Great overview of the topic of semantic search from . A great read, especially, for search marketers just dipping their toes into semantic waters.
GS1, the organization that maintains the global standards for barcodes, has introduced a new generation of package-based product code, the DWCode.
Produced in partnership with digital watermarking company Digimarc, the new code has these attributes and capabilities:
Invisible to ordinary consumers, it can printed across all surfaces of a product package.
Significantly faster to scan at checkouts than barcodes.
Uses GTIN (Global Trade Item Number) to reliably link to product information for that item.
Is scannable by mobile devices.
Provides rich product information to consumers who have scanned the package.
Resolves to a dynamic URL that can be used to provide consumers with product information, coupons, promotions and more.
Pretty interesting initiative that leverages the power of linked data to enhance information about real-life things.
And it's a value proposition in its own right for the power of the GTIN. So for context this just-published article by of Google is an excellent companion piece:
Five reasons why GTIN should be your new favorite Google Shopping acronym
Many thanks to for the heads-up on this!
#barcodes #gtin #linkeddata #gs1
An interesting look by at a recently-granted Google patent, "Semantic selection and purpose facilitation".
The patent describes methods by which a user can readily identify an entity, and then take an action related to that entity. For example - as per the image in the call-out link - a user might identify the restaurant "Ramem Sushi", and then be able to make a reservation at that restaurant.
Bill notes that this "seems to be a fairly aspirational patent, which might require a lot of steps being put into place before it is implemented", but that - given the potential of what such a mechanism could support - it may well be something that the search giant pursues.
#semanticsearch #entities #google
Just one of those Google News mysteries (especially as the cited text doesn't appear in the code).
Good case study from on how he was able to manipulate the Knowledge Graph by modifying an entity's Wikidata statement. And what's interesting here is that he employed only Wikidata, AFAIK leaving the entity's Wikipedia page as-is.
Tony, if you're reading this, can you tell us how long it took before your change was reflected in the Knowledge Panel?
#wikidata #knowledgegraph #google
However, I do take issue with your claim that "Google Authorship was also built on the foundation of schema.org."
The authorship program, of course, was predicated on the use of a rel="author" declaration (later a ?rel=author declaration appended to an href declaration) This is not schema.org (the relevant property for a CreativeWork there is "author"), nor does it bear any relationship to it.
rel="author" as implemented by Google is essentially a flavor of plain old semantic HTML (POSH), where usage is determined by the conventions imposed by the data consumer (such as, in this case, appending a the ?rel="author" parameter to a linked Google+ profile URL), rather than machine readable standards.
schema.org, by contrast, is true linked data. While data consumers like Google can and do impose their own requirements on how schema.org is used, the classes and properties of the vocabulary are defined, along with the expected types for each.
And schema.org launched in June 2011, was in its infancy when authorship in August of that year.
Somewhat ironically, if the vocabulary had been more mature at the time Google began thinking of the implementing authorship (which surely must have been in the works some time prior to the introduction of the vocabulary) they might well have opted to avail themselves of schema.org, since it would have made information declared about the authorship of articles (or any creative work) reliably available to all data consumers. In the current technical environment nothing would be more straightforward then declaring, using JSON-LD, a Google+ profile URL using the sameAs property applied to Person nested under author.
In summary it would have been great if Google Authorship was built on the foundation of schema.org ... but it was not.
Instead of forcing readers to follow a single, linear, flat sequence of sections, just offer different routes between articles or items. Just like there are different ski slopes for beginners and pros, you could even mark the connections with different colors, from blue over red to black.
By offering more than one outgoing connection from one doc item to the next, you allow users to find their individual path through the documentation, following their natural intuition, saving energy for creative tasks.
As announced by (http://bit.ly/1XSuFvy), the Web Data Commons has just released a new RDFa, Microdata, embedded JSON-LD and Microformat data corpus, extracted from the November 2015 version of the Common Crawl (1.8B pages, 14.4M websites):
This is the first Commons corpus to report on embedded JSON-LD discovered in the crawl, so I took a quick look what classes and properties were being most commonly declared.
As per the call-out graphic, there's a pretty direct relationship between what webmasters have been encoding and JSON-LD supported search results features in Google.
Clearly many sites - in the order of a half-million+ - have opted to provide Google with information about their internal search in order to generate an in-SERP sitelinks search box, first announced by Google in September of 2015 (http://bit.ly/1XSvXqt).
Markup in support of social media profile links, announced in January 2015 (http://bit.ly/1CrX6pr), is also much in evidence, and probably accounts for good chunk of the sameAs declarations observed.
Finally, JSON-LD support for logos and corporate contact information, both announced in 2014, is probably responsible for some other top JSON-LD-encoded classes and properties in the extraction, including Organization and sub-classes, ContactPoint and logo.
Interesting to compare this list against the equivalent top lists for microdata, which is quite different and skews much more heavily toward types with long-standing rich snippet support, such as CreativeWork classes, Product and Offer. It will be interesting to see how this distribution over time, with JSON-LD support for these types all having been introduced since this Crawl.
Finally - because we need mystery, right? - how it it that theclothdiaperwhisperer.com has more triples (570B) than all of blogspot.com (538B)? :)
#commoncrawl #webdatacommons #jsonld #schemaorg #structureddata
Past that the Web Data Commons extractions can, depending on the analytical use case, be a little troublesome to work with, as the Common Crawl corpus is neither comprehensive nor consistent (i.e. AFAIK the domains surveyed are not necessarily the same, crawl over crawl). See this paper (PDF) by and - http://bit.ly/1NTqf1K
But for uncovering those trends WDC is an excellent resource.
And in news pertaining to the "semantically categorized" aspect of intelligent content, the International Press Telecommunications Council (IPTC) has just received a grant from the Google Digital News Initiative Innovation Fund to develop EXTRA, "a multilingual open-source platform for rules-based classification of news content."
The goal here is help "publishers to enhance their content with all sorts of metadata services, including enriched search, intelligent recommendations and precise analytics."
An interesting initiative, and one well-suited to be carried out by the IPTC, which has taken a lead role in standards and metadata development in the news industry.
Really interesting analysis of Google's patent application "Evaluating Semantic Interpretations of a Search Query" (http://1.usa.gov/1pSnyb8) by .
Of particular note are Davies' observations about the impact of click-through rate. He rightfully turns the tired trope about whether or not CTR influences rankings on its head and instead turns attention to how CTR can be used by Google to inform their interpretation of a query.
For example, if Google presented results for "hamilton new york history" that were about the village of Hamilton, New York, but then observed that a plurality of clicks were going to pages about Alexander Hamilton in connection with New York City, Google could judge - based on CTR - that its interpretation of the query was problematic, rather than that one or more of those results about the village of Hamilton needed to be re-ranked.
As Davies observes:
"This takes the question of clickthrough rate well beyond whether it impacts an individual site’s rankings and into the realm of the entire result set being valid or not and, if not, a completely different interpretation chosen. This means that if you notice a site going up or down in the rankings it may have little or nothing to do with the individual clichtrough rate or even relevancy to a subject but rather a completely different interpretation of the same query. This is huge."
It's about raising the discussion of the CTR to the next level and starting to think in a more abstract and thus beautiful way.
Google recently facilitated sharing of Knowledge Graph Panels, and some other features (http://bit.ly/1q6hM6o - thanks ).
This comes not too far on the heels of Google's release of the Knowledge Graph Search API in December of 2015 (http://bit.ly/1ZfYwPI).
In playing around with the URL now available from a Knowledge Panel, it became immediately apparent (and wasn't surprising) that the Knowledge Graph identifier there was the same one you can retrieve through a search using the Knowledge Graph API search.
For example, the ID returned from Knowledge Graph for everyone's favorite orange-haired politician is:
Which, transformed into an HTTP address using the "kg" prefix provided in the API search results, is:
Which resolves (via a 301) to:
Compare this to the Knowledge Panel share URL when one searches for "donald trump":
Resolves (via a 301) to:
The difference is, as per the clue offered by the parameter "kponly" in the fully-resolved URL from the Knowledge Graph Search API results, is that g.co/kg/* returns only the Knowledge Graph Panel, without any search context (the query term displayed is actually, for all Panels retrieved via the prefix URL, is "knowledge graph search api") - again:
However, deconstructing the URL to which a shortened Knowledge Panel share resolves exposes some parameters that can be used to expose a Knowledge Panel in a more meaningful content.
By modifying the "hl" parameter you can, of course, change the language of the content displayed - including the content of the Knowledge Panel:
It turns out this also works when using the "kg" URL provided in Knowledge Graph API search results:
But by hacking the structure of the share URL you can provide users with search results and an accompanying Knowledge Panel where the query is only related to the entity in the search result:
Or, as per the call-out image (which is a direct screenshot, not a Photoshop treatment), not related to the Knowledge Panel at all (in this case the query might be related - I'll the reader decide):
I don't know what the other parameters do (though "source" and "entrypoint" seem straightforward enough). "kgmid" and "q" in themselves seem sufficient to generate a search result accompanied by a Knowledge Panel:
None of this, at first blush, has earth-shattering practical implications, although I now know how to generate a Knowledge Panel in a language other than English, and - should it ever become useful - I know now how to send a user to a Knowledge Panel with the search query context of my choosing.
Oh, and worth noting the share URL now provides a method of retrieving a Knowledge Graph ID without using the search API: if you're able to generate a Knowledge Panel via search, you can now simply expose the ID by copying and pasting the share link.
#knowledgegraph #google #identifiers
- University of Alberta
I've long had an interest in classification that has extended in the computer age to meta data and the semantic web - but I definitely wouldn't consider myself a geek (I don't have good enough math skills for that).
For my sins, I also increasingly find myself working on information architecture, user experience and website analytics.
As if that's not enough, I'm an avid observer (and sometimes participant) of the discussion surrounding digital journalism, and the struggles of traditional news media organizations as they try to adopt to 21st century realities. (I'm also very keen on news optimization/SEO for Google News.)
I write on various topics related to Internet marketing on my own blog, as a now-occasional columnist at Search Engine Land and other places around the web. I'm also a prolific tweeter.
Interested in search and the semantic web? This Google+ Community (which I run) is definitely the place for you:
If you're on Twitter, I also recommend following this list (which I curate):
- Semantic Web
600+ semantic web people and organizations on Twitter
- Electronic ArtsSEO and DPO Analyst, 2014 - presentHeading up search engine and digital presence optimization efforts for more than a dozen EA domains, including www.ea.com and www.easports.com.
- AirshockDigital Marketing Consultant, 2013 - 2014Independent digital marketing consultant specializing in improving the visibility and performance of websites in the search engines through the application of semantic web technologies.
- InfoMine.comInternet Marketing Manager, 2011 - 2013
- AirshockSEO Consultant, 2010 - 2012
- Suite101.comDirector of SEO, 2009 - 2010
- Ice.comSEO Manager, 2008 - 2009
- Fivermedia.comSearch Engine Marketing Manager, 2008 - 2008
- Riptown.com MediaSenior SEO, 2006 - 2008
- Riptown.com MediaSEO Specialist, 2005 - 2006
- InfoMine.comSenior Web Designer, 2003 - 2005
- Self-EmployedIndependent Web Designer, 2000 - 2002
- Canadian Forces CollegeWeb Designer/Administrator, 1995 - 1999
- Canadian Forces CollegeLibrary Technician, 1988 - 1995
Why There Aren’t Right-Handed Apes, Or: Handedness and The Evolution of ...
Although most humans are right-handed, other animals don’t seem to show a similar motoric asymmetry. As Corballis mentions in his 2003 BBS a
SEMpdx SearchFest 2015 Mini-Interview: Aaron Bradley | SEMpdx
Aaron Bradley will be speaking on "Semantic Juggernaut" at SearchFest 2015 which will take place Friday, February 27, 2015 in Portland Orego
Should You Put Keywords in the URL? (An Age Old Question)
Since the age of dinosaurs, SEOs and their clients have been asking if they should put keywords in their URLs in order to improve visibility
Why most marketing campaigns fail and how yours can succeed
Marketing campaigns fail mainly because they are executed without any well defined strategies.
Why are you so obsessed with this Semantic Web thing
A lot of nice buzz today in sociala media when Tim Berners-Lee discusses the future of the web in the March issue of Wired UK. The web turns
Powered by EA Customer Service Free Games Free Games Top 5 Free ...
Dawngate is a new MOBA, built from the ground up to look and feel familiar while offering a whole new way to experience MOBA gameplay. Dawng
Powered by EA Customer Service Free Games Free Games Top 5 Free ...
A Galaxy of Possibilities Awaits You! Star Wars™: The Old Republic™ is the only massively-multiplayer online game with a Free-to-Play option
Greenpeace-commissioned report slams plans to develop world's largest co...
"The project's economics don't stack up," authors from the Institute for Energy Economics and Financial Analysis wrote.