Video SEO: ranking on YouTube, and getting into Google results
Google will not index your video until the page it sits on is already ranking, and YouTube never asks you for schema markup. Two systems, two checklists, almost no overlap.
8 minute read
What YouTube says it ranks on
YouTube's help documentation on how its search works is unusually direct. It names three things: relevance, engagement and quality. On relevance it says the system looks at many factors, including how well the title, tags, description and video content match the query. Engagement is described as overall user engagement, such as watch time for a particular video against a particular query. Quality is about signals that suggest a channel demonstrates expertise, authority and trust on a topic. Two details are worth noticing. First, video content sits alongside the metadata, which is why what you say on camera matters to search. Second, engagement does the heavy lifting: no metadata tuning survives viewers leaving in ten seconds. YouTube's own thumbnail and title guidance adds the blunt version, that 90% of the best-performing videos have custom thumbnails.
For Google, the page comes first and the video comes second
This is the part that surprises people. Google's video SEO documentation says that only videos on a watch page are eligible for indexing, and describes a watch page as one whose main purpose is showing a user a single video. It then adds two conditions people miss: the watch page has to be indexed and already doing well in Search before Google will consider the video for indexing, and having the page indexed is no guarantee the video will be. In practice that means video markup is not a shortcut around ordinary SEO. If the page has no links, no text and no rankings, adding schema will not conjure a video result. Get the page ranking as a page, then make the video on it legible.
The minimum useful VideoObject markup
Google's structured data reference for video is short enough to implement in an afternoon. Three properties are required: name, thumbnailUrl and uploadDate in ISO 8601 format. You then need either contentUrl, which Google prefers because it points at the actual video file, or embedUrl for the player. Description and duration are listed as recommended and are worth adding anyway, since both are what a search result ends up displaying. Two constraints catch people out. The markup has to live on a page where the video can actually be watched, and Google states the total video duration must be a minimum of 30 seconds. If you would rather not touch your templates, Google says it also accepts video sitemaps and Open Graph tags as ways to supply the same metadata.
Your thumbnail is a technical requirement, not just an image
A striking share of video indexing failures are thumbnail failures. Google's documentation asks for a single unique and stable thumbnail URL per video, and says explicitly that if the thumbnail URL changes too often the video may not be indexed, which quietly rules out cache-busted or session-signed image URLs. The file has to be reachable by both Googlebot and Googlebot Images, so no robots.txt block and no login wall. Accepted formats include BMP, GIF, JPEG, PNG, WebP, SVG and AVIF, the minimum is 60 by 30 pixels with larger preferred, and at least 80% of the pixels must have an alpha value above 250, meaning near-opaque. Search Console's video indexing report lists each of these as its own named failure, including no thumbnail URL provided and thumbnail blocked by robots.txt.
The video has to be visible, not merely present
Google's guidance is that you should load the video player at its actual size and position when the page loads, without requiring any user interaction, and that you should not rely on clicks, swipes or typing to load it. A player that only appears after someone dismisses a modal, or that sits far below the fold, may simply not be seen. This tightened further in 2023. Google announced in April that year that video thumbnails would appear beside a search result only when the video is the main content of the page, retiring the older format where a thumbnail appeared after a listing for pages with incidental video, and in December it extended the same rule to Video mode. A video buried three sections down a blog post is now, for Google's purposes, close to invisible.
Key moments, transcripts, and what Google does not ask for
Key moments let a result expand into jump links. Google supports two routes: Clip structured data, where you supply exact start times and labels, and SeekToAction, where you tell Google your player's URL pattern and let it work out the segments. If the video is on YouTube, timestamps in the description do the same job with no markup at all. Worth saying plainly: Google's video best practices page does not mention transcripts or captions anywhere. The near-universal advice to publish a transcript is not a video indexing requirement. It is page-level SEO, and it is still worth doing, because the transcript is often the only real text on a watch page, and the page has to rank before the video is even eligible.
A workflow that feeds both systems
The practical setup for most creators is one long video published to YouTube, embedded on your own watch page with VideoObject markup and a transcript beneath it, and then cut into vertical clips for the feeds. That single upload can earn YouTube search rankings, a Google video result through your own page, and a stream of short clips that send people back to it. The clipping step is where the time goes, which is where a tool like FrameOS helps: it reads the transcript, surfaces the strongest self-contained moments, reframes them vertically and burns captions, so the repurposing is a review job rather than an afternoon of scrubbing. Free plan is 300 credits for 3 days · no card.
FAQ
Will adding video schema make my video rank on Google?
Not on its own. Google's documentation says the watch page has to be indexed and already doing well in Search before the video is even considered, and that an indexed page is no guarantee of an indexed video. VideoObject markup makes an eligible page legible to Google. It does not create eligibility where the page has no rankings.
Do I need a video sitemap if I already have VideoObject markup?
No. Google lists structured data, video sitemaps and the Open Graph protocol as three supported ways to supply the same metadata. Pick whichever fits your stack. Structured data on the watch page is usually simplest because it lives with the content it describes and is easier to keep in sync.
Do transcripts help video SEO?
Indirectly. Google's video best practices documentation does not mention transcripts or captions at all, so they are not an indexing requirement. They help because a transcript is often the only substantial text on a watch page, and that page has to rank before the video becomes eligible. On YouTube, spoken content matters separately: YouTube says relevance is judged partly on video content, not just metadata.
Why did Google stop showing a thumbnail next to my page?
Most likely because the video is not the main content of the page. Google announced in April 2023 that thumbnails would only appear beside results where the video is the primary content, removing the older format for pages with incidental video, and extended the same rule to Video mode that December.
Sources
- Video SEO Best Practices · Google Search Central
- Video (VideoObject, Clip, BroadcastEvent) Schema Markup · Google Search Central
- Video indexing report · Search Console Help
- Simplifying video presentation on Google Search Results · Google Search Central Blog
- How YouTube search works · YouTube Help
Related pages
Keep reading
More Video Basics →Social media video specs for every platform
One 1080x1920 export covers most of short-form. This is where that stops being true — the length caps, file limits, and the platforms where the number changes depending on how you upload.
8 minute readSocial media image sizes for every platform
Instagram's help page now allows 3:4, YouTube's recommended thumbnail is 4K, and TikTok publishes almost nothing. Here are the numbers each platform actually documents, and where they leave you guessing.
7 minute readHow to convert landscape video to vertical
A full-height 9:16 crop of a 1920x1080 frame is 607 pixels wide. Which third of the picture you keep, and whether it is allowed to move, is the entire job.
8 minute readTurn one long video into a week of posts
300 credits for 3 days · no card.
Start free