Google Demoted Your Page for a Schema Error You Can’t Find

Your markup passes every validator, the Rich Results Test reports no errors, and yet the rich results are gone. First, be precise about what was lost. Google’s structured data guidelines say a structured data manual action means a page loses eligibility for appearance as a rich result, and that it doesn’t affect how the page ranks in Google web search. So the stars and enhancements disappeared, not necessarily your position. Second, one reason you may not find the error is that it isn’t a syntax error. Validators check structure and required properties; they don’t check whether the markup tells the truth. When accuracy is the cause, recovery means auditing every property for accuracy and keeping only what is true.

Start by ruling out the explicit case. Open the Manual Actions report in Search Console. Google’s Manual Actions report help lists structured data issues such as “Content on page differs from structured data,” and says to select Request Review once all listed issues are fixed on all pages. If there is none, you are in the harder case: the markup is valid, nothing is reported, and the rich results still don’t show.

Two layers: technical validity and content accuracy

Structured data is judged on two axes that are easy to collapse into one.

Technical validity: is the JSON-LD well formed, are the required properties present, does it parse? The Rich Results Test measures this, and passing it is necessary but not sufficient.

Content accuracy: does the markup describe what is on the page, and is it true? Google’s guidelines say it doesn’t guarantee that correctly marked-up structured data will show, and list among the reasons that the structured data is not representative of the main content of the page, or is potentially misleading. A plugin can pass validation while outputting whatever it is fed, accurate or not. The markup can be correct as code and wrong as a claim.

The clearest trigger: ratings that aren’t ratings

An aggregate rating that didn’t come from real reviews is the plainest accuracy failure. Google’s guidelines say not to mark up misleading content such as fake reviews, and note that reviews or ratings not by actual users may result in a manual action. Its review snippet guidelines add that fake or undisclosed incentivized reviews don’t belong on the page or in the markup, and that if you use an aggregate rating, users should be able to see it on the page.

Each of these passes validation and fails accuracy:

• A rating computed from engagement, such as saves, prints or time on page, presented as a user rating.
• A seeded baseline so nothing ever shows as unrated.
• A “quality score” exposed as a star rating.
• A count mismatch, where the markup claims more ratings than the page shows.

The number is a valid number; it doesn’t represent what it claims to. There is also a separate rule for businesses: when the entity being reviewed controls the reviews about itself, pages using LocalBusiness or Organization markup are ineligible for star reviews, however real the reviews are.

Other accuracy traps a syntax check won’t surface

Any property that misstates the page belongs in the same category:

• Counts beyond reviews: a recipe that claims more steps or servings than the page shows, or an item list whose count doesn’t match its items.
• Unrealistic values: a cook time shorter than anyone can achieve.
• Incomplete markup: Google’s guidelines say that if a page includes multiple reviews, the markup should include all the reviews visible on the page, because marking up only part of them would mislead people.
• Aspirational properties: authorship with no real author behind it, or any property that describes what you wish were true rather than what is on the page.

None of these is a syntax problem, so none shows up in a tool that only checks syntax.

A long history of clean rich results isn’t a defense. Years of working markup don’t prove the markup is accurate; they show only that nothing has stopped it so far. “It worked for years” describes the past.

Fix the source, not the symptom. Inaccurate values can come from a template, a plugin setting or a feed, which means the same inaccuracy sits on every page that uses it. Correcting the three worst pages while the template keeps producing the same values on three hundred others leaves the problem in place. Find where each value is generated and correct it there.

Bad data can cost more than no data

Treat accurate-but-absent and inaccurate-but-present as different states. A page with no markup simply isn’t eligible for the rich result. A page whose markup misrepresents it gives Google grounds not to show the rich result and, in the explicit case, grounds for a manual action. It is closer to a bad credit score than to no credit history.

That changes the fix. The tempting fix is to tweak properties so they look better; the durable move is to make every property true, even when that means a lower number or no number.

Recovery: audit, strip, keep the truth, wait

Outside the manual action case there is no button to press, so recovery is a discipline:

1. Audit every property against the visible page, one by one. For each claim the markup makes, confirm the page shows it and that it is true.
2. Strip any rating not sourced from real users, and any property you can’t substantiate. Where the choice is an inaccurate value or no value, choose no value.
3. Keep the corrected markup rather than deleting all structured data. Removing it all makes you ineligible; accurate markup keeps you eligible.
4. Wait for Google to recrawl the pages. If there was a manual action, request review after fixing. If there wasn’t, there is no confirmation message; you will know it worked when rich results return.

Frequently asked questions

Should I file a reconsideration request?

Only if the Manual Actions report shows a structured data manual action. Reconsideration applies to manual actions. If the report is empty, there is nothing to file; fix the accuracy and let Google recrawl.

Will removing all my markup fix it faster?

It removes the inaccurate signal, but it also makes you ineligible for rich results. Keep the markup that is accurate and remove only what you can’t substantiate, so you remain eligible.