This will be my last post on SEO Skeptic. Don't panic – and to get to the action item right away for those that don't care why I'm closing up shop – from today forward you can find me at The Graph Lounge. If you want to know more about my new digital home there's […]
On 12 June Google introduced a new section on its Developers site, Media Actions. These actions "enable users to initiate media content (e.g. songs, albums, movies) on content provider applications via Google Search and the Google Assistant." This Media Action page, which replaces both the Music and TV and Movies Google Search feature guide pages, […]
Enumerations are a feature of schema.org that allow publishers to select a value for a property from a restricted list of values, such as the answer to a question where the only possible responses are "yes", "no" and "maybe". Enumerations are useful both to publishers and data consumers (like search engines) because they provide unambiguous […]
I encountered a tweet today from Dan Sure which provided the text of an intriguing notification from Google concerning the use of podcasts, recipes and news by Google Assistant, Google's virtual assistant that's used by Google Home, Android phones and a host of other devices. (And since then numerous other webmasters have also commented on […]
Google announced today that they have released a new-and-improved version of its Structured Data Testing Tool, available in beta as the Rich Results Test. The name of the tool reflects – as noted in the announcement post – the fact that what Google had previously referred to as "rich snippets, rich cards, or enriched results" will now be called "rich results" moving forward.
(For the record, even though the Google post title starts with "Introducing Rich Results" this is nomenclature that they've been using for over a year.)
Here's my preliminary look at this new tool, especially in comparison to the Structured Data Testing Tool – which continues to be available while the Rich Results Test is in beta.
An important point to note, particularly if you're reading this post right after its publication, is that the tool is in beta, and that it doesn't yet support the full range of data types supported by the Structured Data Testing Tool. These limitations are outlined in the Search Console help document on the "Rich Results Test".
Supported rich result types
This test currently supports only the following rich result types:
- Job posting
Unreported type errors
- This tool does not currently show schema.org syntax errors
- Various other errors (hey, it's beta)
So, for the time being, continue to consider the Structured Data Testing Tool results as the definitive structured test results for Google, especially for the data (rich result) types not yet supported by the new tool.
Eligibility for rich results
The Structured Data Testing Tool never explicitly states whether a given page (or code block) is eligible for rich snippets. The closest the tool comes is to provide a "PREVIEW" link when results are capable of generating a rich result, and only then after you click through to the eligible data type.
Eligibility is now front-and-center in the Rich Results Test response, with pages that are eligible for rich results clearly labelled as such, along which data type ("Detected structured data") for which the page is eligible.
It's equally obvious when a submitted page isn't eligible for a rich result.
There is also a third category pertaining to rich result eligibility, seemingly triggered when a page could be eligible for rich results if errors were corrected.
Details on these errors are provided in a pane on the "Detected structured data" section of the test results.
Again, be cognizant that only job postings, recipes, courses and movies are currently supported by the tool, as per the prominent message at the top of a test results page.
This is a beta version. Not all rich results and error types are supported yet.
Rich results preview
As noted above, the old tool only displays a preview link under the relevant data type, whereas the new tool makes it apparent from the main test results page that a preview is available.
Where there is a single preview available there doesn't seem to be any material difference between what the new and old tools display. However, whereas the old tool provides a desktop view when the test is conducted from a desktop device…
… the new tool renders a mobile preview by default, even on desktop.
There is a "Preview as:" link next to the preview that seemingly gives you the ability to toggle between mobile and desktop results, but for all the tests I conducted (and I tested all supported types) the only option was "Mobile view".
There is, however, one definite difference between the tools when it comes to preview functionality. The old tool only ever provides one preview link, whereas the new tool provides access to multiple previews when the page is eligible for more than one type of rich result.
These are listed as "Result type 1", "Result type 2", and so on in the upper left hand side of the preview pane.
For the recipe page I tested two results were available. A preview with the recipe rich result in a carousel:
And a preview showing a standalone recipe rich result:
Snapshot sharing and page submission
A welcome piece of new functionality is the ability to share test results. In the Structured Data Testing Tool results can only be shared via screenshot or some other manual mechanism. The new Rich Results Test tool, however, provides a simple sharing button by which a sharing link is displayed and made available for copying.
The Rich Results Test also allows webmasters to submit a URL for indexing directly from the tool's test results.
Code validation and error messages
A warning here that my observations here are preliminary, and that I haven't spent a lot of time looking at how the new tool handles code validation or at the messages the tool generates. But with that proviso in mind, here's what I've seen so far.
The two tools provide similar messages when it comes to recommended (i.e. optional) and required properties.
A notable difference, however, can be seen by how the two tools handle an invalid or unrecognized property (I haven't dug into type errors).
When it encounters an invalid property, the Google Structured Data Testing Tool complains. Here's what it has to say about the property
subjectOfStudy, which is not in schema.org.
However a Rich Results Test doesn't provide any messages about this non-existent property at all.
Whether this is an improvement or a drawback depends on how you look at it. It's an improvement if you don't want vocabulary validated that doesn't materially impact the ability of a page to be used for a rich result. And it appears, as the name "Rich Results Test" suggests, that Google has shifted the focus with this tool on the nuts and bolts of Google rich results, rather than more generalized vocabulary or syntax information.
On the other hand, this is useful information for debugging, and for future-proofing code. But this may yet come. The support document says that there are "unreported type errors", noting that the "tool does not currently show schema.org syntax errors" and "various other errors (hey, it's beta)".
Whither editable code?
The Google Structured Data Testing Tool allows you to either provide a URL for testing or to paste code directly into the tool for validation – and then to edit that code directly in the tool, regardless of whether it came from a URL or pasted code, and then re-test the edited code.
The Rich Results Test is, right now anyway, all URLs all the time.
Run the test
Submit the full URL of the page to test. The page must be publicly accessible by an anonymous user in order to test it (it cannot require auth, or be robotted against Googlebot). If your page is behind a firewall or hosted on your local machine, you can still test it by exposing a tunnel.
That's quite a drag, and that's certainly the feature from the Structured Data Testing Tool that I'd most like to see carried over to the Rich Results Tool. The ability to edit and test code on the fly is an enormous time saver.
Other features and summary
The announcement blog post says that the "new tool provides a more accurate reflection of the page's appearance on Search and includes improved handling for Structured Data found on dynamically loaded content".
To this first point, that seems to be true insofar as there are multiple previews available when a page is eligible for them, and that a mobile preview is shown by default.
To the second point I've not yet attempted to validate any dynamically loaded content with the new tool (but for that matter I didn't encounter any particular problems with dynamic content using the Structured Data Testing Tool).
All in all the new tool shows promise. It certainly looks sharper than the old tool, the ability to share test results is encouraging, and I think that we can expect better (and clearer) error and warning messages as the tool matures.
But please bring that code editing function back.
Update, 31 July 2018 On 30 July 2018 (in a post dated 2 Aug. 2018) Bing announced JSON-LD support in Bing Webmaster Tools. Bing had previously announced it was supporting JSON-LD at SMX Advanced in June, but this is the first published confirmation that Bing supports JSON-LD. As of 31 July 2018 the Bing structured […]
Dan Brickley of Google announced a major new release of schema.org today, bringing with it some fairly groundbreaking updates to the hotel and accommodation vocabulary. Dubbed "makemake," schema.org version 3.1 integrates the work of Martin Hepp and Hepp Research GmbH with the existing hospitality schemas, constituting the most significant update for the hospitality industry since […]
Yesterday Google updated its structured data documentation on reviews to include guidelines and code specifications for the creation of critic reviews for local businesses, along with a number of other, smaller changes. Exactly one year earlier Google announced support for critic reviews in the Knowledge Graph (at launch restricted to reviews of movies), and provided […]
I'm just back from Seattle, where I had the pleasure of speaking at yet another sold out SMX Advanced.
The topic of my talk was "What's New with Markup and Structured Data", where I reviewed recent search engine-facing changes relevant to structured data, structured data markup, knowledge bases and other aspects of semantic search. When I say "recent" I mean the past year or so, with an emphasis on developments in the last few months (and in particular Google's May 2016 update, which I covered in some detail in the post preceding this one).
The presentation deck embedded at the bottom of this post is (hopefully) self-explanatory, so I won't go into detail about what you can read there, but here are the major topic areas I covered.
- Just what is structured data?
- Why schema.org plays such a large role in structured data markup for search engines
- Structured data used by distributed content systems (Facebook Instant Articles, Apple News, Accelerated Mobile Pages (AMP))
- The introduction of Google rich results and rich cards
- Rich card previews and other changes to the Google Structured Data Testing Tool
- The rise of JSON-LD as Google's preferred syntax for schema.org
- Structured data for ecommerce, including the GS1 Web Vocabulary
- schema.org extensions
- The changing knowledge base landscape (Freebase, Wikidata, Wikipedia, the Google Knowledge Graph, and the Knowledge Graph Search API)
Aside from this, I'll note that in preparation for this presentation I created a publically-accessible Google Sheet that, for each of the items I discussed, provides the feature announcement and effective dates, the announcement URL, reference URLs and discussion URLs.
- Key Structured Data Events
This is a table I intend to keep updated, so please let me know if you spot any omissions or errors.
On 17 May 2016 Google rolled out a number of changes to how the search engine handles structured data on the web. Among the most important of these changes are the introduction of a new rich results type, Rich Cards; the complete restructuring of its documentation about structured data; and the addition of a new report in Google Search Console, the Rich Cards report.
This follows closely on the heels of changes made to Google's Structured Data Testing Tool just a few days earlier.
Let's take a closer look at some of the more significant of these changes, and what their implications might be for a site owner's search visibility and search engine optimization efforts.
Rich results and user intent
In a new page on search features, Google introduced the term "rich results", a catch-all phrase for instances where attributes about some thing represented on a website "can be featured in a visually compelling way."
These "rich results" are facilitated by the addition of structured data markup to web pages, either in the form of content items ("such as articles, recipes, or movies") or lists of items ("such as recipes and events).
This is the first time Google has made the differentiation between "content items" and "lists of items", and it relates to another concept introduced by Google in their revamped documentation, that of a user's search intent driving presentation.
When you prepare your content for inclusion in Search, it can appear in a variety of visual forms, because it's the user's search intent that determines how Google Search displays information. A user's search intent is a concept describing how our algorithms understand what the user really wants with a specific query within a specific context.
For example, if a user queries for "turkey," the search intent is ambiguous, so our results will display a mix of items, including a Knowledge Graph card about the country of Turkey, news items on Turkey, and turkey as a kind of fowl.
However, if the query changes to "turkey recipes," the search intent narrows to a list query, so a carousel of recipes helps to cluster related recipes together, including a summary carousel for recipes from various providers and a host carousel for recipes from a specific host.
"Cluster" is the key concept here. Google has now explicitly stated that it will make efforts to group results together when it ascertains that the search intent is for a "list query." And this is a good segue to the concept of a "card", and the environment that gave birth to it – the mobile web.
Cards, rich cards and mobile
A card, Google now tells us, "is the fundamental presentation unit for Search results. It appears in the organic results for Google Search either as a single element or a list of elements."
Note here we're talking about cards, not rich cards. Like everything else in the new search guides, the nomenclature here draws heavily from the language of mobile. These, for example, are cards returned for the query "red wine."
In other words – as far as I can judge – "cards" are search result items in mobile, which we used to call a search snippet. On desktop, indeed, one might still call these "search snippets", but its revised documentation and examples for structured data Google has essentially banished the concept of desktop results.
Or will we still be talking of "search snippets" in a week's or a month's time?
There have very recently been multiple sightings of mobile-style cards appearing in the desktop results, such as those you can see in this screenshot provided by Jennifer Slegg in an article on the SEM Post.
If indeed a card is "the fundamental presentation unit for Search results" then this may be less of a test than a preliminary rollout. And while I may have been premature in tweeting "Rich snippets RIP" (they are still mentioned tangentially in the documentation, and the rich cards announcement FAQ gleefully provides the answer "Yes, you can!" to the question "Can I keep my existing rich snippets markup?"), rich snippets themselves may soon be a thing of the past – if only in name, as conceptually this type of rich result lives on in the form of rich cards.
Which brings us to Google's main announcement about these changes, the introduction of rich cards – "a new Search result building on the success of rich snippets." And, in essence, rich cards (as evidenced by Google's "evolution" graphic in their post) are rich snippets re-imagined for the mobile web.
Google's has now started showing two types of rich cards: movies and recipes. Here's a current recipe rich card.
It's actually more accurate to say "here are current recipe rich cards" as what's pictured are multiple cards shown in a carousel. And in some ways, thinking of the "list query intent" discussed above, one might be tempted to say that a rich card is "a rich result that can be displayed in a carousel." But rich cards can also be displayed as single items.
When it comes to movies, however, color me confused. At present every single movie I search for returns ye olde Knowledge Panel (now a "Knowledge Graph card"), and no result bears any resemblance to the movie rich card previews provided by the Structured Data Testing Tool, as we'll see later.
As per the use cases listed the TV and movies data type page, one of the benefits of adding structured data to a movie item is to facilitate its inclusion in a Knowledge Graph card's streaming options, but this isn't a "movie rich card." And in both the announcement post and in the structured data documentation there's no movie rich card to be found.
However, movie rich cards will, apparently, one day appear in a host-specific list carousel (currently in "experiment mode" where you can test your markup for errors, but where the cards themselves aren't yet available in the search results). Which brings us to the next new feature Google announced, host-specific lists.
A host-specific list, as the lists specification page explains in reference to recipes, displays "cards from a single site within a specific category." Here, for example, is host-specific recipe list for curry recipes on the Food Network's website.
As you can see, these results are returned a carousel. Host-specific lists for events are provided in a more traditional "list" format.
As already noted, movie host-specific lists aren't yet available but are coming sometime soon.
And I expect we'll see more types of host-specific lists in the future. Mobile has really opened up opportunities here for Google to innovate with the presentation of search results, as the availability of carousels and lists within rich cards mean they're no longer faced with the conundrum of sacrificing host diversity by presenting a SERP with results all from a single site: those single-site results can be relegated to a carousel, with more diverse listings below.
Rich card and rich snippet reporting Search Console
Announced in concert with rich cards is the addition of new rich cards report within Google Search Console. The new report (which complements rather than supersedes the structured data report) displays errors that render a page ineligible for a rich card ("invalid cards"), cards that are "fully enhanced", and "enhanceable cards" – cards that "can be enhanced by marking up additional fields."
Google also promises that a new "rich results" filter in the search analytics report "will help you track how your rich cards and rich snippets are doing in search: you'll be able to drill down and see clicks and impressions for both." This feature is currently being tested in a closed beta.
As per a post from Google Webmasters on Google+, AMP reporting has been added to the search analytics report on Search Console, but this doesn't pertain to structured data usage specifically. More important in that regard is the 4 May 2016 update announced by John Mueller, in which AMP error information is now aggregated – and this error reporting includes structured data usage.
JSON-LD, JSON-LD and more JSON-LD
As noted in these pages, Google has slowly but surely making improvements for their support for JSON-LD. With these latest changes they're all-in when it comes to favoring JSON-LD for the markup that uses the schema.org vocabulary.
The announcement post states that "[w]e strongly recommend using JSON-LD in your implementation", and the new introduction to structured data page reiterates this recommendation, noting (correctly) that "[s]tructured data markup is most easily represented in JSON-LD format." A table on that page shows that JSON-LD is now supported by Google for "[a]ll public data types except for Breadcrumbs."
However, it's worth noting that, despite this recommendation, microdata and RDFa are still supported for all data types, and there's no suggestion that support for these syntaxes will be deprecated.
A revamped Structured Data Testing Tool and new previews
As noted above, the Google Structured Data Testing Tool was revamped just prior to the rich cards announcement and redesign of the structured data document changes. I wrote about those changes in some detail, so won't repeat the information I already provided in that post.
One thing I didn't notice, however, is that there is limited rich card preview functionality in the revamped Tool. The announcement notes that you can now see a preview "of how the rich card might appear in Search (currently available for recipes and movies)." Rich snippet previews were available in the first version of the Rich Snippets Testing Tool, but haven't been provided since the tool was redesigned in January 2015.
And here's what a recipe rich card preview looks like.
Okay, almost like that: I must confess that I overlaid the image on that screenshot. The actual preview, even though I've declared the image you (which is available at a Googlebot-accessible UR) looks like this.
This is because, as the new help article on the Google Search Preview Tool explains, "you can see your actual image in the preview if we confirm, or verify, that the image URL is in our index", and the asset I've used here hasn't yet been indexed by Google.
It's possible that this restriction is in place because the preview actually runs on google.com (and the help page also includes instructions on how to override regional Google domains using /ncr in order to facilitate the preview), and so might be restricted to returning only indexed images. The preview URL looks like this:
Based on that, and the help page's note that "[i]t helps you preview your structured data markup for web or mobile app URL as it would appear in the Google Search results page" (emphasis mine) It's likely that we'll see the functionality of this tool expand in the future.
For the record here's a preview of a movie item (with, as with my recipe preview, a manually added image). Pretty underwhelming "rich result" in my opinion, but perhaps isn't as lackluster as it appears when you think of these rich cards in the context of a host-specific list of movies. (Updated since publication: see the image of a host-specific movie list that's been added to the host-specific list section above.)
New structured data documentation and examples
Google has completely reorganized how they provide information about structured data use, specifications for specific types, and examples of structured data markup.
Previously, all information was provided from a single hub page on Google Developers. Now all things related to structured data appear on the Google Search site. The most notable change is that informational pages are now separated from pages about specific item types. Here are the most notable pages in the reorganized presentation.
Guides and usage specifications for structured data, indexing and page creation. Includes the new search features page, which gives an overview of rich results type, and the new search gallery, which provides images and code examples for a handful of rich results types (four at time of writing: recipes, events, products and reviews).
- Data types
The hub for individual data types that are eligible for rich results, such as reviews, events and recipes. These were previously served from the single documentation hub on Google Developers, under the heading "Enable Rich Snippets." The new way of organizing these items is much more extensible.
- Pilot features
A new holding area for data types that are being piloted "with a small set of initial providers." Currently the features listed there are live blogs and software apps.
- The Google Search Preview Tool
Information on the Search Preview Tool, as mentioned above. Worth calling out in this list because, as far as I can tell, the link to this page is exposed only after you click the "preview" button in the Structured Data Testing Tool for a type that's eligible for preview.
For articles it's worth noting that the prior pages for "Top Stories with AMP" and article rich snippets have been combined into a single articles page, reflecting the move to organize structured data documentation around data types rather than rich snippet types.
The way in which example code is provided has changed too. Previously code was provided on each specifications page inline. As per the image above, code examples are now linked by a "see markup" button, which opens the code example directly in the Structured Data Testing Tool (though there is still some inline code provided for the local businesses data type).
This doubtlessly makes example management easier for Google, and certainly now relieves users of the burden of copying and pasting code into the Structured Data Testing Tool. And, unlike before, the examples provided mostly come back as error-free in the Structured Data Testing Tool. Current exceptions are an aggregateRating example that, bizarrely, declares something named "Super Book" as a Thing rather than some more specific class (declared as such in examples for all three supported syntaxes), a VideoObject example where the thumbnail URL provided is not valid, and a MusicGroup example where the content declared by and event property is a URL, rather than the Event type that is expected.
One vocabulary to rule them all
Finally, while this isn't new, it's worth noting that all of Google's structured data guidelines and documentation reference schema.org, and only schema.org. Google has gone all-in with the vocabulary as the basis for structured data markup they support, just as schema.org's fifth anniversary approaches.
Noise is the data you don’t want to analyze. On the Web — specifically in the link graph — noise is the data no one pays attention to. It is the trivial stuff ignored machine learning algorithms ignore. An example of trivial link noise would be two Websites, Mike and Charlie, who each link to...
If you ask me what my favorite part of optimizing for search is in 2020 (or before), you wonâ€™t find link building anywhere near the top of the list. I say that as someone who has built thousands of links and who’s written multiple how to do linkbuilding outreach blog posts. I created my first...
Vectors have become the new SEO magic. I see people writing about and speaking about vectors everywhere and few of them really have any idea of what they are talking about. Vectors are not “math”. They are used in math but they are used in many other disciplines, too. We use math to precisely describe...
Itâ€™s not often I get invited to write a guest post by a mogul in the SEO industry that I happen to respect. So, thank you, Mr. Martinez. I appreciate the hot mic. Letâ€™s jump into it. Remembering Eric Ward (AKA LinkMoses) The fact Michael readily agreed to let me pay homage to the late...
Web marketers are locked in a death embrace spiral with 3rd party analytics tools like Ahrefs, Majestic, Moz, and others. Every year I see more people sharing screen reports of traffic estimates from these tools when asking for helping or bragging about how great their SEO strategies are. These 3rd-party traffic reports are unreliable, highly...
Finding a good marketing forum to help you with the challenges you face isn’t easy. Some forums won’t allow you to post a link to your Website, which makes it virtually impossible to get good advice from the people visiting the forum. Some forums go even further into making life difficult for you by preventing...
Michael Martinez asked me to contribute some SEO tips to this blog, and I haven’t written about SEO in a long time. I (reluctantly) said okay. But I wasn’t sure where to start. I finally thought maybe I should start with the same activity I start my SEO activities with: Keyword research. Here’s how to...
SEO Theory is now running on two sites. One is a subdomain. This division of content was necessary for several practical reasons. My decision to archive older content on a subdomain wasn’t based on anything to do with SEO, but there will be some practical SEO benefits from doing so. I don’t want to dwell...
You are here: Index Webmaster Tips and Tricks