What gets printed, and what gets dropped

Most pages with this heading describe a testing bench. This one cannot, because there is no bench: the catalogue is built from what the retailer publishes, and the honest version of “how we choose” is a description of how that data was handled. It went like this, in order.

1. Read the whole storefront, not a selection

   Every listing published under the brand was pulled from the retailer’s own product feed in one pass — thirty-seven of them. Nothing was excluded at this stage, because a selection made before the data is understood is a selection made by taste.

2. Group the listings by their rating pool

   This is where the shape of the range appeared. Listings that share one pool of ratings are one design in the retailer’s own bookkeeping, and grouping on that basis turned thirty-seven listings into seven groups without a single judgement call from us.

3. Demote colour from a product to an attribute

   Each group became one page, and the shades inside it became swatches on that page. Thirty-seven colourways survive as swatches, two of them marked as out of stock at the time of reading.

4. Take the rating from the group, not from a listing

   A score copied off one colour would be the same pooled number wearing a narrower label. The figure on each page is the pool for the whole design, and the count is printed beside it, because a 4.7 average over 106,115 ratings and the same average over a dozen are not the same claim.

5. Read one price per product and stamp it with the month

   The lowest live price across the group, read in July 2026. Colour variants can differ by a dollar or two; where they did, the cheaper reading is the one printed, and the page says which colourway it came from.

6. Write down the contradictions instead of settling them

   Where two fields of one listing disagree, both figures go on the page with the field each came from named. There are two such cases in this range, and every one of them is printed rather than resolved by us.

7. Drop what cannot be published honestly

   Foreign-language bullets, and a category rank that belongs to a family of listings rather than to any single product. Both are described below, because a catalogue that quietly drops things is indistinguishable from one that never had them.

What “choose” does not mean here

It does not mean curation. Nothing was left out for being poor, and the lowest-rated product in the range — door mats at 4.3 — has the same kind of page as the highest. A catalogue of one brand that quietly omitted the brand’s weaker products would be an advertisement with a reading room attached.

It also does not mean recommendation. The front page picks four products by four printed rules — largest rating pool, highest score, lowest price, widest colour choice — and each card says which rule put it there. Those are positions in a list, not opinions, and you can check every one of them on the shelf pages, where the same figures stand in the same columns.

A rank that belongs to a family, not a product

The category position on the storefront appears on four listings as the identical place in the identical category. A number that is the same for several listings is a number about all of them, so it is printed only where it belongs to a single product.

What we genuinely cannot tell you

Whether the silicone smells when new. Whether a seam gives out in the second year. Whether the cuff clears your forearm, which depends on your forearm. Whether the board warps in your dishwasher, on your cycle. None of that is in a product feed, and no amount of careful reading turns a feed into experience.

What the data does support is narrower and still worth having: how long each cuff is, how many pieces are in which box, what the care field says as opposed to what a marketing bullet says, how many people have rated a design and how strongly, and where the seller’s own paperwork disagrees with itself. Those are the questions the seven pages here try to answer well.

Last reviewed July 2026.