, with warranty terms that are word for word identical — the same sag threshold, the same slat requirement, the same structure. A shopper comparing them believes they are weighing two options.

Elsewhere, a major high street retailer has been owned by one of the world’s largest mattress manufacturers since 2021. Another bought a well-known online brand out of administration in 2022 and now sells it alongside its own ranges.

None of this is hidden and none of it is wrongdoing. But a market that looks like a dozen competing choices is, in places, three or four companies wearing more hats than that.

What the rules actually require

Worth knowing, because the standards are higher than the sector’s general practice suggests.

• Affiliate content is advertising. The ASA treats affiliate links as advertising regardless of whether anything was gifted, and holds both the publisher and the brand responsible. The label must be up front and unambiguous — a disclaimer at the foot of the page is explicitly insufficient, because the reader meets the link first. The ASA has specifically rejected “we may earn a commission” as too vague, along with “affiliate” on its own and “Aff”.⁴
• Reviews must be real. Under the Digital Markets, Competition and Consumers Act 2024, content that “purports to be, but is not, based on a person’s genuine experience” is banned outright. Commission counts as an incentive that must be disclosed. The Competition and Markets Authority has been enforcing this directly, without needing to go to court, since July 2025.⁵
• Search engines expect testing too. Google added a rule in July 2026 prohibiting review markup for reviews not based on genuine experience of the product, and its guidance asks reviewers to document how many items were tested and how.¹⁶

The gap between those requirements and what is actually published is wide. It has not so far been vigorously policed — almost every mattress-sector ASA ruling we can find concerns pricing rather than review conduct — but the rules exist and the enforcement powers have got sharper, not softer.

So what can you actually compare?

Not the price, because exclusive models have no second price.

Not the customer reviews, because they are held by the retailer and recycled by everyone else.

Not the scores, because they are assembled from the two things above.

The written terms. Those are published, binding, and identical wherever you read them — which makes them the one genuinely comparable thing in this market.

None of those require anyone to have slept on the mattress. They require someone to read seven sets of terms and write down what they say. That is unglamorous work, and it attracts little commercial funding, because a page comparing guarantee exclusions persuades very few people to click Buy.

Reading a mattress review from here on

1. Scroll to the bottom first. Find the funding disclosure and read it before the verdict.
2. Look for a photograph of the actual mattress in a real room. Press images mean no contact with the product.
3. Check the customer review count against the retailer’s. If they match, you are reading the shop’s reviews.
4. Ask who slept on it, and for how long. If the page will not say, that is the answer.
5. Check a fact against the retailer’s own page. Spring counts and depths are easy to verify, and disagreements are common.
6. Treat the score as decoration. A number out of ten with no stated method is a design element, not a measurement.

And apply all six to this page as readily as to any other. We have told you how we are funded and what we have not done; the rest you should check.

The comparable facts — guarantee thresholds, trial terms, slat requirements and the rest — are collected on our UK bed facts page, and the guarantee small print is unpacked in what a mattress guarantee actually covers.