Google Just Updated Its Video Markup Rules. Does It Affect You?

Google Just Updated Its Video Markup Rules. Does It Affect You?

As of September 2026, Google has updated its VideoObject documentation to add the creator property and to spell out exactly which interaction statistics it supports. If you host video on your own pages, this is a small, cheap change worth making. If your video lives only on YouTube, it changes nothing for you.

The update landed on 24 September 2026 in Google's Search Central documentation changelog. It is not an algorithm update and nobody is going to lose traffic over it. It is a clarification, and clarifications are the most useful kind of documentation change, because they tell you what was previously ambiguous.

We want to be precise about what changed and what did not, because plenty of write-ups are about to overstate it.

What Exactly Did Google Change?

Two things. The documentation now describes support for the creator property, with a note about the author property alongside it. And it now states which interaction types Google actually understands inside interactionStatistic.

That second half is the more interesting one. Before, you could put an interaction count in your markup and hope Google made sense of it. Now the documentation lists four supported types: WatchAction for view count, LikeAction for likes or upvotes, CommentAction for comments, and ShareAction for reshares.

Four named types is a much better brief than "user statistics applied to the video." You now know what to emit and, just as usefully, what not to bother emitting.

Are the Required Properties Any Different?

No. The requirements are unchanged, and they are short. Google asks for name, which is the title of the video. It asks for thumbnailUrl, pointing at the video's own thumbnail image file. And it asks for uploadDate, in ISO 8601 format.

Everything else, including creator and interactionStatistic, sits in the recommended list along with contentUrl, description, duration, embedUrl, expires, hasPart, ineligibleRegion and regionsAllowed.

This matters for prioritising. If your video markup is currently missing a required property, fix that first and ignore the news entirely. Recommended properties improve how a result can appear. Required properties decide whether it can appear at all.

Why Would You Add the Creator Property at All?

Because it attaches the video to an identifiable person or organisation, and that attribution is the kind of signal both search and answer engines are increasingly built around. Google's documentation describes it as the person or organisation that created or published the video, given as a Person or an Organization with a name and a url.

That url subproperty is the part people skip, and it is the part that does the work. A name is a string. A name with a url is an entity you can connect to a profile, an about page, or an author page you already maintain.

We treat this the same way we treat article bylines. The point is not the label. The point is giving a machine one unambiguous way to decide who is behind the content, which is exactly what our guide to schema markup argues is the whole job of structured data.

Should You Publish Your View and Like Counts?

Only if they flatter you, and only if they are real. This is the honest part of the interactionStatistic conversation that vendor blog posts tend to skip.

A video with a large, genuine view count benefits from saying so. A video with forty views does not, and marking it up invites a comparison you will lose. There is no penalty for omitting a recommended property.

The hard rule is that the number has to be true and it has to stay true. If your markup says one thing and your page says another, you have created a mismatch a reviewer or an automated check can find. If your counts are static because nobody wired them to anything, they will be wrong within a month. Either generate them from the real source or leave the property out.

What Does This Mean If Your Video Is on YouTube?

Very little. If a video exists only as a YouTube URL, YouTube handles its own structured data and its own appearance in search. You are not competing for that result and you cannot improve it with markup on your site.

Where it does matter is the page on your own domain that embeds the video and surrounds it with content. That page can carry VideoObject markup describing the embedded video, and it can rank on its own terms for queries your YouTube listing will never reach.

The practical split is simple. Video hosted on your infrastructure, or embedded on a page where the video is the point, deserves markup. A video casually dropped into a blog post does not need it.

What Are Clip and SeekToAction Actually For?

Key moments. Both are ways of telling Google that specific timestamps inside a long video answer specific things, so a result can deep link into the middle of it rather than to the start.

Clip is the explicit version. You nest it inside your VideoObject, and each clip needs a name, a startOffset and a url, with endOffset optional. You are hand labelling the chapters.

SeekToAction is the automatic version. Instead of listing moments, you tell Google the URL pattern your player uses for timestamps, using potentialAction with target and a startOffset-input, and let Google work out the moments itself.

Pick Clip when you have a webinar or a demo with genuinely distinct sections that you are willing to label. Pick SeekToAction when you have a long back catalogue and no appetite for labelling any of it.

Does Any of This Help You Get Cited in AI Answers?

Indirectly, and not as much as the transcript does. Structured data tells a machine what a thing is. It does not tell a machine what the video says. An answer engine that wants to quote you needs words.

So the ordering we use has not changed. Publish a real transcript or a written summary on the page, in text a parser can read. Then mark up the video properly. The markup makes the page legible; the transcript makes it quotable.

We have written more about that division of labour in our piece on whether video content can get you cited in AI answers, and the short version is that the text around the video does most of the work.

What About Live Video?

There is a separate mechanism for it, and it is easy to get wrong. To be eligible for a LIVE badge, Google looks for BroadcastEvent nested in the publication property, with isLiveBroadcast, startDate and endDate.

The trap is endDate. A broadcast that has finished but still claims to be live is a mismatch between your markup and reality, and it is the sort of thing nobody notices for months because the page looks fine to a human.

If you run webinars, wire these dates to whatever system schedules them. If you run one live event a year, set the dates by hand and put a reminder in the calendar to correct them afterwards.

What Should You Actually Do This Week?

Almost certainly nothing urgent. This is a documentation clarification, not a deadline. But it is a good prompt for a ten minute check.

Find your pages that host or embed video as their main content. Confirm each has name, thumbnailUrl and uploadDate. Add creator with a real url pointing at a page you control. Add interaction counts only where the numbers are both genuine and generated from a live source. Test each page in Google's Rich Results Test and see what it reports back.

Then leave it alone. Structured data is maintenance work, and the teams that do best with it are the ones who set it up correctly once and wire it to real data, rather than the ones who revisit it every time a changelog entry appears. If your video pages are slow as well as under-marked, that is the bigger problem, and our notes on video and page performance cover it. If you would rather hand the whole thing to someone, we are happy to talk it through at phoenix.studio.