Rank Math turns on Article schema by default for every WordPress post, and Google’s own documentation says none of that schema’s properties are technically required — so it is easy to assume the default output is already fine. It usually isn’t: Google’s guidance lists several properties it recommends specifically because they unlock richer search results, and Rank Math’s defaults only cover some of them automatically. This is a property-by-property checklist comparing what Rank Math fills in on its own against what Google’s Article guidelines actually ask for, built from both parties’ current documentation rather than from a generic schema list.
Why “no required properties” is not the same as “nothing to check”
Google’s structured-data documentation for Article is explicit: there are no required properties, and the advice is simply to add whatever properties apply to the content. That framing matters because it shifts the question from “is my schema valid” (it probably is, since almost nothing is required) to “is my schema doing anything for me” — a technically valid Article block with just a headline is not going to be the thing that gets a story a larger image or an author byline in search results. The properties below are the ones Google calls out by name as recommended, and each row notes whether Rank Math adds it for you.
The property-by-property checklist
| Property | What Google recommends | What Rank Math does by default |
|---|---|---|
headline | A concise title; long titles may be truncated on some devices | Defaults to the post’s SEO title (%seo_title%); editable per post in the Schema Generator |
image | Multiple images, at least 50K pixels (width × height), in 16:9, 4:3, and 1:1 aspect ratios, hosted at crawlable, indexable URLs and representing the actual content (not a logo) | Not documented as covering multiple aspect ratios automatically — check your theme’s featured-image size against the 50K-pixel minimum before relying on it |
datePublished / dateModified | ISO 8601 format, with timezone included (otherwise Google falls back to Googlebot’s timezone) | Not documented as a configurable field in the Schema Builder — worth confirming in the rendered page source rather than assuming |
author | Person or Organization type; every author shown on the page listed as a separate author entry, with author.name containing only the name (no titles, no “posted by”) | Includes only the primary author by default; additional authors and reviewedBy require the PRO plan’s Advanced Editor |
author.url | A URL (or sameAs) that uniquely identifies the author | Added automatically only if author archive pages are enabled in WordPress — if archives are off, this property is silently missing |
publisher | An Organization (not a Person) | Added automatically; Rank Math enforces the Organization type for you here |
keywords | Not part of Google’s Article guidance at all | Defaults to the post’s focus keyword; Rank Math’s own documentation notes Bing treats meta keywords mainly as a spam signal, so this property is close to decorative for SEO purposes |
The two gaps worth fixing first
Of the rows above, two are worth checking on every post rather than assuming the default is correct. First, author archives: if your theme or a privacy setting has author archive pages disabled, Rank Math does not add author.url, and the Article schema ships with an author name but no way for Google to disambiguate who that is — the fix is either turning author archives back on or setting a custom author URL per post if your theme supports it. Second, image dimensions: Google’s 50K-pixel minimum (for example, roughly 250×200 or larger) and its three aspect ratios are about what image actually gets pulled into a rich result, and a featured image cropped tightly to one aspect ratio by your theme may not satisfy all three even if it clears the pixel minimum.
Multi-author posts hit a wall that single-author posts don’t
A site with one byline per post, which covers most of what Rank Math documents by default, never notices that only the primary author is included. A site that runs co-authored posts, or wants reviewedBy to show editorial oversight on AI-assisted content, runs into the PRO-only Advanced Editor immediately — the free plan’s Schema Builder does not expose a way to add a second author entry. If editorial review is part of your content process, as it should be for anything published at volume, this is worth confirming before publishing day rather than after, since the schema will validate fine with just the primary author and give no error indicating a second one is missing.
How to actually check what shipped, not what the settings page implies
- Run the live, published URL (not a draft preview) through Google’s Rich Results Test and read the parsed properties, not just the pass/fail banner — it shows exactly which of the properties above actually made it into the rendered page.
- Use Rank Math’s own Code Validation feature in the Schema tab as a faster first pass before going to Google’s tool, since it flags obviously malformed JSON-LD without a round trip to an external service.
- View the page source and search for
"@type":"Article"(orBlogPosting) directly — this is the only way to confirmdatePublished,dateModified, and the exactimageURLs Rank Math chose, since none of those are documented as editable settings. - Re-check after any theme change. Featured-image dimensions and whether author archives are enabled are both theme- and settings-dependent, not fixed by Rank Math, so a schema that passed last month can silently lose the
imageorauthor.urlproperty after an unrelated theme update. - Cross-check Search Console’s structured-data reporting periodically rather than only right after publishing — it surfaces properties Google saw at crawl time, which can lag a recent fix by days and will catch a regression the Rich Results Test alone, run once, would not.
Where this fits with the rest of a technical SEO pass
Article schema is one input among several into how a page is understood and displayed, not a ranking lever on its own. For the content-quality side of that same question, see auditing a site against Google’s scaled content abuse policy and signaling experience and expertise in AI-assisted content, both of which cover what schema alone cannot fix. For the crawler-facing side of the same page, robots.txt vs. meta robots vs. X-Robots-Tag and what the llms.txt spec actually requires round out the technical checklist beyond structured data. And if the goal behind cleaning up Article schema is building out a cluster of related posts rather than fixing one page in isolation, building topical authority with a content cluster strategy is the piece that ties individual technical fixes like this one to a larger plan.
Don’t switch the Article type without reading what it implies
Rank Math’s Schema Builder lets you pick among Article, Blog Post (BlogPosting), and News Article for the same piece of content, and it’s tempting to pick whichever sounds most authoritative. Don’t: Rank Math’s own documentation is explicit that NewsArticle is meant only for factual news coverage, that it is set automatically for posts marked as news in a News Sitemap, and that misusing it can produce errors in Google’s Rich Results Test. A how-to guide or an explainer — most of what a prompt-engineering or AI-tools blog publishes — belongs under Article or BlogPosting, not NewsArticle, even if the post covers something that happened recently. Picking the wrong type doesn’t just risk a validator error; it misrepresents the content to whatever is reading the schema.
Why this matters beyond the classic search result
Structured data was built for rich results in traditional search, but the same JSON-LD block is also one of the cleanest signals available to anything summarizing a page for an AI-generated answer: a clear headline, an accurate datePublished and dateModified, and a real author entry give a summarizer unambiguous facts to quote instead of forcing it to guess from the visible page layout. A post with a missing or stale dateModified is a specific, avoidable way to look out of date to any system deciding whether to cite it, independent of whether the on-page content was actually updated. None of the properties in the checklist above were written with AI answer engines in mind, but getting them right has the same effect either way: it removes one more source of ambiguity about what the page is, who wrote it, and when.
This is also where the gap between “valid” and “complete” schema costs the most. A validator only checks that the JSON-LD parses and that the types used make sense together; it has no way to flag an author entry that’s missing its url, or an image that technically exists but is cropped to one aspect ratio instead of the three Google’s guidance recommends. Those are exactly the kind of gaps that pass every automated check while still leaving real value on the table, which is the reason for checking the rendered output directly rather than trusting a green checkmark.
None of this is a one-time task. Schema output depends on settings — theme, author archives, image sizes — that change independently of the content itself, so a post that passed every check on its publish date can quietly drift out of compliance later without anyone editing the post. Treating the checklist above as a recurring spot check, not a one-off launch task, is the difference between schema that was correct once and schema that stays correct.
