W.W. Grainger

The four pillars

Scored against the published framework — fourteen signals, no model judgement anywhere in the scoring.

What the site probe found

The pages we sampled

Every one, openable. Each was taken from the middle of a different category listing — never a featured or promoted placement — so the sample reflects the catalog rather than its best shelf.

Page score is out of 69 — the portion of the framework scored per page. The remaining 31 points are site-level: consistency, gating, and the four agent-readiness signals.

What would move this number

Derived from the score itself, largest gap first — Buyer Answerability lost the most points.

1. 1A GTIN or UPC appeared on 0% of the sampled pages. Identifiers are what let a marketplace, a search engine, or a procurement system recognise your listing as the same item everyone else is selling.
2. 2Bot protection returned no readable page to a standard browser client, so structured data could not be observed at all. Whatever markup exists behind that wall, a crawler or shopping agent hits the same barrier — which is worth weighing against what the protection is buying.
3. 3robots.txt explicitly blocks named AI crawlers. That is a legitimate and deliberate choice, not an error — but it is worth revisiting deliberately rather than by inheritance, because it decides whether your catalog can be cited when a buyer asks an assistant what to buy.
4. 4The richest and thinnest pages sampled differed by 14 attributes. A spread that wide usually points at categories onboarded under different rules rather than at any one bad page.