Capture the markup readers and crawlers can receive
Start with the real canonical URL and save its initial HTML. Extract the JSON-LD scripts from what you captured. If the app adds markup after JavaScript runs, inspect the rendered document separately and label each sample. An empty source extract tells you about that sample, not every rendering.
Record all blocks that describe the item. If a plugin and a custom template both publish the same article entity, compare their identifiers and facts. Multiple blocks can be intentional when they describe different entities or link related ones through a shared identifier.
Separate parsing errors from vocabulary and eligibility
First fix JSON syntax. A missing comma or unmatched quote prevents further inspection of that block. The browser validator provides syntax and selected structure checks. It does not fetch remote contexts or validate every Schema.org property. Compare type and property choices with the vocabulary and application requirements separately.
A valid Schema.org type can be useful data even when Google has no rich feature for it. Check Google's current requirements for the intended feature, then compare the markup with visible content. Passing a rich-result test does not establish that a result will appear.
Fix the template that produced the stale fact
In this fictional example, an article template now displays a corrected heading but a separate metadata field still emits the old heading. Correct the mapping at the source and re-extract another article to catch template-wide effects. The following minimal block illustrates the corrected fact; check feature-specific fields before calling a real article eligible.
{
"@context": "https://schema.org",
"@type": "Article",
"@id": "https://example.com/guides/room-layouts/#article",
"headline": "Plan a room layout for a small workshop"
}Re-extract after deploy and document what changed
Compare the deployed markup with the intended version, using both initial and rendered views if they differ. Check the visible facts and expected canonical URL before rerunning a consumer’s validation. Keep the before and after samples with the template change so the next content update can be checked quickly.