Skip to content
Desk and Daylight

Our methodology

By Stephen V.Last reviewed

Units we have tested: 0.

We don’t have a lab. We haven’t sat in these chairs, stood at these desks, or switched on these lamps. Plenty of sites say they tested everything. We’d rather tell you exactly what we did instead, so you can check it.

Every number we publish comes from the manufacturer’s own spec sheet, manual, or a published standard. We cite each one. If a number isn’t published, we write “—” rather than guess. Our rankings come from arithmetic you can reproduce: adjustment ranges against published anthropometric data, weight ratings against real monitor weights, warranty terms read from the actual warranty document.

If we get something wrong, tell us and we’ll fix it and say we did.

Why we do it this way

Because the alternative would be a lie, and it would be an easy one to catch.

The publishers who rank above us compete on test labs and credentialled reviewers. We have neither. We could buy a stock photo of someone in a lab coat and write “we tested 40 chairs”, and for a while it would probably work. It is also fraud, it is trivially falsifiable by any competitor who cares to look, and it would destroy the only asset this site has.

So we compete on the one axis a new site can actually win: published, reproducible methodology. Our numbers come from documents anyone can open. A lab’s “we sat in it and liked it” cannot be checked by anyone.

Here is what that turned up in practice, and it is why we think the bet is right. When we went through the ergonomics category properly, we found that OSHA prescribes no desk height at all — despite the number of sites that say it does. We found that the desk-height formula the whole internet uses cites no source — not a study, not a standard, nothing. We found that ANSI/BIFMA X5.1’s own scope explicitly excludes “ergonomic considerations”, which means “BIFMA certified” on a chair tells you nothing about whether it is ergonomic. And we found that the CIE, the body that defined CRI, has published a statement saying the metric doesn’t work properly for LEDs and needs replacing — while every lamp on the market still leads with it.

None of that required a lab. It required reading the sources. That is the whole product.

What we measure, and why those numbers

Every product on this site carries a Spec Line: three numbers, and the third is always a longevity number. That is the site’s thesis in a component — slots one and two are “does this still feel right after eight hours”, slot three is “is it still fine in three years”.

  • Chairs— seat-height range, weight capacity, warranty years. The seat range decides whether a chair can fit your body at all, which is prior to whether it is comfortable.
  • Desks— height range, weight capacity, frame warranty. The floor height, not the ceiling: every listing advertises how high a desk goes, and the low end is what decides fit.
  • Monitor arms — weight range, VESA support, warranty. The minimum, not just the maximum: a counterbalanced arm below its minimum will not stay up.
  • Lighting— colour temperature, CRI, rated life.
  • Cable management— capacity, material, and whatever load or cycle rating exists.

Where the numbers come from

In this order of preference:

  1. The manufacturer’s spec sheet, manual or warranty document. Linked, and dated with the day we read it.
  2. Published standards — ANSI/HFES 100, ANSI/BIFMA X5.1, VESA FDMI, IES/CIE lighting metrics, ISO 9241.
  3. Government and peer-reviewed guidance — OSHA, CCOHS, Cochrane, published anthropometric datasets.
  4. Amazon’s Creators API — for price, availability and product images, exactly as returned.
  5. Aggregated owner reviews— for the subjective half we cannot assess, always labelled as owners’ reports and never as our finding.

Never a source:our memory, another affiliate site’s numbers, a manufacturer’s marketing adjective, or a language model’s recollection of a spec.

The dash rule

A spec we cannot verify renders “—”, never an estimate. An honest gap costs nothing. One invented number costs the entire methodology position, which is the only thing we have.

The dashes are not failures. They are the most informative thing on some pages. Three of the six chairs in our chair roundup have no published seat height — the single most important dimension on a task chair, and half the category treats it as optional. Two of the four bars in our light bar roundup have no CRI figure: one manufacturer’s own page claims Ra98 and Ra90 simultaneously, and another blocks access to its own specs while its two regional sites disagree.

You would not learn any of that from a site that filled the gaps in.

How ranking works

There is no score. We do not publish a 9.2/10, because we have not touched these products and a number like that would be an invented judgement dressed up as a measurement.

What we do instead, in order:

  1. Fit first. Can this product physically work for the person reading? A desk that will not descend to your elbow height is wrong for you at any price.
  2. Then longevity. What does the manufacturer commit to in writing, and how specific is that commitment?
  3. Then documentation.When two products are close, we rank the one that told you what you were buying above the one that didn’t. That is a real quality difference, and it is the one we are actually equipped to judge.

You are welcome to disagree with that weighting. It is written down so that you can.

What we can’t tell you

This matters as much as the rest, so it gets its own section rather than a footnote.

We cannot tell you whether a chair’s mesh digs in at hour four, whether a desk wobbles at full extension, how loud a motor really is at 7am, whether a lamp flickers, or how any of it holds up after a year of real use. Those need hands and time, and we have neither.

For that half, owner reviews will serve you better than we will. We are not going to pretend otherwise to keep you on the page. And a generous return window is often worth more than a warranty: it answers the comfort question with your own body, in a fortnight, which is something no amount of spec-reading can do.

Prices

Prices come from Amazon’s API and nowhere else. They are stamped with the date they were fetched. If our data goes stale — older than 48 hours — the number disappearsand the button falls back to “Check price on Amazon”. It never quietly shows you an old number.

Right now: 33 products carry a live price, last verified Jul 17, 2026. We never write a price into an article by hand, and we never publish a rating or a review count — if the API doesn’t supply it, the page doesn’t claim it.

How we make money, and what it doesn’t buy

We earn a commission when you buy through our links, at no extra cost to you. That is the whole business model and it is on a page of its own.

No brand pays to appear here. No brand has been given a preview, a veto, or a phone call. Commission rates play no part in ranking — and the easiest way to check that is to notice that we rank products we cannot earn much on, tell you not to buy things regularly, and left the biggest name in the standing-desk category off our roundup entirely because we could not read its specs.

Corrections

If a number here is wrong, we want to know, and the form is here. When we correct something we log it visibly on the page rather than quietly editing it — a correction you can see is worth more than a page that has never admitted one.

Every page carries a “last reviewed” date. That is the date a human actually re-checked it, not the date the site was rebuilt. An automated date bump with no review behind it is a lie, and a cheap one to get caught at.