Watsco

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 — Machine & Agent Readiness lost the most points.

1. 1No schema.org Product markup was found in the server-rendered HTML. This is usually the cheapest fix on the list and it is the single clearest signal to a crawler or an AI agent that the page describes a product.
2. 2robots.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.
3. 3The richest and thinnest pages sampled differed by 29 attributes. A spread that wide usually points at categories onboarded under different rules rather than at any one bad page.
4. 4A price was published on 20% of sampled pages. Withholding price is a legitimate commercial strategy, and it is also a hard stop for any automated buyer comparing options.