How Video Schema Markup Gets Your Videos Cited by AI
- Harold Bell
- Jul 22
- 6 min read
Updated: Jul 23
Key takeaways
|
Search engines and AI systems can't watch video. Every decision they make about your embed, whether it ranks, whether it earns a video rich result, whether an AI answer cites it, comes from the text and structured data around the player. Video schema markup is the structured half of that equation, and after more than 16 years of shipping content for enterprise brands, I can tell you it's the highest leverage twenty minutes in video publishing.
The stakes went up with AI search. YouTube citations in Google's AI Overviews grew 34% in six months according to Search Engine Journal data, and the selection signals are structural, metadata, chapters, transcripts, not popularity. VideoObject is how you hand engines that structure on your own domain. Here's the complete implementation, including the copy ready code and the Wix specific deployment steps.
What is video schema markup
Video schema markup is structured data, usually JSON-LD using the VideoObject type from schema.org, added to pages that host video. It describes the video’s title, description, thumbnail, upload date, duration, and location in a machine readable format, enabling video rich results and helping AI engines understand and cite the content. |
Think of it as the difference between an engine inferring what your video might be from surrounding text and being told outright. Google uses VideoObject to qualify pages for video rich results, key moments, and carousel placements. AI engines use the same structured facts when deciding what a page contains and whether it answers a query, the mechanism we unpacked in our broader guide to getting cited by AI.
It belongs to the same family as the other structured data on your site, and the general principles from our schema markup examples library apply here, JSON-LD format, one coherent graph per page, and claims that match visible content exactly.
How does VideoObject schema work
VideoObject schema works by declaring a video’s properties in a JSON-LD block in the page head. Crawlers parse the block, associate the declared facts with the embedded player, and use them to index the video, qualify it for rich results, and inform AI systems assembling answers about the topic. |
Here's a complete, working example for a video embedded on a blog post. Swap the placeholder values and it's production ready.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "VideoObject",
"name": "How to Build a B2B Video Marketing Strategy",
"description": "A practical walkthrough of building a video
marketing strategy that drives pipeline for B2B tech teams.",
"thumbnailUrl": "https://www.example.com/thumb.jpg",
"uploadDate": "2026-07-22T08:00:00-05:00",
"duration": "PT8M12S",
"embedUrl": "https://www.youtube.com/embed/VIDEO_ID",
"publisher": {
"@type": "Organization",
"name": "MQL Magnet",
"logo": {
"@type": "ImageObject",
"url": "https://www.example.com/logo.png"
}
}
}
</script>
A few mechanics worth knowing. duration uses ISO 8601 format, so PT8M12S means eight minutes twelve seconds. embedUrl points to the player, contentUrl points to the actual video file, and you should provide whichever you have, or both. And the block can live in the same graph as your BlogPosting and FAQPage schema, which is how we deploy it on episode posts.
Which VideoObject properties matter most
Required for eligibility
Google requires name, thumbnailUrl, and uploadDate for video rich results. Miss any and the markup validates but earns nothing.
Strongly recommended
Description, duration, and embedUrl or contentUrl. The description is doing more work than it used to, since AI systems weigh declared descriptions when matching videos to queries.
The citation multipliers
The hasPart property with Clip objects declares chapters, each with its own name and start offset. This mirrors on platform chapters, which OtterlyAI's research shows are the strongest structural correlate of repeated AI citation, with 78% of timestamped videos cited across two to five distinct chapters. The transcript property, while not required for rich results, hands engines the full text, and pairing it with a visible on page transcript is the strongest configuration.
SeekToAction
For videos on your own player infrastructure, this enables Google's key moments without manual chapter declaration. Most B2B teams embedding from YouTube can skip it.
Does video schema help AI citations
Yes, indirectly and meaningfully. AI engines select citations based on structural clarity, transcripts, chapters, and declared metadata rather than popularity. VideoObject schema supplies that declared structure for embeds on your domain, making your video pages easier to understand, match to queries, and cite in generated answers. |
The evidence for structure over popularity keeps piling up. WebFX reported analysis showing 40.83% of AI cited YouTube videos had fewer than 1,000 views at study time. BrightEdge data puts YouTube's citation share in Google AI Overviews at 29.5%, the largest of any domain. Low view videos getting cited at that scale tells you the selection mechanism reads structure, and schema is structure you control completely.
There's also an academic anchor here. The Princeton GEO study presented at ACM KDD found that content with authoritative citations and clear structure increased AI visibility by as much as 115% for some optimizations. Structured, well sourced pages get quoted.
VideoObject extends that machine readability to the video layer, which is why it's a standing item in our video SEO services scope and a lever in every engine of the Engine Optimization Matrix.
How to add video schema to your site
On Wix, which is what we run
Open the page's SEO settings, go to Advanced SEO, then Custom Meta Tags, and paste the JSON-LD block there. Never put schema in an HTML embed element, because Wix renders embeds in sandboxed iframes that crawlers don't read. This is the same deployment path we use for every schema type on the site, including the FAQ schema on our question driven posts.
Everywhere else
WordPress accepts JSON-LD through most SEO plugins or directly in the head. Custom stacks should render it server side in the initial HTML rather than injecting it client side, since some crawlers and most AI retrieval systems read the raw response.
Then validate
Use validator.schema.org for structural correctness. Google's Rich Results Test still reports video eligibility and is worth a check for rich result features specifically, but schema.org validation is the source of truth for the markup itself, especially for properties Google's test ignores.
Common video schema mistakes
Declaring metadata that doesn't match the visible page is the big one, engines treat mismatches as spam signals. Marking up pages where video isn't the primary content earns nothing and dilutes trust. Missing uploadDate silently kills rich result eligibility.
Client side injection means many crawlers never see the block. Schema on a page with no transcript wastes half the opportunity, since the declared structure points at text that isn't there. And forgetting to update duration and thumbnailUrl when you swap a video leaves engines describing content that no longer exists.
None of these are exotic. They're publishing hygiene, which is the actual secret of most technical visibility work, done consistently on every video page rather than heroically on one.
Frequently asked questions
What is video schema markup?
Video schema markup is structured data, typically JSON-LD using schema.org’s VideoObject type, added to pages hosting video. It declares the video’s title, description, thumbnail, upload date, duration, and location so search engines and AI systems can understand, index, and cite the content.
Is VideoObject schema required to rank videos?
It is not strictly required, but Google requires name, thumbnailUrl, and uploadDate in VideoObject markup for video rich result eligibility, and pages without it compete at a disadvantage. For AI citation purposes, declared structure is one of the few selection signals you fully control.
Does video schema markup help with AI citations?
Yes. AI engines select sources based on structural clarity rather than popularity, and studies show large shares of AI cited videos have minimal views. VideoObject supplies declared, machine readable structure for video embeds on your domain, improving how reliably engines match and cite your pages.
Which VideoObject properties are required?
Google requires name, thumbnailUrl, and uploadDate for video rich results. Strongly recommended additions are description, duration, and embedUrl or contentUrl. For citation performance, add hasPart chapter declarations and pair the markup with a visible transcript on the page.
How do you add video schema in Wix?
Open the page’s SEO settings, choose Advanced SEO, then Custom Meta Tags, and paste the JSON-LD block. Do not use an HTML embed element, because Wix sandboxes embeds in iframes that crawlers cannot read. Validate the deployed page at validator.schema.org.
Can you combine VideoObject with BlogPosting and FAQPage schema?
Yes, and you should when a post contains a video. Place all types in a single @graph within one JSON-LD block so engines receive one coherent description of the page, the article, its FAQs, and the embedded video, with no conflicting duplicate declarations.
What is the hasPart property in video schema?
hasPart declares chapters within a video as Clip objects, each with a name, start offset, and URL. It mirrors on platform chapters and matters because chaptered videos function as multiple citable units, with research showing timestamped videos frequently earn citations across several distinct chapters.
How do you validate video schema markup?
Validate at validator.schema.org for structural correctness of the markup itself, then optionally check Google’s Rich Results Test for video rich result eligibility. Always test the live published URL rather than pasted code, since deployment method determines whether crawlers actually receive the block.
Comments