How we choose and score picks

Two kinds of pick, labelled honestly, and one published rubric.

Tier 1: Hands-on

A Hands-on pick is a product we own and use at our own desk. It gets our own photographs, measurements where they matter (actual height range, actual weight, actual noise), and notes from weeks of use rather than an afternoon. Hands-on picks are marked with a badge on the comparison table and beside their score, with a note on what our own use confirmed. Owning something can only move two of the four product marks - Versatility and Build, the two things use can actually prove - and the note says what moved and why. It never changes Capability, which is what the specification says, or Value, which is about the job the pick is doing on that page, and it never adds a bonus on its own, so a product we own can still lose to one we do not.

Tier 2: Spec-checked

Most of the market is out of reach of one desk, so most picks are spec-checked: built from the manufacturer's published specification, the standards it supports, warranty terms, known platform constraints and brand track record. We read the spec sheet, not the marketing page, and we verify the retailer link lands on the exact product we describe. A spec-checked pick is never presented as something we have handled.

Link verification applies to both tiers, and it is the check we take most seriously: before a guide goes live, every recommendation is checked against the live retailer page so the name matches the exact product we describe, and we re-check every link on the site on a rolling basis. When a listing dies or is replaced, we replace the pick and rewrite the guide around a verified alternative rather than leave a dead recommendation up.

How we score (the HSP Score)

Every pick carries an HSP Score out of 10. This is our own editorial score, worked out from a fixed rubric applied to the product's specifications and intended use. It is not a retailer star rating and it is not taken from or matched to customer reviews.

The rubric scores the product on four things, each from 1 to 10, and a fifth mark rates its fit for this guide. Every guide shows the scorecard under each pick so you can see the working. The four product marks belong to the product and are the same in every guide you meet it in; Fit is set per guide. Together the product marks carry 70% of the total and Fit the other 30%; the share of the total each mark carries is in brackets:

Because every product here has already cleared the checks above, the scores sit in the upper half of the scale: the total and the scorecard are there to separate strong options from each other, not to rate them against the whole market. Four marks rate the product; the fifth rates its fit for this guide. The total is exactly the weighted mean of the five printed marks at the weights above, so you can check the arithmetic yourself. No two picks on a page ever share a score, so the order is always clear - and the size of the gap is worth reading. A pick at 9.5 or above carries a short What keeps it off a 10 line, because at that level you deserve to know exactly what is missing; below that, the honest lows live in the scorecard and the cons. Nothing here scores below a 7: a product we would rate lower than that would not be on the page at all. We revise scores when we update a guide.

The badge leads; the score supports it. Our Best Overall pick is the one we would buy for most people at its price and size, and it carries the top score in its guide because fit is part of the score. A premium pick is often the more capable product, and the four product marks show exactly where it is stronger; its Fit mark says how much of that the guide's reader actually needs. A Best Budget pick can top its page when it genuinely fits the brief better than the overall pick, and the page says so.

How the guides are written

Every product fact on a guide is traced to a published source: the maker's own specification page or the live retailer listing. The prose around those facts is drafted with AI assistance, then checked before it is published, by automated checks (port and cable counts against their sources, other figures flagged for review, each link against the live listing, banned claims such as prices in the prose or star ratings) and by a separate AI review against the same sources. AI tools help find candidate products, but a product reaches a page only after it clears the quality bar on the live listing, and a model never supplies a fact: a figure we cannot trace to a source is cut or sent back for review.

What we will not say

Ergonomic gear attracts medical-sounding marketing. We describe what a product is designed to do (an adjustable lumbar support, a sit-stand range that suits a given height) and, for a Hands-on pick, what it was like to use. We do not claim that any desk, chair or accessory treats, fixes or prevents back pain, posture problems or any other condition, and we do not repeat such claims from manufacturers. If you have a medical question about your setup, ask a professional.

What we deliberately leave out

We do not quote prices in our write-ups. Where a price appears in a comparison table it comes from the retailer's own product feed and is dated; prices change, so always check the retailer page before buying. We do not display star ratings, review counts or stock claims either: ratings belong to the retailer, not to us, and a stale availability line is worse than none. Check the retailer page via any of our links for the current price, availability and reviews. Our job is the part that stays put: what the product actually is, and whether it is the right pick for you.

Links: About • Disclosure • Privacy