How we score products

Every product page leads with one of four answers. They are a closed set, deliberately — a fifth option meaning “it depends” is where every undecided verdict would quietly end up, and a recommendation that can say anything says nothing.

• Buy now — worth its current price today. Not “the best in the world”; worth what it costs, now.
• Wait for a sale — the right product at the wrong price. The product is sound, the number on it is not, and history says the number moves.
• Skip — not worth it at any price we expect it to reach. A recommendation nobody can fail is not a recommendation.
• Consider an alternative — nothing wrong with it, but something else does the job you are buying it for noticeably better. The alternatives block at the bottom of the page names which, and why.

The verdict is a judgement, not a threshold. A 7.1 can be a Buy now if it is cheap for what it does, and an 8.6 can be a Wait for a sale if it has just launched at a launch price.

The PickD Score runs from 0 to 10 with one decimal place. It is our evaluation of a product against the standard for its category — not against the other products that happen to be in the same round-up, and not against everything ever made.

Roughly what the bands mean:

• 9.0 and above — exceptional. Rare, and getting rarer as categories mature.
• 8.0–8.9 — very good. Few real compromises at its price.
• 7.0–7.9 — good, with trade-offs worth knowing about before you buy.
• 6.0–6.9 — fair. Competent, but beaten in its price band.
• Below 6.0 — weak. There is a better use of the money, and the page will say what it is.

A score carries the date it was set. Prices move, firmware changes things, and successors arrive — a score from eighteen months ago is a historical fact, not a current recommendation, and the date is there so you can weigh it yourself.

Before any product in a category is scored, the criteria for that category are fixed. Headphones are judged on noise cancellation, call quality, comfort and battery. A mouse is judged on sensor accuracy, ergonomics, buttons and build. A single five-star scale stretched across unrelated things measures nothing at all.

Every criterion is scored on its own, 0 to 10, against the category standard. The overall score is a weighted combination of those criteria — the weights are part of the category’s configuration, not something an editor picks per product to reach a number they had in mind.

The breakdown is published on the product page. If the overall looks generous, you can see exactly which criterion carried it.

Two numbers appear on a product page and they measure different things. The PickD Score is ours, on a 0–10 scale, and it is an evaluation. The community rating is yours, on a 1–5 scale, and it is a report of experience.

They are never averaged into one figure. Blending an editorial judgement with an owner poll produces a number that means neither, and hides the far more useful signal — the cases where they disagree. A product we score well that owners rate badly is telling you something specific, and merging the two would erase exactly that.

No brand, advertiser, or affiliate partner has any input into a score or a verdict. There is no rate card, because there is nothing on sale.

The order is fixed and not negotiable: the research finishes, the verdict is written, and only then is a retailer link attached. Whether a product has an affiliate link, and which retailer pays the better commission, are facts nobody scoring it is working from.

Where a product page earns us a commission, the link itself says so — not just the footer. See how we make money and what our editors may and may not do.

Most verdicts on this site are research verdicts. That means published specifications, independent measurements, long-form reviews, teardowns, and the pattern in what owners report once a product has been out long enough to have one — assessed against the category standard and against the price.

Where somebody here has physically used a product, the “How we reviewed this” box on that page says so. Where it does not say so, nobody has. We would rather under-claim on every page than let one page imply a test that never happened.

No verdict on this site is generated automatically. Software helps gather specifications and track prices; a person decides what a product is worth, writes the line that says so, and is accountable for it.