Hi everyone  - I just posted a presentation on how the website of Apache Marmotta (an open source Linked Data Platform implementation) has been scraped and "featured" as Answer box 2.0 by Google. 

The presentation also  highlights how we're starting using +WordLift 3.0 (a plug-in for WordPress currently being developed) to publish your entities as linked open data and adding owl:sameAs to link the equivalent Freebase entities. - presentation is actually dedicated to this group :) 
Add a comment...
Is there a smart way to extract Organization and/or LocalBusiness types, as a list, from (or somewhere) which will remain current?

My knowledge of SPARQL is limited (but improving) and I'm assuming that using it would be the best way to extract the list of business types as well as other types i.e Events, rather than keep such a list manually updated.

Hopefully someone here can point me in the right direction, or provide a universal solution.

The use case/scenario I have in mind is that 'the list' could then be used in plugins/forms etc to help webmasters make good current choices.

thx in advance for any ideas and feedback.
Kevin Polley's profile photoAaron Bradley's profile photo
The simple approach works for me :) Thanks so much +Dan Scott, add 'brilliant' to 'outstanding'.
Add a comment...
"Pages with integrations rank better by an average of four positions compared to pages without integrations"

This being probably the most notable finding from the final version of the  +Searchmetrics "study of tens of thousands of keywords and more than half a million domains" headed " in Google search results" (see the request button at the bottom of the page for the full PDF):

The study authors are quick to note that correlation does not necessarily imply causation, although they correctly acknowledge that there's a logical connection between structured data use and search engine visibility.

"Whether the use of HTML markups does genuinely affect the ranking of a domain, as it would appear from the data, is hard to say. In fact, it may be due in part to interaction with other factors."

"However, it can be said that the inclusion of HTML tags from to clearly flag up the content of the page will greatly help search engines better understand the Web’s content – and that is an essential consideration for a good ranking."

Other notable factoids from the study:

- Only 0.3% of domains in their sample included [1]
- "Google enhances search results with markups in more than 36% of keyword queries" [2]
- Movie, Offer, TVSeries and Review were the item types that most often ended up generating rich snippets in the SERPs (compare this to the most common classes by domain from the lastest Common Crawl data - WebPage, Article, Blog, Product - and most common classes by no. of entities - Product, Offer, Person, PostalAddress, Organization)

[1] Making for the best pie chart evah
[2] I'm interpreting this as "for 36% of our sample queries we observed a rich snippet in the top 50 results [their query depth] that appeared to be based on the presence of markup."  The language of this study leaves a lot to be desired.

UPDATE Some further information on the study can be found in a post from Searchmetric's +Marcus Tober:
(Marcus, if you're reading this note that the links to your LinkedIn and G+ profiles on are malformed.)

#schemaorg   #structureddata   #google   #seo  
Alex H Yong's profile photoHenry Zeitler's profile photo
Add a comment...

Hasan Deniz

Introductions  - 
List of sites using Microdata

A few example sites for to study Microdata implementation 

#schemamarkup   #microdata   #listing  
Ethan Hays's profile photoHasan Deniz's profile photo
Following up the link you have provided, I have updated the page with more accurate numbers.. :)
Add a comment...

Maxime Deland  - 
I'd like to markup a store that fits in both "ClothingStore" and  "ShoeStore".

If I divide the "Store" in 2 departments, I'll only have a description under these 2 departments. Is that acceptable SEO-wise?

Is it acceptable to equally have 2 itemtypes for a single location or I should leave it to the more generic "Store" itemtype ?
Jarno van Driel's profile photoBarry Rolapp's profile photo
+Jarno van Driel I also frequently forget we can use multiple types an props like that. Good call! 
Add a comment...
altough most of you probably know the differences and relationships between, Microdata, Microformats and so on, but i think this article is still worth sharing as it gives a very nice and simple overview
Vincent Velema's profile photoAlex H Yong's profile photoHenry Zeitler's profile photoHasan Deniz's profile photo
Thanks! I actually don't have knowledge in those things so I need this!
Add a comment...

Mark Davey  - 
More on actions and use cases from +Alexandre Passant
Aaron Bradley's profile photo
Great share and nice post +Alexandre Passant!
Add a comment...
Question for the community... allows implicit or missing information to be used with the meta tag with content. Using this technique prevents the information from being displayed but helps tell search engines about the additional information. "Lat longs" is a type of data used as an example for this scenario.

It might look like this:

<meta itemprop="latitude" content"40.806862"/><meta itemprop="longitude" content="-96.681679"/>

So, my question is, is it reasonable to expect a search engine will index the "content" used in this tag (i.e. the lat longs) and use it as a relevance signal for that page?

Follow-up question: Would this have any benefit in local search? (in the case of locational data?

David Deering's profile photoAaron Bradley's profile photo
Some great answers here - I'll just add some further color and a couple of links.

Most importantly, I think, you need to distinguish between truly hidden data, and data that is "hidden" from humans in one form, but exposed to them in another form.

So if you provide a location like "our store in Exampleville, CT" to humans, but then provide geocordinates in "hidden" <meta> tags to the search engines, then you're basically separating the presentation and data layer, without looking to deceive.

When data is not visible in any form to humans, that's generally when one is crossing the line.  As per the FAQ on this subject (

"As a general rule, you should mark up only the content that is visible to people who visit the web page and not content in hidden div's or other hidden page elements."

Sometimes there's situations where data that humans can understand needs to be formally declared for search engines, and this is the other class of "allowable" non-visible markup.  Other documentation ( provides a great explicit example of this, where an image of star ratings - readily understandable by humans - needs to be expressed as a structured value for search engines.
Add a comment...

Hasan Deniz

Structured data markup  - 
Semantic coded. Microdata is being added day by day.

Please do make comments as they are so valuable.

#semantic #microdata #schema
Hasan Deniz's profile photo
A quick question if you have time +Aaron Bradley 
How to use TravelAction for multiple stops?
fromLocation - toLocation , distance 18km
fromLocation - toLocation , distance 25km
fromLocation - toLocation , distance 12km

Or can we use 
fromLocation - toLocation , distance 18km
toLocation - toLocation , distance 25km
toLocation - toLocation , distance 12km

Doesnt seem right but could it be used for multiple stops on one trip ?
Add a comment...

Evgeniy Orlov

Introductions  - 
There is a funny domainname with PR4, registered in 2011, which will expire. Interesting for somebody?
Add a comment...

Kenichi Suzuki

Structured data markup  - 
I'm a newbie of Linked Open Data (LOD). Would you help me to learn it?  

Is an implementation described below regarded as using LOD?
- I have a profile page on my site.
- I include samaAs property of pointing to URL of my twitter account in it.

I know Twitter is not a LOD but couldn't it be a kind of LOD? I could connect my profile on my site with data that reside on Twitter.
Michael Atherton's profile photoKenichi Suzuki's profile photo
+Michael Atherton Thank you for your thoughts. I really appreciate them. I'll look into FOAF.
Add a comment...

About this community

A community for the discussion of topics at the intersection of search engine marketing and semantic web technologies This includes - but is not limited to - and means of marking up HTML with structured data, such as microdata and RDFa. While "search marketing" is part of this community's name, discussions about digital marketing and the semantic web outside of the context of SEO and PPC are also welcome and encouraged. New to this Community? Select "Introductions" as the post type and tell us something about yourself! #schema #schemaorg #rdfa #microdata #knowledgegraph
There!  I just became #667.  I felt that as the grandson of a minister it was my duty to throw off the beast.

You're welcome. :-))
Christopher Vogelmann's profile photoKevin Polley's profile photo
Thank you +Kevin Polley! Such a friendly bunch. 😎👍
Add a comment...

Aaron Bradley

Discussion  - 
Quick - find us some new members!

Until then we remain the "beastliest" Community on Google+.  If we remain 666 for very long I will start posting videos from this Iron Maiden tribute album.
Jarno van Driel's profile photoKevin Polley's profile photo
If you like DubFX you should also check out MC Xander ( +Kevin Polley.

And if we move away from rock to dubstep than I can't help throwing some hometown hardcore (Neophyte - Execute) into the mix as well. How's that for a U-turn?

Which isn't all that strange actually, since I also like Industrial (Ministry, Nine Inch Nails, Rammstein). Now both genres have their root in EBM (Kraftwerk, Front 242), which's roots ly with Pink Floyd - which I also like. 
Add a comment...

Barry Rolapp

Structured data markup  - 
I have a quick question I am hoping someone can help me with.

I am implementing JSON-LD on location pages for some of my clients - however contrary to how the examples at, my client has different hours for different days of the week - i.e. open Monday - Friday 9 am to 8pm, open Sunday 9am to 6pm.

When I input two instances of the openingHours itemprop and then test the JSON-LD in, it does not throw an error, however it won't display that it recognizes the second openingHours itemprop.

Would an implementation like "openingHours": "Mo-Fr 09:00-20:00, Sa 09:00-18:00", be an incorrect implementation JSON-LD.
Jarno van Driel's profile photoBarry Rolapp's profile photo
As long as you don't program chances are slim you'll ever be responsible for generating the JSON-LD output yourself +Barry Rolapp. But the way I see it, somebody who understands semantics will probable always be needed to doublecheck the outcome and give feedback where needed.

So my personal aim is to feel comfortable enough with it I can do such. And thus sooner or later I will have to bite the bullet and get comfortable with the JSON(-LD) terminology and specs enough to be able to get the message across to programmers, as for now they're 'kicking my proverbial ass' still. Hence my fiddling around with JSON-LD.
Add a comment...

Aaron Bradley
owner  - 
"If I had to guess..."

Wow, the already-awful "reporting" and "expert commentary" on just hit a new low.

On the plus side, I just deleted the better part of a rant which you, gentle reader, now don't have to read (or abandon as you discover it's a rant).

So two parting points:

1)  Aspiring young journalists, people turn to news sites when they hope to have something explained.  You should refrain from offering up your guesses when you don't know something, as your guess is as good as mine.

2) I beg +Barbara Starr to write some sense about Actions in her next +Search Engine Land column, and so do a little bit to redeem this publication's blighted good name. announced a new form of Schema they have introduced with support from Google, Microsoft Bing, Yahoo and Yandex named Actions. Schema Actions are a way to communicate via markup on your web page the actions they enable and how these actions can be invoked. Technically, describes it as: An action performed by a […]
Leeza Rodriguez's profile photoAaron Bradley's profile photo
+1. next time just go cook some of those delicious popovers  
Add a comment...

Jerimee Richir  - 
Hello. I have a question. I have a restaurant and on our menu we offer three things: pie, cake, and bacon. On my website I display the menu for those three things, but I want to add the microdata for each of the three offerings. Can someone point me to an example of that?

Same question articulated differently: What is the correct way to add schema microdata to this page?

Thanks so much in advance! I'm excited about learning more about incorporating schema into my sites!
Aaron Bradley's profile photoJerimee Richir's profile photo
If you notice any changes to your search visibility as a result of making changes - especially the appearance of any menu items in the SERPs - please let us know.

(To date the only sightings of menus have been based on linked 3rd party menus as far as I know -
Add a comment...
"Perhaps the most interesting addition to since launch"

As announced today by +Dan Brickley on ( 1.2 is live, and now includes the complete Action hierarchy - including the potentialAction hierarchy.

Say the primary authors in their blog post (call-out link):

"... the Web is not just about static descriptions of entities. It is about taking action on these entities - from making a reservation to watching a movie to commenting on a post."

"Today, we are excited to start the next chapter of and structured data on the Web by introducing vocabulary that enables websites to describe the actions they enable and how these actions can be invoked." 

Says Dan of potential actions:

"Today's publication builds on the types that we added last August by providing a way of describing the capability to perform actions in the future."

Two years in the making, the work on actions been very much collaborative, with Jason Douglas and +Sam Goto of Google, +Steve Macbeth and +Jason Johnson of Microsoft, +Alexander Shubin of Yandex and +Peter Mika of Yahoo leading the charge.!

#schemaorg   #actions  
When we launched almost 3 years ago, our main focus was on providing vocabularies for describing entities --- people, places, movies, restaurants, ...  But the Web is not just about static descriptions of entities...
Frank Gainsford's profile photoMariya Moeva's profile photoAaron Bradley's profile photoWissam Dandan's profile photo
+Alexander Shubin Oops, I knew that and meant to type "of Yandex" but somehow didn't (I guess in my excitement:).  Corrected now Alexander!
Add a comment...
"In some cases, it is desirable to serialize in the reverse direction"

So says the JSON-LD recommendation, but no such reverse property mechanism exists for microdata.

Accordingly, a proposal to create a reverse property - "itemprop-reverse" - has been created, and is available on the WebSchemas Wiki at the call-out link.

The simplest summary, I think, is this sentence from the proposal:

"The new attribute @itemprop-reverse will be equivalent to the existing @itemprop, except for the fact that the subject and the object of the statement are swapped."

Great to see this, and I'm impressed with the speed with which this this discussion morphed into a solid proposal.  +1's all around - +Jarno van Driel+Martin Hepp, +Thad Guidry+Dan Brickley.

Quick (two part) question for the principles (one day I avail myself of the "discussion" tab, I guess).  All the examples use <link> to declare itemprop-reverse, and all the values used in the examples are URIs.  Is this inherent to the proposed new attribute?

For example, instead of this:
<div itemprop="creator" itemscope itemtype="">
    <span itemprop="name">William Shakespeare</name>
    <link itemprop-reverse="creator" href="">

Could one say this (the strangeness of this construction notwithstanding)?

<div itemprop="creator" itemscope itemtype="">
    <span itemprop="name">William Shakespeare</name>, creator of <span itemprop-reverse="creator">Romeo and Juliet</span>

#microdata   #inverseproperties   #schemaorg  
Jarno van Driel's profile photoMartin Hepp's profile photo
+Jarno van Driel You are right: The good thing about @itemid is that you can use the same entity (@itemtype) with different properties, while with @itemref, you have to use the same property for linking to the entity.
Add a comment...
Question -- what is the difference between, and
when you have a website for a LASIK surgeon?
Kevin Polley's profile photoAaron Bradley's profile photo
A suggestion definitely worth a +1 +Aaron Bradley.  At the same time maybe you can also suggest the inclusion of as a type of Intangible.  This would be in keeping with recent changes away from pluralized types and help avoid misconceptions people may have about their mythical status :)
Add a comment...

Jarno van Driel
moderator  - 
Prepare the bomb shelters - inverse properties for are up for debate [censored]
Duck! bbgun pellets are coming your way - inverse properties for are up for debate

If the ongoing discussion about Socialnetworks of a person or organization at the public-vocab isn't spectacular enough for you, than you might want to get ringside tickets for the new thread +Martin Hepp just started at the public-vocab about inversed properties (

A mechanism to be able to indicate reversed properties is something I regard as a good thing because it adds a lot of value as to what we can express with For example: With Microdata we currently only can express that an Article has an 'author'. But there is no way to express a Person is the 'author of' an Article. Yet RDFa and JSON-LD do provide ways to accomplish this with

The problem is, if we try to resolve this issue from within, we would need to almost double the amount of properties since many properties would need to have an inverted version of it.

But if we want to prevent suffering from obesitas then what do we do? Do we use HTML properties like 'rel' and 'rev' or do we maybe even need to change the Microdata specifications to accommodate inverted relations? Questions this thread hopefully will answer, but which potentially could be quite the semantic showdown.
Jarno van Driel's profile photoGregg Kellogg's profile photo
I just posted a suggestion to the Web Schema's list that proposes that we interpret the JSON-LD context definition (forthcoming shortly, I'm sure) when processing Microdata. This would allow the context to define reverse terms, which could be used as the value of an @itemprop in Microdata markup. It also addresses the problem of keeping the Microdata registry up to date, by abandoning it in favor of a JSON-LD context.

+Martin Hepp I think this might simplify the @itemref issue too, as it doesn't really change the core processing model, just they way triples are emitted.
Add a comment...