Colquhoun.AI
← All guides

Research Guides

Structured Data QA: A Worksheet for Claims, Code and Search

· Colquhoun.shop editorial

A structured data check should leave you with a defensible account of what the page says, what the markup says and what a test actually verified. A green result alone leaves several questions unanswered: is the named author correct, is the information current, and did you test the page visitors now receive?

This manual review method is designed for a small website or a clearly defined sample of a larger one. It combines a reusable evidence worksheet with a practical release check. It does not require a paid SEO platform, and it does not claim that Colquhoun.shop runs these checks automatically.

1. Define the page and the intended search feature

Start with one page template and one reason for reviewing it. For an article, that might be checking whether the author, headline and dates agree with the published story. For a product page, it could be checking whether the offer information describes what a customer can actually buy.

Structured data gives machines explicit information about a page and its content. Google uses it to understand pages and may use it for enhanced search appearances. It belongs on the page the information describes. Google's introduction to structured data

Write down the relevant Google feature guide before choosing properties. The current Search Gallery links to the supported features and their implementation requirements. A type existing in Schema.org is not, by itself, evidence that Google offers a particular search presentation for it. Google's supported structured data features

Keep your initial scope concrete: “Five recently edited article pages using the current publishing template” gives a reviewer a useful boundary. It does not justify a claim that every page on the site has passed.

2. Compare each meaningful value with the visible page

Read the page before reading a test summary. Then inspect the markup and ask where each important value comes from. Does the author match the byline? Does the headline identify this story? Do product details describe the selected item rather than a different variation?

Google's quality guidelines require relevant, up-to-date markup that represents the page content. Hidden, misleading or invented content can prevent eligibility even when the syntax is correct. These quality questions are not fully resolved by automated technical tests. Google's structured data policies

For an article, Google's documentation distinguishes the publication date from the most recent modification date and describes author identification. Use the actual editorial record when resolving a mismatch. Do not change a date merely to make an old article appear new. Google's Article documentation

Trace each discrepancy to its source. An outdated value might come from a template default, an old CMS field or a second plugin generating overlapping markup. Record both the symptom and the likely source, then confirm the source before changing it. Deleting a second block without understanding its role can remove useful information as well as the conflict.

3. Copy this evidence worksheet

Use one row per finding. Several findings may refer to the same page or item. Keep a separate list of all reviewed URLs so a page with no finding is still counted.

This is an evidence log, not a numerical SEO score. A blank field should mean “not recorded” rather than “passed.” If you cannot access a page or its markup, retain that row as unverified.

Keep private previews, unpublished customer information and credentials out of public testing services. Use local validation for private material, and submit only information you are authorised to share.

4. Match the validator to the question

The Schema.org Markup Validator extracts structured data and identifies syntax problems. It is useful for seeing the data that was recognised. Record what it found rather than treating that output as a Google appearance promise. Schema.org's validator documentation

Google's Rich Results Test accepts either a URL or pasted code. It reports detected supported items, errors and suggestions, and allows smartphone or desktop testing. For a public page, save the chosen mode and device with the result. Google's Rich Results Test guide

A code test checks the submitted sample. A URL test checks what the service can retrieve at that address. If you fix a code sample, deploy through the site's normal authorised process and test the public URL again. Compare the resulting values with the worksheet; a corrected draft does not establish that a template change reached production.

Separate three outcomes in your notes: the page could not be retrieved, the markup could not be parsed, or a detected item failed a feature requirement. They lead to different investigations. For a “no items detected” result, confirm the expected type and inspect the tested content before concluding that no structured data exists anywhere on the page.

5. Keep page counts, item counts and factual accuracy separate

Fictional example: A team checks three public pages. Each contains two structured data items, so the review covers six items. The tests report five without critical errors and one with a critical error. A manual comparison also finds an outdated author value in one of the five technically valid items.

The review summary is: “Three pages reviewed, six items checked, one technical failure and one factual correction identified.” It should not become “five pages passed,” because the unit was an item. Nor should it say that all five technically valid items accurately describe their pages.

One completed worksheet finding could read: “Example article; BlogPosting author.name; visible byline Ava Taylor; markup says Editorial Team; no technical error reported; editor confirms Ava Taylor is the credited author and the stored field is outdated; content owner to correct that field; recheck the deployed byline and markup together.” All names and findings in this example are fictional. The row connects a discrepancy to a responsible person and a specific recheck instead of stopping at “schema passed.”

Search Console's rich result reports likewise count items rather than pages and show a sample rather than a comprehensive inventory. An item can have multiple issues, so adding issue counts can count the same item more than once. Keep the exact report and its units alongside any exported figures. Search Console's rich result report guide

For a small audit, report coverage plainly: pages selected, pages retrieved, items reviewed and unresolved findings. Avoid blending a local sample with Search Console totals into one unexplained percentage.

6. Close the review with a dated deployment check

After an approved correction, inspect the affected live pages and at least one unchanged example using the same template. Recheck the visible content as well as the markup. Save the before-and-after evidence and note whether the change resolved a technical error, a factual mismatch or both.

For a site you manage, URL Inspection can show Google's indexed information and provide a separate live test. The indexed information reflects a previously processed version; it is not a live reading of a change you just shipped. Record the relevant crawl and test dates. Google's URL Inspection guide

Even correctly implemented structured data does not guarantee a rich result. Close the task with the checks you completed and any questions still open, rather than a promise of rankings, clicks or a particular search appearance. A useful handoff lets the next person reproduce the review and understand exactly what the evidence supports.

This guide describes the built-in Web Reader and Keyword Density behavior verified for this publication. Available data depends on the page response; these tools do not measure visits or search rankings.