How the Rich Results Test works
You give it either a live URL or a block of pasted code. It fetches the page as Googlebot, runs the rendering step so that markup injected by JavaScript is included, and compares what it finds against Google’s published requirements for each search feature.
Results are grouped by feature type rather than by tag. For each type it lists the items it detected and flags what is missing. Errors are required properties that are absent, and they disqualify the item. Warnings are recommended properties that are absent — the item still qualifies, but with less available to display. You can switch between the smartphone and desktop crawler, and open the rendered HTML to see exactly what was parsed.
Why the Rich Results Test matters
Structured data is invisible. Nothing on the page looks different when it breaks, so without a validator you discover the problem weeks later, when a rich result quietly disappears from search. This test shortens that feedback loop to a few seconds.
It also settles a recurring argument about JavaScript. A plugin or a tag manager that writes structured data after the page has loaded may or may not survive Google’s rendering step. Testing the live URL rather than reading the page source is the only dependable way to find out.
Common mistakes with the Rich Results Test
The biggest is reading “eligible” as “will appear”. The test judges the markup, not the page’s chance of being chosen. Google decides feature by feature and query by query, and it has retired whole feature types before. Promising a client a rich result on the strength of a green tick is a promise you cannot keep.
Two smaller mistakes follow. Testing pasted code instead of the live URL hides everything that caching, rendering or the live template could break at publication. And driving every warning to zero is rarely worth the effort — required properties are not negotiable, optional ones are a judgement about how much detail you want available.
How to act on it
Test a template once rather than every page built from it, then spot-check a few live URLs after each release. Most structured data faults are template faults repeated at scale, so fixing the template fixes the site. Clear all errors first, then decide which warnings earn the extra fields.
After that, hand continuous monitoring to Search Console. Its enhancement reports watch the whole site over time, while this test examines one page on demand, and the two answer different questions. Keep the markup and the visible content in step: if a price, a stock status or an opening time changes on the page, the schema markup has to change with it, or you are describing a page that no longer exists.
