How We Research
Quick answer
Every Honest Review Hub article starts from real reader questions, checks published trials and clinical guidance, and cites its sources. If no good source exists for a question, we say so openly. Marketing copy is never a source, and no number ships without a citation you can check.
Applies to: anyone who wants to verify the method behind any review on this site.
Most pillow "research" you've read was written by a marketing department. Ours starts in medical journals.
The pipeline, in six steps
-
We start from real questions
We collect the actual questions people ask search engines about pillows and sleep — hundreds per topic — and deduplicate them so each article answers a distinct question. No article exists just to capture a keyword.
-
We read the primary sources
Published trials first, then systematic reviews, then clinical guidance (NHS, Mayo Clinic, FDA), then expert consensus. Marketing copy and other review sites are never sources.
-
Every number traces to a citation
If a page states a figure, the source is named on the page. Our internal evidence ledger holds the exact claim and URL behind every fact we publish — 269 sourced rows at last count.
-
If no good source exists, we say so
Some questions — for example, how often to replace a pillow — have no clinical studies behind them. We label those answers as convention or consensus and say plainly what the evidence doesn't support. We never dress up a guess as a finding.
-
We check the complaint record
For every product we review, we check independent complaint sources — Trustpilot, consumer organizations, and what Google's own AI answers report — and we publish what we find, including refund problems and dark-pattern tactics.
-
We update
Pages carry publish and review dates. Claims are rechecked on a cycle and whenever major new evidence appears. Outdated findings get corrected with a dated note, not silently edited away.
What counts as evidence
Not all sources are equal. Here's the hierarchy we follow — from the top:
Manufacturer claims sit at the base: treated as claims to check, never as evidence. Text equivalent: randomized/blinded trials → systematic reviews → clinical guidelines → large observational studies → expert consensus → manufacturer claims.
| Evidence tier | How we treat it |
|---|---|
| Randomized / blinded trials | Strongest. Example: Gordon 2010, the only blinded 5-material pillow trial — latex ranked best, feather worst. |
| Systematic reviews | Weight of the evidence across trials; preferred over any single study. |
| Clinical guidelines | NHS, Mayo Clinic, and similar guidance for safety and care-seeking lines. |
| Large observational studies | Used with caveats — correlation, not proof. |
| Expert consensus | Labeled as consensus, never as proven fact. |
| Manufacturer claims | Treated as claims, never as evidence. Checked, then labeled. |
What we don't do
- No paid placements. Scores and rankings cannot be bought.
- No brand bidding. We run organic content and never bid on brand keywords; our product names stay in body copy only.
- No invented statistics, experts, or testimonials. Every number is cited; every byline is real.
- No cherry-picking. We never pit one favorable study against the weight of evidence to flatter a product.
- No health-outcome promises. We describe what research found; we never promise a pillow will treat pain, snoring, or any condition.
- No fake lab testing. We do not run a sleep lab, and we never imply we do. Our "testing" is reading the evidence and checking the product record.
Corrections policy
Errors are corrected visibly, with a dated correction note on the page where the error appeared. If you spot a mistake — a wrong number, a broken citation, an outdated claim — write to [email protected] and we will check it and publish the fix.
Who enforces all this?
One person is accountable for every page on this site — and the standard he enforces on himself.
Last reviewed: 2026-11-01 by Honest Review Hub Editorial Team
How this page was made
This page documents the site's own research pipeline from the published methodology records — it describes how articles are built, not a health finding. It cites no external studies because it makes no health claims; its job is to let any reader audit the method.
The Gordon 2010 trial is referenced as an example of our top evidence tier, sourced from our internal evidence ledger.