Structured data is how you tell search engines exactly what a page is: a product, an article, a recipe. Done right it earns rich results. Done wrong it is silently ignored. Here is what the errors and warnings mean, and how to fix them.
Structured data is a machine-readable description of a page, added as JSON-LD in the page head. It does not change what a visitor sees. It gives search engines an unambiguous label for the content, which is what makes rich results possible.
Google reads structured data to decide whether a page is eligible for an enhanced listing: star ratings under a product, a recipe's cook time, an FAQ expander, a breadcrumb trail. These enhancements take up more space in the results and draw more clicks, which is the entire reason to add schema. But eligibility is conditional: the markup has to be valid, and it has to contain the fields Google requires for that result type.
That is where the two audit findings come from. A validation error means the markup is malformed or breaks schema.org's rules, so Google discards it entirely. A missing-field warning means the markup is valid but lacks a field Google recommends, so the page works but is not eligible for the richer version of the result. The example below is a minimal, valid Article block:
The three ways structured data shows up in an audit.
The markup breaks schema.org rules: a wrong data type, a malformed date, a required property missing entirely, or invalid JSON. Google discards the whole block, so you get none of the benefit. Fix by correcting the flagged property and re-testing until it validates clean.
The markup is valid but lacks a field Google recommends for that result type. The page still works; it just is not eligible for the enhanced listing. Fix by adding the recommended property, using real data. Never invent a value such as a rating to satisfy the warning: fabricated schema is a policy violation Google penalizes.
The page has no structured data but its content clearly fits a type: an article, a product, a how-to with numbered steps. Adding the matching schema makes it eligible for a rich result it currently cannot get. This is an addition, not a fix, and often the highest-value one.
Schema must describe what is actually on the page.
Google's single firmest rule for structured data is that it must reflect the visible content of the page. A Product schema must sit on a real product page; a Review or aggregateRating must come from real reviews that are actually shown to users. Marking up content that is not there, or inventing a rating a page does not have, is exactly the kind of manipulation Google's spam systems are built to detect, and it can get structured data ignored site-wide or trigger a manual action.
This is why a missing-field warning is not something to paper over with a made-up value. If a page has no reviews, the honest fix is either to leave the enhancement unclaimed, or to switch to a schema type that does not require a rating. The warning is a prompt to add real data or the right type, never fake data.
Not directly. Google has said structured data is not itself a ranking factor: adding an Article schema will not lift a page's position. What it changes is the appearance of the listing once you rank, and a listing with stars, an image or a price attached earns a higher click-through rate at the same position.
So the value is real but indirect. Schema does not move you up the page; it makes your existing spot on the page work harder. That is often a larger practical gain than a small ranking change, because click-through compounds across every impression.
Schema can be added three ways, but Google recommends JSON-LD: a single script block in the head, separate from your HTML. It is easier to add, easier to maintain, and does not tangle markup into your page templates the way inline microdata does.
Keeping it in one block also makes it auditable. A crawler can read and validate a self-contained JSON-LD script far more reliably than markup scattered across dozens of HTML attributes, which is part of why validation errors are easier to trace when you use it.
After changing schema, validate it two ways: schema.org's own validator confirms the markup is well-formed, and Google's Rich Results Test confirms the page is eligible for a specific rich result and lists any remaining warnings. Only the second tells you whether Google will actually use it.
Then re-crawl. An audit that re-checks structured data on every run turns schema from a one-time task into something monitored, so a template change that quietly breaks a valid block is caught before it costs you rich results across a whole section.
Free to start. The audit validates JSON-LD and flags Google-style missing-field warnings per page.
Start my free audit