Built around the height a desk drops to, not the one it rises to

A desk works for a person when it can reach two heights: their elbows at roughly a right angle sitting down, and the same thing standing up. Those land at about 0.4 and 0.62 of a person’s stature, which are the constants behind every desk-height calculator ever written. Run them backwards and a desk’s own two heights tell you the span of people it serves — its minimum divided by 0.4 is the shortest person who can sit at it, its maximum divided by 0.62 the tallest who can stand.

That span is drawn on every card and every desk page as a lit segment on a 4′8″–7′0″ rule, so two desks can be compared by eye. It is an estimate and is labelled one everywhere it appears: torso-to-leg proportion varies between people of the same height, a keyboard tray lowers the working surface and a thick top raises both ends. The arithmetic and its limits are set out in full. Desktop converters get a one-sided band, because a converter sits on a desk you already own and cannot change how low that desk sits — 3 of the 53 entries here are converters and none of them claims a shortest user.

7 construction types, from three-stage electric columns down to counterbalance arms, chosen because they predict behaviour: stages decide how far a leg collapses, leg count decides how much the surface sways at full height, and a converter is a different object from a desk. Converters are held out of every height statistic on the site rather than dragging the catalog minimum down to a floor they do not own.

The 14 feature chips on the listing are predicates over the structured fields, never string matches against a maker’s own wording. One brand writes “memory presets”, another “4 programmable positions”, and a filter that reads prose quietly finds half of what is there. Brands with fewer than 2 models get no brand page — 16 of 23 clear that — because a hub listing one desk is a page pretending to be a category.

Manufacturers’ own product pages and documentation, then published spec sheets and named third-party measurements where a maker leaves a gap. Height ranges get their own field recording the basis each figure is quoted on, because some makers publish the bare frame and some the frame with their standard top fitted, and the gap between the two conventions is over an inch. Comparing one against the other is the easiest mistake to make when cross-shopping two desks, so every desk page states which it is using.

The catalog spans a 22″ floor to a 52.6″ ceiling, and 49 of 53 desks reach high enough for someone 6′4″ to stand at. How we compare sets out the sourcing rule for each field and how the shortlists are sorted.

Travel speed is unpublished for 20 of 53 desks and noise for 20. Those fields read “not published” rather than carrying a plausible-looking guess. A missing figure is missing data, not a low score, and a desk is never ranked down for a number its maker declined to print.

No desk here shows a star rating. A rating appears only behind a named, non-retailer source, and 0 of 53 currently have one — so the stars stay hidden and there is no aggregate review markup in the structured data anywhere on the site. Product photography is at 41 of 53; a desk without one shows a plain panel carrying its brand and model.

Public owner discussion and long-form write-ups by people who bought the desk and lived with it, read rather than summarised second-hand, and labelled as owner accounts wherever they appear. 24 of 53 desks have enough genuine discussion behind them to carry an owner section.

The pros and cons on a desk page carry the same distinction: 10 desks draw on owner accounts, 43 are written from the specification and the design alone, and each page says which. Nothing here says “owners report” about a desk whose owners have not been read. It is why the wobble complaint recurs on the two-leg frames and is absent from the four-leg ones — that is what people actually wrote, not a house theory applied evenly.

Two people, at desks they paid for themselves. No office, no test lab, nobody on staff. No desk here has been assembled, loaded, pushed on or measured by us, and the site never claims otherwise — where a judgement rests on hands-on experience it is attributed to the owners who had it.

Nothing is paid placement. There are no affiliate links, no buy buttons and no retailer links of any kind on this site; no manufacturer has paid to be listed, ranked higher or described favourably, and none has any say in what appears. There is no advertising network and nothing to sign up for. What the site does collect — server logs, analytics and a filter preference kept in your own browser — is set out on the privacy page.

The catalog carries one date and everything on the site reads it: 2026-10-07, the last full pass over the 53 desks. It moves when the data moves, not on a schedule, and no page bumps its own date to look fresh. Prices are hand-recorded snapshots from that pass, shown with a tilde, and will differ from what a seller is charging today; several are the entry configuration of a range that costs considerably more once specified, which each desk page says outright.

The 12 shortlists, 67 head-to-head pages and 8 explainers are all generated from that same file, so none of them can quote a desk or a number the catalog does not hold. Last updated October 2026.

A wrong height range is the correction that matters most here, since those are the figures makers document least consistently and the ones the whole site is built on. Send it to the contact page, and it gets checked against the manufacturer’s own documentation before anything changes.