The question the listing never asked: what are you drilling into?

Every negative review said the same thing in different words: it pulled out, it spun in the hole, it did nothing. The quality team tested and retested the product and found nothing wrong with it, which was correct and beside the point. The anchors were being bought by people fixing a shelf to drywall, a rail to brick and a bracket to hollow block, and only one of those three had bought the right product.

The listing offered no help whatsoever with that choice. It described the load rating, the material and the pack count — all facts the shopper could not act on, because the variable that decides whether the anchor works is a property of their wall, not of the anchor. A 3.7 rating on a good product was the arithmetic result of letting two thirds of buyers guess.

We made the wall the first thing the page talks about.

• Substrate as frame two — a simple visual test for identifying drywall, masonry and hollow block, before any specification appears.
• One anchor type per listing, with the wall it is for stated in the title, ending the habit of selling three jobs from one page.
• A cross-reference block pointing a shopper with the wrong wall to the right product in the range, rather than letting them buy anyway.
• Keyword work by wall material, which turned out to carry far more demand than the generic anchor terms and almost none of the competition.

One-star reviews fell 71% and the hero rating climbed from 3.7 to 4.1 in six months, with no change to the product at all.

Revenue grew 80% on the back of the rating and a 39% conversion improvement: a shopper who is shown how to identify their own wall and then told plainly that this is the anchor for it needs no further persuading. Sending the wrong buyer to a different listing cost a few orders a week and bought back the reviews those orders were producing.