How we review

Comfort District sells completed research. We check the documents, mine the owner reports, read the warranty fine print, search the safety databases, and tie every published claim to a logged source. This page shows the whole method, because a research claim you cannot inspect is just marketing.

What kind of reviews these are

Every review on this site carries one of three badges, and the badge means exactly what it says. A research review synthesizes manufacturer documentation, regulatory records, warranty terms, and owner evidence gathered across public platforms; nobody sat in the chair. A spec audit is narrower: we verified the product's published specifications against primary sources on a stated date. The hands-on badge appears only when someone actually used that exact generation of the product, and until that happens for a given model, you will never see it there. We never imply testing that did not happen.

The source checklist

For each product, where applicable and available, we investigate: the current manufacturer page and spec documentation, the product manual, the actual warranty document with its exclusions, replacement-parts availability, manufacturer support channels, CPSC recall records, CPSC SaferProducts.gov incident reports, FDA regulatory status for red light devices, BBB complaint themes and company response behavior, archive.org records of prior-generation listings, major retailer listings and their Q&A, verified-purchase owner reviews, Reddit owner threads, specialist forums, YouTube owner material and comments, Trustpilot as a directional signal, and secondary editorial coverage. A required check means the category was investigated; when a search returns nothing, we say so and state the search scope rather than inventing findings.

How sources are ranked

Primary and regulatory sources win. Manufacturer documents, CPSC records, FDA databases, and other government data settle spec conflicts unless the evidence shows the primary source is stale or describes a different generation. Retailer and dealer data comes second. Owner evidence ranks third and works as patterns: a hundred owners repeating a myth never overrides a spec sheet, but recurring independent reports of the same failure are exactly how real problems surface. Editorial coverage from other sites ranks last, useful for leads and context, never as proof because it happens to rank well on Google.

How owner reports are counted

Two numbers appear on our pages and they measure different things. Sources checked counts distinct documents, domains, and datasets. Owner reports sampled counts the individual reviews, comments, and threads we actually read; a thousand comments on one platform still count as one source. Where a review displays platform counts, those are reports we located on that platform as of the stated date. Amazon's owner base gets described in words only, because Amazon's data may be used solely through its official API, which this site does not have access to yet. Every displayed number is stored in a per-product evidence ledger in our repository, and our publishing system refuses to ship a page whose numbers cannot be reconstructed from that ledger.

What Owners Report and Known Issues

Each researched review synthesizes owner experience in our own voice: fit, intensity, noise, comfort, setup, delivery, reliability, controls, warranty service, and long-term ownership. We report consensus praise and recurring complaints together, and we separate repeated themes from one-off anecdotes. Known issues carry one of three labels. An isolated report means one or a few credible accounts. A recurring owner complaint means independent repetition across owners or platforms, which is still not a confirmed defect. A documented issue means a manufacturer acknowledgment, service bulletin, recall, or equally strong corroboration, and we describe exactly what is documented and by whom.

Regulatory checks, worded precisely

When we search CPSC records we distinguish official recalls from consumer incident reports, and we never write "CPSC found" unless CPSC made the finding. For red light devices we record exact FDA status: cleared, approved, exempt, registered, listed, or no matching clearance located, and we never inflate any of those into "FDA approved." Finding no records is reported as no records found in our search as of the date, never as proof the records do not exist.

Warranty and parts research

Marketing pages advertise warranties; the legal document defines them. We read the actual terms and log the coverage period, parts and labor split, in-home service, shipping responsibility, exclusions, transferability, registration requirements, and the claim procedure, and where the marketing language and the legal language differ, the review says so. We also research whether replacement parts such as remotes, motors, pumps, and boards can actually be obtained, and we distinguish documented availability from assumption.

Product identity rules

Before research starts we pin down the brand, exact model, model number, generation, and release period, because comfort equipment brands reuse names across very different hardware. Owner reports, incidents, and specs from a different generation never get attached to the current model; prior-generation evidence is labeled as such wherever it appears. When generation identity is uncertain, the review discloses the uncertainty instead of papering over it.

Grades and the District Verdict

Reviews grade each dimension in words, never invented decimals: Excellent, Very good, Good, Fair, or Poor, and a dimension with no verified evidence stays ungraded with a note saying so. Every major review then ends in a District Verdict of BUY, CONSIDER, or SKIP. For search engines' structured data, that verdict maps to a fixed whole-number scale, BUY five, CONSIDER three, SKIP one, out of five; the number is our own editorial call restated, never an aggregate and never anyone else's rating.

Spec checks and public corrections

When current manufacturer data conflicts with older data, archived pages, or specs that other sites keep repeating, the review publishes a SPEC CHECK box: the disputed spec, the current primary value, the conflicting value, generation notes, sources, the verification date, and our conclusion. Before calling anything a correction we rule out generation differences, regional variants, SKU differences, listing errors, and unit conversions. Every meaningful correction is also listed on our public spec corrections index.

Manufacturer contact and corrections

Before publishing a safety concern, a recurring failure pattern, a warranty-denial pattern, or any other materially damaging allegation, we contact the company, log the date and the question, and allow three business days for a response where practicable. Their answer is represented accurately, or the review carries a documented no-response line; we never claim a company refused to comment unless it refused. If we get something wrong, the fix is made on the page, recorded in the review's visible update history, and reflected in the corrections index. Update histories record material changes only, never cosmetic freshness edits.

Affiliate independence

Comfort District earns commissions when readers buy through our links, mostly at Amazon, at no extra cost to the reader. Retailers and manufacturers never pay for placement, never see reviews before publication, and never influence a verdict; every comparison we publish names a loser, which is not how paid rankings behave. Buttons say what they do: the price lives at the retailer, and no purchase ever happens on this site.

Where to start

The method shows up best in the work. Walk the massage chair aisle, open any review, and check the sources box at the bottom against the claims above it.