Google published a Search Off Record episode about structured data. An interesting topic they touched on was how Google might use structured data that goes beyond what Google’s developer pages recommend. Martin Splitt said that more data can be better.
But there are caveats to adding more structured data beyond what Google recommends.
Schema.org and Google’s Structured Data Recommendations
The relevant part of the podcast began when Lizzi Sassman, a tech writer on the Search Relations group, observed that the structured data documentation at Schema.org was larger than the documentation that is found on Google’s developer help pages.
Google’s structured data recommendations only use a fraction of all the available structured data that is documented at Schema.org.
The structured data ecosystem is larger than just Google and there are many uses of structured data that go beyond what Google recommends and according to what Google shared in this podcast, Google can use some of that extra non-recommended structured data that’s off the books in terms of what Google recommends.
The founding companies behind Schema.org are:
- Microsoft
- Yahoo
- Yandex
However there are major contributors from Drupal and many others in the greater web community.
So Schema.org and the structured data vocabulary that they develop is far greater than the essentially small amount of structured data that Google recommends.
Ryan Levering, a Google software engineer concurred on how Schema.org contains a far more expansive available vocabulary than what Google uses.
Levering observed:
“So it’s just a way of expressing information.
…So it’s very expansive, and it has to account for all of the different use cases that could be used for it.”
Lizzi Sassman, the Search Relations tech writer noted how limitless Schema.org could be:
“And the application of those things could be limitless or like other people can then use this to then say, “We need this type of information to do this thing.” Or that’s the Google aspect, it’s that it’s adding the “This is what we can do with it if we know this information.”
Google’s Use of Structured Data
They next noted how the recommendations found on Google’s developer pages are limited in scope but that Google might actually use structured data that goes beyond their recommendations.
It’s important to note that the purpose for Google’s structured data recommendations is more or less to help publishers obtain rich results by making it easy for Google to identify which images to show in rich results and which data to use in that context.
What the Googler’s next talked about was about using structured data to help Google know what a page is about.
This use goes beyond what Google recommends for the purpose of showing rich results and it’s something that John Mueller has mentioned in the past.
But this additional use of structured data is not something that is officially documented.
So it’s kind of like, if you know about it then you can take advantage of it.
Martin Splitt , a Developer Advocate at Google asked:
“…But is structured data useful beyond what we use for, say, certain search features such as rich results?”
Levering responded that Google could potentially use non-recommended structured data.
Ryan Levering:
“We can potentially use that for some things at Google.
…I never advise people to not put structured data on their web page if it makes semantic sense.
…We also have some things that we do to generally understand the topic of the page. And sometimes the data you put on that can go into that.
Now that’s a very ML sort of process, where we look at all of the text on the page and we look at other things that have to do with the page.
So structured data is just one signal in that overall calculation.
But it can help us with certain disambiguations in terms of what the actual page is about. So it is useful but just in a more implicit sense right now.”
Non-Recommended Structured Data Can Be An Extra Signal
Lizzi Sassman said that the additional structured data could help Google better understand what a web page is about.
Ryan Levering observed that they could probably figure out what a web page was about without structured data.
But that in some cases it might be harder to know what a page is about and that’s where additional structured data might be helpful.
Levering observed:
“So it’s hard to convey that in some of our reporting and stuff that we actually find this useful. Because it’s a nuanced calculation.
But when there is problems detecting it, we can use it as an extra signal.
So it’s usually on the edge cases where we find that stuff useful.”
Martin Splitt commented on how good it was to add more data if it helps to clarify what a page is about.
Splitt commented:
“That’s really, really cool, and I think it’s generally easy to say like more data, as long as it is correct and reflects what’s shown to the user on the page, is never worse, right.
It’s always better to add more data to clarify what the content on the page is.”
Implementing “Undocumented” Structured Data
A caveat inherent in what Martin Splitt said is that everything that’s in structured data must also be in the visible portion of the web page. Google considers it spam when the content in the structured data does not match the content that’s in the visible part of the web page.
Citation
Listen to the Google podcast on YouTube at the 11:56 minute mark:
