Do University Lecture Videos Speak Google's Language? A Webometrics Study of VideoObject Structured Data Adoption

Executive Summary

On 24 September 2026, Google Search Central updated its VideoObject structured-data documentation, formally adding a creator property alongside the existing author property and clarifying that interactionStatistic supports four named interaction types WatchAction, LikeAction, CommentAction, and ShareAction in place of the older, retired interactionCount guidance. Universities routinely publish video content of exactly the kind this markup describes: recorded lectures, open courseware, admissions and research-explainer videos, most of it hosted directly on institutional domains. Webometrics attempted to examine 13 university web pages across eight countries for evidence of VideoObject structured data, collected 25 September 2026. Of the 13 pages attempted, 8 were confirmed, by direct inspection, to host an individually identifiable video on the page examined. Among those 8 confirmed video-hosting pages, none contained a JSON-LD structured-data block of any kind, let alone one describing a VideoObject. Consequently, none exposed VideoObject’s three required properties, and none exposed the creator/author or interactionStatistic properties Google’s documentation newly clarified. This study reports these as direct observations from a small, purposive, non-random sample collected on a single date; it does not, and by its design cannot, establish why markup was absent at any individual institution.

Key Findings

1. Of 13 university web pages Webometrics attempted to examine, 8 (61.5%) were confirmed, by direct inspection of the page’s HTML source, to host an individually identifiable video an embedded player, an iframe, or a directly offered video file on the specific page examined.
2. Among those 8 confirmed video-hosting pages, 0 (0%) contained a <script type="application/ld+json"> block of any @type at the time of collection, 25 September 2026 not only no VideoObject markup, but no JSON-LD structured data of any kind.
3. Because no JSON-LD was present, none of the 8 pages exposed VideoObject’s three properties Google’s documentation lists as required name, thumbnailUrl, and uploadDate in machine-readable form, and none exposed the creator/author or interactionStatistic properties Google clarified in its 24 September 2026 update.
4. The mechanism used to deliver video content varied across the 8 confirmed pages: 5 (62.5%) embedded the video through a third-party YouTube iframe (Massachusetts Institute of Technology, Yale University, the University of Melbourne, the University of Cape Town, and the Universidad Nacional Autónoma de México); 2 (25.0%) offered a directly downloadable video file without an embedded player (the University of Oxford and the University of Cambridge); and 1 (12.5%) used an institution-operated video-channel page (the University of Tokyo).
5. Five further institution pages examined in this study a course-catalog page, a news article, a feature page, and an institutional overview page did not themselves embed or offer an individual video on the specific page examined and were excluded from the video-hosting results; one additional page (Stanford Online) could not be retrieved at all, returning an HTTP 403 response, and is reported as inconclusive rather than as an absence of markup.
6. At the University of Oxford, descriptive video metadata episode duration, contributor names, and keywords was present as ordinary visible page text, but none of it was expressed in the specific machine-readable JSON-LD form Google’s documentation associates with video rich-result and video-search eligibility.

Why This Matters for Universities

Video is not a marginal format for university web presence. Recorded lectures, open courseware, admissions and campus-life videos, research-explainer content, and event recordings are common across institutional websites, and Webometrics’ own current methodology explicitly recognizes “video/audio, podcasts, digital publications, presentations, recorded lectures, educational media, public events” as part of the Distributed & Multimedia Presence indicator within Impact. Video is also an increasingly AI-search-relevant content type: video results, video carousels, and video-grounded AI answers all depend on a search or AI system being able to identify, describe, and attribute a specific video, which is precisely the function VideoObject structured data is designed to serve.

Google’s own documentation states, as a general principle for structured data, that publishers can help ensure the best representation of their content in Google Search results by using structured data. The question this study asks is narrow and directly testable: on pages where a university has already gone to the effort of publishing an individual video, has the institution also published the specific machine-readable description that lets a search or AI system understand what that video is, when it was published, who created it, and how audiences have interacted with it.

Authoritative Background

Google Search Central’s VideoObject structured-data reference page carries a confirmed “Last updated 2026-09-24 UTC” stamp, re-verified directly on the page before this study’s data collection began. Google’s own documentation-updates changelog records the corresponding entry for that date: “Added the creator property (noting support for author) and updated interactionStatistic to document supported interaction types in the VideoObject structured data documentation.”

The live page confirms three properties as required for VideoObject eligibility: name (“the title of the video”), thumbnailUrl (“a URL pointing to the video’s unique thumbnail image file”), and uploadDate (“the date and time the video was first published, in ISO 8601 format”). Among the recommended properties, the page describes creator and author as interchangeable: “creator (or author)… the person or organization that created or published the video,” presented as alternative properties rather than a mandatory pairing. For interaction data, the documentation states that Google supports four named interactionType values under interactionStatistic: WatchAction (“the number of times the video has been watched”), LikeAction (“the number of likes or upvotes”), CommentAction (“the number of comments”), and ShareAction (“the number of reshares”) the structured replacement for the older, simpler interactionCount property the documentation previously recommended.

None of this changes what a page must minimally include to be eligible for video-specific search treatment: the three required properties have not changed. What changed on 24 September 2026 is the documented vocabulary available to a publisher that wants to describe who created a video and how audiences have engaged with it both directly relevant to how recorded academic and institutional video content, most of it created by an identifiable university, department, or research group, could in principle be described.

Webometrics has previously examined general structured data and AI-crawler configuration at universities, including a study of robots.txt and llms.txt adoption across 18 research universities and a study of digital-presence signals mapped against Google’s helpful-content and structured-data documentation. Neither of those studies tested video-specific markup; this study is a distinct, video-focused extension of that existing structured-data coverage rather than a repeat of it.

Research Question

Among university web pages confirmed to host an individually identifiable video, do those pages implement VideoObject structured data including the properties Google’s documentation newly clarified on 24 September 2026 and if so, to what extent?

Methodology

Collection date: 25 September 2026, a single point-in-time snapshot, one day after the Google Search Central documentation update this study is framed against.

Sample: Webometrics attempted to locate and examine one video-content page at each of 13 research universities across eight countries and six world regions, drawing on institutions previously examined in Webometrics’ structured-data and AI-crawler-access research: Massachusetts Institute of Technology, Harvard University, and Stanford University (United States); the University of Oxford and the University of Cambridge (United Kingdom); Yale University (United States); the University of Tokyo (Japan); the Australian National University and the University of Melbourne (Australia); the National University of Singapore (Singapore); Sorbonne Université (France); the University of Cape Town (South Africa); and the Universidad Nacional Autónoma de México (Mexico). This is a purposive convenience sample selected for the practical ability to locate an identifiable, publicly accessible video page on the institution’s own domain within this study’s tooling; it is not a random or Webometrics-rank-stratified sample, and its findings should not be generalized beyond the institutions examined.

Selection method: for each institution, Webometrics searched for a plausible on-domain video content page an open courseware lecture, a recorded-lecture or podcast series, or a news/research-communication article and, where more than one candidate existed, selected the first page that could be retrieved and directly inspected, without regard to whether it was expected to carry structured data.

Inclusion rule: a page was counted in the video-hosting results only where direct inspection of the page’s HTML source confirmed the presence of an individual video an embedded player (iframe or <video> tag) or a directly linked, downloadable video file on that specific page. A page discussing video content, linking to a general video portal, or describing a video-related initiative without itself embedding or offering a specific video was not counted.

Variables: presence of any <script type="application/ld+json"> block, of any @type; presence of a VideoObject-typed block specifically; presence of the three properties Google documents as required (name, thumbnailUrl, uploadDate); presence of creator and/or author; presence of interactionStatistic and its named interaction types; and the video delivery mechanism observed.

Evidence source: direct retrieval and inspection of each page’s HTML source, collected 25 September 2026, cross-referenced against Google Search Central’s live VideoObject documentation and its 24 September 2026 changelog entry, both independently re-verified before this study’s findings were finalized.

Analysis approach: descriptive frequency counts with stated denominators; no statistical inference; no causal interpretation of any pattern.

Limitations: disclosed in full below.

Results

Table 1. University pages confirmed to host an individual video (8 of 13 attempted), collected 25 September 2026

Table 2. Pages attempted but excluded from the video-hosting results (5 of 13 attempted)

Across all 8 confirmed video-hosting pages, the result for every structured-data variable examined was the same: no <script type="application/ld+json"> block of any kind was present, and therefore none of VideoObject’s required or newly documented recommended properties could be found.

Analysis

The single most consistent finding in this study is not that universities implement VideoObject incompletely, but that among the pages examined, none implemented it at all and, further, that none of the 8 confirmed video-hosting pages carried JSON-LD structured data of any type on the video-hosting page itself. This is a narrower and more specific finding than “universities have poor structured data,” which prior Webometrics research has not supported as a general statement; it is a finding about one particular markup type, on one particular page per institution, on one collection date.

A recurring pattern in the results is the reliance on third-party video platforms. Five of the 8 confirmed pages delivered video through an embedded YouTube iframe rather than a locally hosted player. Google’s VideoObject documentation does not require a video to be self-hosted, and structured data can in principle describe a video regardless of where the underlying file is served from but the JSON-LD block itself, where present, is a property of the page it sits on, not of the video platform. A university page that embeds a YouTube video without adding its own VideoObject markup leaves the page’s own machine-readable description of that video absent, even though a description of the same video may separately exist on YouTube’s own page. This study did not examine YouTube’s own structured data for the videos referenced, since the research question concerns the university’s own domain and its own institutional web presence, consistent with Webometrics’ Impact and Openness indicators, both of which are attributed to the institution’s own eligible web assets.

The two pages that offered direct video downloads without an embedded player the University of Oxford and the University of Cambridge illustrate a related distinction. Both pages presented rich, human-readable descriptive information about the recording (a title, a lecture series, contributor names, and, at Oxford, an explicit duration field under “Episode Information”), demonstrating that the underlying descriptive data an institution would need to populate VideoObject’s required and recommended properties was already available to the page’s own content-management system. What was absent was not the information itself, but its expression in the specific JSON-LD format Google’s documentation specifies.

This study’s design does not permit a causal explanation for the uniform absence observed. Plausible, non-exclusive explanations include that structured-data implementation has, to date, more commonly targeted organizational, course, or article-level schema types than video-specific schema; that content-management or video-hosting platforms in common institutional use may not generate VideoObject markup by default; or that video content is frequently managed by units (communications offices, teaching-and-learning centers, library media services) that operate separately from the technical teams responsible for a site’s broader structured-data implementation. Distinguishing between these, or other, explanations would require evidence beyond what this study collected.

University Examples

The University of Oxford’s “Creative Media Lecture 01” page, part of its long-running podcast series, presents duration, contributor, and keyword information under a clearly labeled “Episode Information” section and offers separate video and audio download links for the same recording a page evidently designed with careful, human-readable metadata, but one whose HTML source, at the time of collection, contained no JSON-LD structured data of any kind.

Yale University’s Open Yale Courses page for PSYC 110, Lecture 1, embeds the lecture recording through a YouTube iframe positioned within the course page’s own layout, alongside a lecture transcript and course materials also hosted on the Yale domain. The page’s own JSON-LD, at the time of collection, did not describe the embedded video.

The University of Cape Town’s news article on a doctoral researcher’s work similarly embeds a YouTube video within an otherwise conventional news-article layout. As with the other examples above, this is reported as a direct observation of the page’s published configuration on the collection date, not as an evaluation of the page’s overall quality or communications effectiveness.

Implications for Higher Education

University communications, library media, and digital-strategy teams that already produce or curate video content have, on the evidence of this small sample, an available and currently unused opportunity: the descriptive information VideoObject requires a title, a thumbnail image, a publication date, and, optionally, a creator and interaction statistics closely matches metadata that a video-publishing workflow typically already captures for editorial or cataloguing purposes.

IT and web-platform teams may find it useful that Google’s documentation permits VideoObject markup on a page regardless of whether the video itself is self-hosted or embedded from a third-party platform such as YouTube, meaning that adopting this markup does not require migrating away from existing video-hosting arrangements.

Research communications offices in particular may want to note that the creator property, newly clarified alongside author in Google’s 24 September 2026 update, is well suited to attributing a research-explainer or lecture video to the specific researcher, department, or laboratory that produced it an attribution that machine-readable markup can express even where a video is otherwise presented without a prominent named byline.

Recommended Actions

• Inventory pages that already host an individual, institution-relevant video recorded lectures, open courseware, admissions and research-explainer content and check whether each carries VideoObject JSON-LD, using Google’s required-property list (name, thumbnailUrl, uploadDate) as the minimum baseline.
• Where a video is embedded from a third-party platform, add VideoObject markup to the institution’s own page rather than relying solely on the platform’s own structured data, since Google’s documentation ties the markup to the page it appears on.
• Where creator attribution is meaningful a specific researcher, department, or production unit use the creator (or author) property Google’s 24 September 2026 update newly clarified, rather than leaving the video’s creator undocumented in machine-readable form.
• Where genuine, available interaction data exists (view counts, likes, comments, shares), consider interactionStatistic with the four interaction types Google’s documentation names, rather than the retired interactionCount property.

These actions are not presented as guaranteeing improved search visibility, video rich-result eligibility, AI-citation frequency, or ranking outcome. No source cited in this study makes such a claim, and this study did not test for one.

Limitations

This study examined a small, purposive sample of 13 attempted pages at 13 institutions, of which 8 were confirmed to host an individual video and are the basis for this study’s quantitative findings; this sample cannot be generalized to the global university population or to any specific country’s higher-education sector. Only one page was examined per institution; a different video page on the same domain, or a different video platform used elsewhere on the same site, could show different markup, and this study’s findings describe the specific pages examined, not each institution’s website as a whole. Five pages were excluded because the specific page selected did not itself embed or offer an individual video; this reflects the page selected for examination, not a conclusion about whether video content with structured data exists elsewhere on that institution’s domain. One page (Stanford Online) could not be retrieved due to an HTTP 403 response and is reported as inconclusive rather than assumed to lack markup. This study’s page-retrieval method inspects a page’s delivered HTML source; if a page relies on client-side JavaScript to inject JSON-LD structured data after the initial page load, this study’s tooling may not detect markup added that way, and the absence reported here reflects what was present in the retrieved source, not necessarily what a fully executed browser session would show. This study reflects a single snapshot collected 25 September 2026; structured-data implementation can change at any time, including in direct response to the documentation update this study examines. This study did not assess whether the absence of VideoObject markup has any measurable effect on search visibility, video rich-result eligibility, or AI-system citation of the pages examined, and makes no such claim.

Conclusion

Among 8 university web pages confirmed, on 25 September 2026, to host an individually identifiable video, none carried VideoObject structured data, and none carried JSON-LD structured data of any type on the page examined. This finding holds across institutions in six world regions and across three different video-delivery mechanisms third-party embeds, direct file downloads, and an institution-operated video channel suggesting the absence observed is not confined to any single delivery method in this sample. Google Search Central’s own VideoObject documentation, updated one day before this study’s collection date, continues to specify three required properties unchanged since before this update, while newly clarifying a creator property alongside author and a four-type interactionStatistic model. On the evidence of this study’s small sample, that vocabulary remains, as of the collection date, unused on the specific university pages examined.

Sources

• Google, “Video (VideoObject) Structured Data,” Google Search Central, last updated 2026-09-24 UTC developers.google.com/search/docs/appearance/structured-data/video
• Google, “Search Central documentation updates,” Google for Developers developers.google.com/search/updates
• Schema.org, “VideoObject” schema.org/VideoObject
• Webometrics’ companion study on AI crawler access at universities (18 September 2026), source of several institutions in this study’s sample
• Webometrics’ companion study on university website digital-presence signals (23 September 2026)
• Direct inspection of the HTML source of 13 university web pages, collected 25 September 2026.

Suggested Citation

Webometrics by UNIRANKS (2026). Do University Lecture Videos Speak Google’s Language? A Webometrics Study of VideoObject Structured Data Adoption. Webometrics.org, 25 September 2026.

Frequently Asked Questions

What is VideoObject structured data?

VideoObject is a Schema.org vocabulary type that Google Search Central documents for describing an individual video on a web page in machine-readable form, including its title, thumbnail image, and upload date, so that search and AI systems can identify and describe the video accurately.

What changed in Google’s VideoObject documentation on 24 September 2026?

Google’s Search Central documentation-updates changelog records that the creator property was added, noting its interchangeability with the existing author property, and that interactionStatistic was updated to document its supported interaction types: WatchAction, LikeAction, CommentAction, and ShareAction. The three properties Google documents as required for VideoObject name, thumbnailUrl, and uploadDate were not changed by this update.

Does the absence of VideoObject markup mean a university’s video will not appear in Google Search?

Not necessarily. This study examined only the presence of a specific structured-data type; it did not test search rankings, indexing, or visibility outcomes, and Google’s own documentation does not state that VideoObject markup is required for a video to be indexed or to appear in ordinary search results. VideoObject is documented as relevant to video-specific search features, such as video rich results.

Were all universities in this study found to lack VideoObject markup?

Among the pages this study could confirm host an individual video, yes 0 of 8 examined pages contained VideoObject structured data. This is a small, purposive sample of one page per institution at 13 attempted institutions, not a comprehensive audit of any university’s full website, and should be read accordingly.