What I cannot do, said plainly

The best air purifier sites buy the units and measure the output. HouseFresh runs particle counters. AirPurifierFirst names its instrument, a Temtop LKC-1000S, on every review it publishes.

I do none of that. There is no lab here, no test chamber, no shelf of samples. When a page says a purifier covers 200 square feet, that figure comes from the manufacturer, and I have not stood in a 200 square foot room measuring whether it does.

That is a real limitation and it decides what this site is useful for. It is good at questions with published answers: what size fits, what the clearance has to be, which part number belongs to which model, whether a listing is worth trusting. It is not good at how quietly a fan runs, and any page claiming otherwise would be making it up.

Where the measurements come from

The dimensions and clearances on this site come from published trade and design references rather than from other blogs. The ones that come up most often are the NKBA kitchen planning guidelines, Dimensions.com, Fine Homebuilding, and manufacturer documentation for specific models.

Two rules go with that. Every figure links to where it came from, and every page carries a sources block at the foot listing them. Where I could not find a source for a number that seemed worth including, the page says so in the text instead of quietly presenting it as fact.

Around twenty figures were removed or relabelled on that basis during a review in August 2026. Some were reworded so the number was not needed. Some were traced to a source and kept. A few are still on the site marked as unsourced, because they are useful and honest labelling beats deletion.

What a product has to clear

The bar is deliberately boring, and it rejects more than it accepts.

That last row is the one that does the most work. Every product on this site has a stated flaw. If a listing has genuinely nothing wrong with it, I have not read it carefully enough, and a recommendation with no downside attached reads as advertising because it usually is.

Consumables are the exception, and I will say so on the page when it applies. Filters, mats and cartridges attract very few reviews, because they are bought by people who already own the unit. Several of the genuine AeraMax listings have no one or two star reviews at all, so there is nothing to quote. On those pages the drawback comes from what I can check instead: how much stock is left, who is actually selling it, what the manufacturer documentation says, and whether the genuine part can be bought in the US at all.

Commercial and contract-grade products

The bar above assumes a product people buy one at a time and then review. Office furniture sold through dealers and specifiers does not work that way. A monitor arm that ships by the pallet to a facilities manager will never collect fifty consumer reviews no matter how good it is, so applying that threshold to it does not filter quality, it just filters out an entire category.

So contract-grade products are judged on published evidence instead, and land in one of three places.

The checks behind those three:

What BIFMA compliance actually is, and what it is not

I had this wrong for the first months of writing this site, so it is worth stating carefully.

BIFMA does not certify furniture. Its own licence agreement says so in as many words: “BIFMA does not test or certify products and BIFMA is not responsible for the safety and performance of products listed in this Compliant Program.” The programme is a way for a manufacturer to self-declare that a product meets a standard, publish that declaration on a public register, and be held to it.

That is not nothing. To list, a company signs a licence, pays, and has the product tested in a lab accredited to ISO 17025. BIFMA audits a fifth of a new entrant’s listings and a slice of everyone’s annually, and anyone can file a challenge the manufacturer then has to answer with test data. But the audit reads the paperwork rather than retesting the chair. A listing is a claim with a receipt attached. Not a seal.

Which means “BIFMA certified” is a phrase with nothing behind it, and I have stopped using it.

The register is at compliant.bifma.org and I check it directly. In August 2026 it held about 3,390 products from 86 brands. Herman Miller, Steelcase, Haworth and Knoll are in it, as you would expect. So are IKEA, Staples, Office Depot, Crate and Barrel and Sauder, which matters, because it kills the excuse that only contract brands bother.

So absence from the register gets read three ways here:

That third row cost me several verdicts. Ergotron has no products in the register, so failing its converters on a standards check was measuring whether Ergotron had joined a voluntary scheme, not whether its hardware holds a monitor. Eurotech is not in it either, including the one chair this site recommends, and that page says so.

Where a safety listing actually lives

This one I got wrong twice, on the two pages where being wrong mattered most, so it gets its own section.

For a powered desk or a lamp, the listing usually does not belong to the brand on the box. It belongs to whoever made the control box or the power supply, and it often appears nowhere on the seller’s own website. A brand can ship a properly listed product and never print a mark on any page you will read.

So the question is not whether the brand publishes a mark. It is whether a listing exists anywhere in the chain, and the places worth looking are the component maker’s manual, the dealer channel’s spec data, and the certification registers themselves.

Those registers are the honest weak point in this method. UL Product iQ, Intertek’s ETL directory and CSA’s product listing are all gated or blocked to me. When I cannot query them, I say the mark is absent from the documents I can reach, and I never say it does not exist. An absence I could not test is not a finding.

A certification can be superseded, and the manual will not tell you

Registers get reissued. AHAM’s directory of certified air cleaners comes out in editions, and a product can be retested between them, so the number a manufacturer prints and the number currently on file are not always the same figure.

The Fellowes AeraMax range shows how far that can drift. The manuals were printed in 2021 and quote the April 2020 certification. AHAM published a new edition in October 2023 and the numbers had moved. On the AeraMax 90 the certified smoke rating went from 68 to 71 and the pollen rating from 75 to 87. On the 290 smoke went from 191 to 199.

Check the current edition before accusing anyone of inflating a number. Held against the 2020 file, Fellowes appears to overstate its pollen performance by sixteen percent. Held against the 2023 file, which is the one in force, its published figures match its certification exactly. The manufacturer was right and the paperwork in the box was out of date, which is a different story with a different villain.

Which standard belongs in a nook

BIFMA publishes different standards for different duty. The ones people quote are X5.1 for office chairs and X5.5 for desks and tables, both written for commercial use. X5.1 targets “an estimated product life of ten years based on single-shift usage”, roughly twenty thousand hours in a seat. A chair that passes it will not be troubled by a spare room. The bar sits well above what a home office asks of it.

But BIFMA also publishes a tier for the way most people reading this actually work. ANSI/BIFMA X6.5 covers home office and occasional-use desks, tables and storage, and ANSI/SOHO S6.5 covers furniture whose “styling, marketing and chain of distribution are intended to address usage in a residential, home office or small office environment”. There is an occasional-use seating standard as well, X6.4.

X6.5 is the honest yardstick for a corner of a spare room, and it is the one I cite first. Where a product clears the commercial tier instead, I say so. That is the higher bar, not a substitute for this one.

Two things worth separating out. X7.1 is an emissions standard, covering formaldehyde and volatile compounds, and it turns up in marketing next to the word BIFMA while saying nothing about whether a desk holds weight. And BIFMA publishes no standard for task lighting at all, so judging a lamp against BIFMA is judging it against a rule that does not exist.

A product recommended this way is labelled on its own page as a pick, with the standard it is listed against named. It is a different test from the review-count bar, not a softer one. Consumer reviews tell you what happens after six months in somebody’s house, and no amount of compliance paperwork replaces that.

Where a product misses this bar, it stays on the site with the verdict it earned, and the page says which check it missed and what you would have to accept to buy it anyway.

What gets rejected

A few patterns come up constantly and all of them get passed on.

Perfect ratings on tiny samples. Twelve reviews at five stars tells you nothing. It could be a good product or a new one with helpful friends, and there is no way to tell them apart from the outside.

Famous names with thin listings. A well-known brand carrying 46 reviews and a seven-hundred-dollar price through a third-party seller does not clear the bar just because the name is familiar.

Whole brands sold through one reseller. When every listing for a manufacturer runs through the same marketplace storefront with a handful of reviews each, that is a signal about the channel rather than the product, and the honest answer is to send you somewhere else.

Categories where nothing is good. Task lighting is the clearest example. Hundreds of near-identical listings, a few dozen reviews each, nothing with real evidence behind it. Where that happens the page says the category is weak instead of picking a winner at random.

How this content is made

Google asks publishers whether the use of automation, including AI generation, is self-evident to readers. Here it is.

These pages are researched and drafted with AI assistance, then checked, corrected and edited by me before anything publishes. The research is where it does most of the work: pulling specs, comparing listings, reading through review patterns at a volume I could not do by hand.

What it does not do is decide what gets recommended, and it does not go up unchecked. A review of this site in August 2026 caught four factual errors that had made it into drafts, including a footprint calculation that dropped a clearance and was wrong by more than two feet. Those were found and fixed before publication, which is exactly the argument for a human reading every page rather than trusting the draft.

If you find a number here that looks wrong, it may well be. Tell me and I will check it and either fix it or show you the source.

How this site makes money

Some links here are affiliate links, which means a purchase made through one pays a small commission at no extra cost to you. Pages carrying paid links say so at the top, above the content, rather than burying it in the footer.

Two things that follow from that, and they matter more than the disclosure itself.

Recommendations are not ordered by what pays best. Several pages on this site point you at a retailer that pays less, or at no retailer at all, because that was the right answer. One tells you the machine has a washable filter and therefore no repeat purchases, which is the opposite of what a site funded by repeat purchases wants you to know.

And prices are shown as bands rather than figures. Affiliate programs generally prohibit stating a price you are not reading live, because a stale number is a misleading one. Where you see a range, that is why.

Questions

Have you used the products you recommend?
Mostly no, and the pages say so. I ran a contract furniture business until 2020, so the office seating and contract fittings are things I sold and handled. Everything else is assessed from published specs, seller history and review patterns.

Why are there no prices?
Because a price printed on a page is out of date within days, and affiliate terms generally forbid stating one you are not reading live. Bands are honest. Exact figures would not be.

Why does a page sometimes tell me not to buy the thing it is about?
Because that was the answer. It happens most often when a good product only sells through a bad listing.

How current is any of this?
Every page carries a last-updated date at the top and stock was checked on that date. Listings change faster than pages do, so treat anything about availability as of that date rather than as of today.

I ran Corner Nook as a contract-grade furniture store from 2020. It now covers furniture for small spaces, corners, nooks, and rooms that have to do two jobs. More about who runs it, or go back to corner nook ideas and how to measure a corner.