WebProNews

Tag: Knowledge Graph

  • Google Now Offers A Knowledge Graph Search API

    Google Now Offers A Knowledge Graph Search API

    Google’s Freebase recently announced the Knowledge Graph Search API, which will enable developers to query the Knowledge Graph for entities.

    The API uses standard schema.org types and is compliant with JSON-LD specifications. Freebase says it’s working on a replacement for the Freebase Suggest Widget that leverages the Knowledge Graph Search API. They hope to release that early next year.

    “In December of last year, we announced that we would wind down the Freebase service and transfer the data to Wikidata,” Freebase says in a Google+ update (via Search Engine Land). “We also made a commitment to continue to support users of the Freebase API by creating a replacement API that would help with searching the Knowledge Graph.”

    “We will continue to support the Freebase API and widget for three full months after the Suggest Widget replacement is released, at which point we will shut it down,” they say.

    You can stay up to date on developments via the Google Developers channel.

    Image via Google

  • Google Pilots Local Business Markup for Knowledge Panels

    Google Pilots Local Business Markup for Knowledge Panels

    Last week, Google launched the Google My Business API to let big businesses and third parties integrate with the Google My Business platform and publish updates to customers on Google Search and Google Maps.

    Businesses can use the API to set special holiday hours, for example, and update them across all locations. The special hours feature was added to Google My business last month.

    Developers can also use the API to create locations with names, addresses, phone numbers, categories, business hours, etc. They can also mark a business location as permanently closed, manage business photos, list/invite/remove managers on locations and business accounts, read listing state to identify Google update, duplicate and suspended locations, search/filter locations by name/category/label, and set the service area for a business by specifying a point and radius or Place ID.

    To work with the API, you’ll need to have a basic knowledge of RESTful APIs.

    Documentation for the API is here.

    Search Engine Land points to a Google help document in which it talks about a way to provide local business information via structured data markup. When utilized, uses who search for a business may see a Knowledge Panel card with details about the business.

    Google says, “We are currently piloting this feature with a restricted set of initial data providers. We hope to soon open up the feature so any provider who implements spec-compliant markup is eligible to participate.”

    “Use markup on your official website to provide Google with information about local business locations,” it says. “These attributes appear on the card and elsewhere throughout Google Search and Maps. For instance, a restaurant’s menu URL appears prominently on the card, letting users click through to see what kind of food and drink is available.”

    More about adding the markup here.

    Images via Google

  • Google Now Displays Fast Food Nutritional Info

    Google has displayed nutritional information in search results for about two years now – making it easy for you to quickly pull up the calorie counts for a head of cauliflower, or a bunch of grapes, for instance.

    Now, Google is expanding this to include nutritional info on popular fast food items.

    If you search “calories in big mac”, for example, here’s what you’ll see:

    Screen Shot 2015-06-16 at 11.26.51 AM

    Searching a more generic phrase like “calories in McDonald’s” will put up a scroll box of all the fast food chain’s menu items, complete with calorie counts.

    Screen Shot 2015-06-16 at 11.29.20 AM

    Right now this only works for a handful of restaurant chains inside the Google app, as well as on desktop search.

  • SEO Vet Bill Slawski On Google’s Knowledge And Answers

    SEO Vet Bill Slawski On Google’s Knowledge And Answers

    In case you haven’t noticed, Google has changed a lot over the years, and much of the search engine’s focus is on showing users answers and information about what they’re looking for directly on search results pages through things like direct answers and Knowledge Graph.

    Is the Google experience improving as a result? Share your thoughts in the comments.

    On top of that, Google thinks it can effectively dig into websites and distinguish what is fact from what is not, and use that as a ranking signal. This was explained in a recently released research paper about what Google calls “Knowledge-Based Trust,” though the company maintains it is not using it at this time.

    Google owns a ton of patents, and there are clues sprinkled throughout that trove about its methods for implementing this stuff. With that in mind, we reached out to THE person who spends more time analyzing Google patents than anybody else on the web, to get some perspective on how Google is doing with these initiatives.

    If you follow the SEO industry, Bill Slawski needs no introduction, but just in case, he is the president and founder of SEO by the Sea as well as the Director of Search at Go Fish Digital. His blog has been going strong for the past decade, and is without a doubt the best place to read analysis of Google’s search-related patents.

    Let’s get to the Q&A.

    In your opinion, how is Google doing with “knowledge” right now? Are they getting it right? Are there too many errors? How do you feel, overall, about how Google has progressed here since first implementing Knowledge Graph?

    Bill Slawski: Google is trying a multitude of approaches in responding to Knowledge Graph answers. One of the biggest areas of change happening at Google and at Bing right now is the evolution of search results. With Google, we are seeing increasing numbers of Direct answers, Structured snippets, and rich snippets in response to queries. The Answer Box has been around at Google for a few years, and when it first started out, it tended to be filled with vertical search results that it thought might be appropriate in response to a query – responses filled with News results, Local results, Book results. and others that were different from the 10 blue links that Google had been showing searchers. In Google’s Financial 10 K statement for 2014 they stated that they would be trying to provide more direct answers for natural language queries:

    It’s been that way from the beginning; providing ways to access knowledge and information has been core to Google and our products have come a long way in the last decade. We used to show just ten blue links in our results. You had to click through to different websites to get your answers, which took time. Now we are increasingly able to provide direct answers — even if you’re speaking your question using Voice Search — which makes it quicker, easier and more natural to find what you’re looking for.

    Google did have a program that was run by Andrew Houge, now the director of Engineering at FourSquare, which he referred to in an online copy of his Resume, the “Annotation Framework” which resulted in a number of knowledge Web based patents being developed at Google, including Google Maps, which seems like a proof of concept for the creation of a knowledge-based index. He then was involved in the Acquisition of Meta-Web , which ended up bringing a number of new search engineers to Google focused upon Semantic Web Technology.

    Google has had other people involved in bringing knowledge graph technology to the search engine, including their Head of structured search Alon Halevy, who was involved in The WebTables project that is being used for query refinements in response to different queries and Google’s Structured Snippet, which enrich snippets with content from tables found on pages that are being indexed. Also Google’s Ramanathan Guha, inventor of Google Custom Search Engines, and Google Rich Snippets (his name is on the Google Blog post that introduced rich snippets to searchers.)

    It’s clear from that financial statement that the audience Google is responding to with queries are searchers and not site owners.

    How about direct answers? Are they doing an adequate job or is there a great deal of room for improvement? Any particular niches you think Google is doing a better job at than others?

    BS: Google has been increasing the number of direct answers they show searchers, and recently overhauled the medical answers they were showing searchers, improving those with input from people at the Mayo Clinic and with Google Medical Staff. A White paper that Google released on “How-to” type direct answers described how Google was using a Semantic Sense framework to try to understand the sources of such content better. See: Cooking with Semantics.

    What do you make of the recently released research paper about “knowledge-based trust” as a ranking signal? Google has said it’s not using this in search right now, but do you expect them to? Do you believe they should, and that it should carry more weight than links/PageRank?

    BS: The team that released this Knowledge-based Trust paper seemed very familiar – Many of them were among the people releasing a paper on Google’s Knowledge Vault, which identified a number of ways that Google could potentially improve the quality of information in Google’s knowledge Base. If you were following along at the time that paper was released, it described a number of approaches that the knowledge graph could be improved, and people were talking about the knowledge Vault being a replacement for Google’s Knowledge Panel at the time. They did describe a number or ways of improving the quality of information in the Knowledge Panel. As I look at more and more patents and papers about Google Knowledge, it does appear that “authoritiative source” information for things like direct answers are located based upon things involving Google’s link graph, like the rankings of pages in response to certain queries, or how often those pages get clicked upon when they are displayed as a search result for a query that might be relevant to the topic of a direct answer. If that is how Google is selecting “authoritative” sources for answers, than using a Knowledge-based Trust approach doesn’t sound like a bad way to go.

    Roughly how many patents by your estimate does Google own that you can see directly applying to Knowledge Graph and related features? How many do you see specifically related to Knowledge-based trust?

    BS: Questions like this one are why the idea of a Web where better use of Data on the Web might be helpful. Patents at the USPTO (US Patent and Trademark Office) are classified into different categories, but there’s no easy way to sort those patents into those different categories. It would be nice if it were easier to use that classification data to answer a question like this one. Instead I’m left to try to guess. I’ve been searching for patents related to the Knowledge graph. There are a number that have been published at WIPO (The World Intellectual Property Office). I’ve written about a number of Patents published (released as patent applications or granted patents) by Google that directly apply to the knowledge graph, and there are some that I haven’t written about because i had wondered how helpful it might be to write about them. The annotation Framework patents I mentioned earlier, there may have been around 30 or so, and there may have been at least as many others. I haven’t seen any patents that I can say are specifically related to “Knowledge-Based-Trust, though it’s possible that some may have been filed, and not yet made public by the USPTO or WIPO.

    It seems like Google does a pretty good job (for the most part) of determining factual information for its own direct answers. It also seems like Google could apply some of this to KBT. Do you see KBT and the work Google has done with direct answers as related at all?

    BS: It does seem like the Knowledge-based-Trust approach could lead to better direct answers.

    In general, how much of what Google patents ends up being put to use in your estimation?

    BS: There’s so much range to what Intellectual property that Google attempts to protect with patents, that I think it’s impossible to make that estimation. I don’t usually spend too much time on patents from them that have little to nothing to do with search (unless they cover something like a cure for cancer).

    Is Google’s work with direct answers a serious detriment to webmasters or is this blown out of proportion?

    BS: Google’s work with direct answers appears to be a natural evolution of what searchers appear to want in search results – and make it more likely that searchers will continue to use a search engine to perform searches on Mobile devices and using spoken queries.

    What are some search-related Google patents you’ve analyzed that you don’t see Google utilizing, but that you think it should?

    BS: It’s difficult to answer this because it’s sometimes hard to determine whether or not Google implemented some patents. Google announced recently that they would be taking action against doorway pages on site. I wrote a post about a Google patent granted in 2007 that was originally filed in 2003, that seemed like it addressed many issues related to doorway pages, and yet they are announcing they are going to come out with a new algorithm to address that problem?

    —–

    Indeed they are. You can read more about that here.

    Slawski recently wrote a series of posts about Google’s Direct Answers, which if you should also take a look at. Those start here.

    Do you want to see Google implement the Knowledge-Based Trust signal? Let us know in the comments.

    Image via SEO by the Sea

  • Drink Up – Google Wants to Make You a Bartender

    No, Google won’t serve you drinks – but it’ll tell you how to make them.

    Google has added cocktails to the list of instructables it’ll show with the Knowledge Graph – so if you’re wondering how to make a tequila sunrise, Google has your back. Alternatively, you could just put the tequila in a glass and drink it. That’s the preferred method of tequila consumption.

    Searching for the phrase “how do I make an old fashioned” or even just “how to make old fashioned” yields a new card that explains the basics of the drink-making process.

    The recipes aren’t really recipes, as with the ones for food – they’re more generic instructions (mix some bourbon) – but it’s a pretty nice basic outline of how to make the drink.

    Image via Thinkstock

  • Should Facts Outrank Links As A Google Signal?

    Should Facts Outrank Links As A Google Signal?

    Google has a new paper out that discusses how it might rank pages based on facts as opposed to links. If this were to become the case, it would represent a huge move for the search engine, which has historically used links as a major indication of relevance. In fact, it was the PageRank algorithm that really put Google on the map in the first place, and led to the search engine overtaking other players like Yahoo years ago.

    Do you think Google’s algorithm would be better off with the approach discussed in the paper? Let us know in the comments.

    These days, Google has at least over 200 signals it uses to rank content, but links are still a significant part of that. Just how significant they is debatable, particularly as Google includes more and more content and answers directly in its search results.

    Of course just having this paper doesn’t mean that Google has implemented such a ranking strategy, nor does it necessarily mean that it will. The company has countless patents, and not all of them are in use. That said, the fact that Google has been researching this, and has indeed authored a paper on it, combined with the moves the search engine has already made, suggest that this is something Google could implement at some point.

    You can read the whole paper here. The abstract reads as follows:

    The quality of web sources has been traditionally evaluated using exogenous signals such as the hyperlink structure of the graph. We propose a new approach that relies on endogenous signals, namely, the correctness of factual information provided by the source. A source that has few false facts is considered to be trustworthy.

    The facts are automatically extracted from each source by information extraction methods commonly used to construct knowledge bases. We propose a way to distinguish errors made in the extraction process from factual errors in the web source per se, by using joint inference in a novel multi-layer probabilistic model.

    We call the trustworthiness score we computed Knowledge-Based Trust (KBT). On synthetic data, we show that our method can reliably compute the true trustworthiness levels of the sources. We then apply it to a database of 2.8B facts extracted from the web, and thereby estimate the trustworthiness of 119M webpages. Manual evaluation of a subset of the results confirms the effectiveness of the method.

    So, they’ve confirmed the effectiveness of this method. That’s interesting. And if that wasn’t enough to get you thinking about where Google might be headed, the opening paragraph of the paper’s introduction pretty much discredits links as a valuable signal:

    Quality assessment for web sources is of tremendous importance in web search. It has been traditionally evaluated using exogenous signals such as hyperlinks and browsing history. However, such signals mostly capture how popular a webpage is. For example, the gossip websites listed in mostly have high PageRank scores, but would not generally be considered reliable. Conversely, some less popular websites nevertheless have very accurate information.

    Curious about which “gossip sites” they’re referring to? Well, the section it points to points readers to this list of the top 15 most popular celebrity gossip sites, which include: Yahoo! OMG!, TMZ, E Online, People, USMagazine, WonderWall, Gawker, ZimBio, PerezHilton, HollywoodLife, RadarOnline, PopSugar, WetPaint, MediaTakeOut, and FishWrapper.

    Later in the paper, it notes that among these fifteen sites, fourteen have a PageRank among the top 15% of websites due to popularity, but for all of them, the KBT are in the bottom 50%.

    “In other words, they are considered less trustworthy than half of the websites,” it says. It also says that forum websites tend to get low KBT, specifically calling out an example of inaccurate info found on Yahoo Answers, which you’ve probably seen ranking highly in Google results repeatedly.

    The paper does also note that KBT as a signal is orthogonal to more traditional signals like PageRank. It also appears to hint at identifying content that is irrelevant to the main topic of a website.

    This all really just scratches the surface of what the paper itself gets into, so feel free to jump in there for a deeper dive into what we’re dealing with.

    In theory, what Google is proposing here could lead to some major improvements to search rankings. It makes some really good points. Chief among them is the one that popularity isn’t necessarily the best indicator of relevance.

    Questions will remain, however, about just how well Google really is able to distinguish fact from fiction and/or fact versus outdated information. We’ve seen Google struggle with this time and time again with its Knowledge Graph. If Google’s “knowledge” is to become the backbone of ranking in the way that PageRank has been historically, it could open the algorithm up to potential errors.

    That said, given that Google uses so many signals, and this would still just be one of them, I personally feel like this could be a more legitimate signal than PageRank. It’s been well-documentecd how links can be manipulated while Google plays whack-a-mole both manually and algorithmically. This might be harder for evildoers to game. Facts would certainly be harder to buy, although you have to wonder how the native advertising/sponsored content industry will play into this.

    For now, it’s all theoretical anyway. You should really be more concerned with getting your site mobile-friendly. This is an actual signal Google will launch next month. If you have an Android app, you should get it set up for app indexing. These are the things that can make a difference in the near term.

    Do you want to see Google take the fact-over-PageRank approach? Let us know in the comments.

  • Google Starts Returning Quick Medical Facts

    Google is about to start displaying more medical information in its Knowledge Graph, enabling users to quickly search and retrieve important health info without necessarily having to click through to other sites. The company says it has been working with a team of doctors led by its own Dr. Kapil Parakh, M.D., MPH, Ph.D. to compile, curate, and review the information it shows.

    “All of the gathered facts represent real-life clinical knowledge from these doctors and high-quality medical sources across the web, and the information has been checked by medical doctors at Google and the Mayo Clinic for accuracy,” says product manager Prem Ramaswami. “That doesn’t mean these search results are intended as medical advice. We know that cases can vary in severity from person to person, and that there are bound to be exceptions. What we present is intended for informational purposes only—and you should always consult a healthcare professional if you have a medical concern.”

    “Think of the last time you searched on Google for health information,” Ramaswami says. “Maybe you heard a news story about gluten-free diets and pulled up the Google app to ask, ‘What is celiac disease?’ Maybe a co-worker shook your hand and later found out she had pink eye, so you looked up ‘pink eye’ to see whether it’s contagious. Or maybe you were worried about a loved one—like I was, recently, when my infant son Veer fell off a bed in a hotel in rural Vermont, and I was concerned that he might have a concussion. I wasn’t able to search and quickly find the information I urgently needed (and I work at Google!).”

    Ramaswami says his son was indeed OK (hopefully a doctor’s assessment rather than Google’s), and notes that 1 in 20 Google searches is health-related.

    The new information sounds like a major step up from what Google has offered in the past. Before the Panda update, there were some pretty questionable articles ranking for some health-related queries. We’re talking brain cancer articles from eHow written by non-medical professionals.

    Google has been working on improving health search for years. In February fo 2012, the company started displaying lists of possible health conditions when the searcher typed a query for a symptom.

    Interestingly enough, results for that same query look more like an old school SERP these days:

    Later that year, Google added new medical info to the Knowledge Graph, specifically for medications:

    The following year, the Knowledge Graph began to show nutrition information for foods:

    The latest Knowledge Graph additions seem like a major improvement to Google’s health-related search results. It’s good that all of this information is being reviewed by a team of doctors before inclusion, which would suggest a better review process than some of the other Knowledge Graph info has been subjected to in the past.

    Still, if it’s important, don’t rely on Google.

    Images via Google

  • Your Brand’s Social Profiles Can Now Appear In Google’s Knowledge Graph

    As some have noticed, Google has started displaying social media profiles in the Knowledge Graph area of search results pages for many brand searches. If you search for Wendy’s, for example, you can very clearly see links to their profiles on Facebook, Twitter, YouTube, LinkedIn, and Google+.

    The same holds true for Chase, Nike, Starbucks, and any other number of brands you can think of. This is good news for consumers who wish to engage with brands, without sifting through search results trying to find the appropriate channel. This makes them very easy to find.

    It’s also good for businesses for that very same reason, and the even better news is that it’s not limited to major brands like those mentioned. Google is offering businesses markup to enable the search engine to display social profiles for their brands.

    Google’s Zineb Ait Bahajji made the announcement on Google+ (via Search Engine Land):


    In case you didn’t feel like clicking to expand that post, supported profiles include Facebook, Twitter, Google+, Instagram, YouTube, LinkedIn, and Myspace. That’s right. You can even proudly display your Myspace profile. Google says it encourages you to specific additional social profiles as well, but for now, they won’t appear in the Google search results the way the others do.

    Google says its algorithms process the social profiles yo specify, and display the most relevant ones based on individual user queries, so people won’t always necessarily see the same ones. It’s also worth noting that Google will only show verified profiles for sites that have a verification process. That’s not good news for brands wanting to display their Twitter profiles, as Twitter is notoriously stingy about handing out verifications.

    You can find the documentation for setting up your structured markup here.

  • Google Is Shutting Down Knowledge Graph Source Freebase

    When Google first announced the Knowlege Graph, it named Freebase as one of its primary sources of structured data. It was named as one several public sources of information, which also included Wikipedia and the CIA World Factbook.

    Google announced that it will be closing Freebase, and moving toward the Wikimedia Foundation’s Wikidata, which is described as “a free linked database that can be read and edited by both humans and machines”. It acts as “central storage” for the structured data of Wikipedia, Wikivoyage, Wikisource, and others.

    WIkidata has simply been improving better than Freebase, so Google has decided to support that instead, and will be working to transition Freebase data over to Wikidata appropriately.

    “When we publicly launched Freebase back in 2007, we thought of it as a ‘Wikipedia for structured data,’ So it shouldn’t be surprising that we’ve been closely watching the Wikimedia Foundation’s project Wikidata] since it launched about two years ago,” Google’s Freebase says in a Google+ update (via Search Engine Roundtable). “We believe strongly in a robust community-driven effort to collect and curate structured knowledge about the world, but we now think we can serve that goal best by supporting Wikidata — they’re growing fast, have an active community, and are better-suited to lead an open collaborative knowledge base.”

    We should note that while Freebase has been around since 2007, it’s only belonged to Google since the company acquired Metaweb in 2010.

    Google is helping transfer Freebase to Wikidata, and in the middle of next year, it will wind down the Freebase service as a standalone project. They’ll also launch a new API for entity search powered by Google’s Knowledge Graph.

    “Loading Freebase into Wikidata as-is wouldn’t meet the Wikidata community’s guidelines for citation and sourcing of facts — while a significant portion of the facts in Freebase came from Wikipedia itself, those facts were attributed to Wikipedia and not the actual original non-Wikipedia sources,” Freebase explains. “So we’ll be launching a tool for Wikidata community members to match Freebase assertions to potential citations from either Google Search or our Knowledge Vault, so these individual facts can then be properly loaded to Wikidata.”

    You may remember hearing about Knowledge Vault earlier this year.

    It’s a system of Google’s, which stores information so that machines and people can read it. It’s basically Google’s giant database of facts. When you ask Google questions and get those direct answers, they’re likely coming from there. Rather than relying on crowdsourcing info like the Knowledge Graph, it uses an algorithm to pull info from the web, and turn it into raw data. You can get more in depth into it here.

    “We believe this is the best first step we can take toward becoming a constructive participant in the Wikidata community, but we’ll look to continually evolve our role to support the goal of a comprehensive open database of common knowledge that anyone can use,” Freebase says.

    Before the end of March, Google will launch a Wikidata import review tool and announce a transition plan for the Freebase Search API & Suggest Widget to a Knowledge Graph-based solution.

    On March 31, Freebase will become read-only, the website will no longer accept edits, and they’ll retire the MQL write API.

    On June 30, they’ll retire the Freebase website and APIs.

    Image via Freebase

  • Stephen Colbert Might Just Sue Larry Page’s Ass Over Knowledge Graph Screwup

    Apparently, Stephen Colbert is 5’11” – a height that puts him on par with the likes of Brad Pitt, Richard Gere, and Orlando Bloom. Colbert’s proud of this.

    But there’s a problem. The jerk-offs over at Google are slandering his good height. According to Google and its knowledge graph, Colbert is only 5’10”.

    Might as well be a Oompa Loompa.

    He took issue with this on last night’s episode of The Colbert Report, addressing Google’s Larry Page specifically.

    “Larry, I demand a retraction, and investigation, an apology, and a substantial cash settlement – or I will see your ass in court,” said Colbert, pointing into the camera.

    Colbert should know that this isn’t isolated to his height. Google has had plenty of issues keeping its Knowledge Graph, well, knowledgable. Seriously. You may as well call it the misinformation graph.

    Image via Colbert Report, Comedy Central screenshot

  • Here’s Some Potentially Troubling News About Google Search

    So you know those answer boxes Google has been showing in search results for a while now? The ones that extract text from third-party websites (possibly your own), to answer users’ queries without them having to click through to the site? Well, it looks like they’ve dramatically increased the frequency with which they’re doing this, and at times it’s inaccurate or outdated.

    Is this feature really the best for users? Does it harm webmasters? Let us know what you think in the comments.

    Moz has a new report finding that last week, Google jacked up the number of direct answer boxes it’s showing by as much as 98%. The report includes this graph showing the jump:

    The report says that many of the answers come from Wikipedia, as you’d expect, but Google clearly gets its answers from all kinds of sites. For example, it once turned to phillytown.com for this gem:

    It should be noted that after that one got some attention, they stopped showing a direct answer box for that query, in favor of the classic ten blue links-style results, which are led by Urban Dictionary with an even more disgusting (though not necessarily inaccurate) description.

    I guess Google didn’t want to own that one. This is followed by another Urban Dictionary result for “Upper Decker Double Blumpkin,” which you probably don’t want to read if you’re easily offended. Interestingly, the original phillytown.com result that Google once considered “the answer” to the question, isn’t even in on the first page of results.

    Not to get too far off base here, but the point is that these “answers” could easily come from any number of sources.

    Dr. Peter J. Meyers, the report’s author, says, “Many of these new queries seem to be broad, ‘head’ queries, but that could be a result of our data set, which tends to be skewed toward shorter, commercial queries.”

    He notes that one four-word query with a new answer box was ‘girl scout cookies types’.

    The fact that Google is increasing the number of direct answer boxes so drastically (and will probably continue to do so) is concerning to webmaters as it could mean Google sending them less traffic – particularly for their content that the search engine is actually showing to users. Remember, Google is also extracting data from websites now for its new “structured snippets,” which provide users with bits of information that keep them from having to click through to learn.

    Another concerning angle to Google’s approach is the accuracy of the answers it’s actually displaying.

    Google launched its Knowledge Graph over two years ago. We’ve seen quite a few times that it can provide questionable, outdated, and/or inaccurate results.

    In August, we learned about the “Knowledge Vault,” which is apparently the source of these third-party site-based answer results. From the sound of it, these answers are even riper for outdated and/or inaccurate data.

    Meyers points to one of his own articles that Google uses for one of the direct answers. He notes that the information in question was outdated, as it was an older article. He went in and updated the content to reflect accurate information, but Google hadn’t caught up with it even after two months.

    Google’s Knowledge Graph doesn’t always update as quickly as it should, but this appears to be even worse. Much worse, and that’s troubling considering how much they’re cranking up the volume on this type of search result.

    “At this point, there’s very little anyone outside of Google can do but keep their eyes open,” concludes Meyers. “If this is truly the Knowledge Vault in action, it’s going to grow, impacting more queries and potentially drawing more traffic away from sites. At the same time, Google may be becoming more possessive of that information, and will probably try to remove any kind of direct, third-party editing (which is possible, if difficult, with the current Knowledge Graph).”

    It’s interesting to think about Google becoming “possessive” of information it’s getting from other sites.

    Of course Google isn’t for sites, as the company frequently reminds us. It’s for users. Unfortunately, inaccurate and outdated information isn’t good for them either. Sadly, much of the inaccuracy and outdated information is likely to go unnoticed, as people aren’t likely to dig deeper into other results a lot of the time. The point is, after all, to get users the info they’re looking for without them having to dig deeper.

    Is this the right direction for Google Search to be taking? Share your thoughts in the comments.

    Image via Moz

  • New Google Feature Could Keep Traffic From Websites

    Google introduced a new feature for search results pages this week, which displays so-called facts in the snippets of specific results. It’s yet another way of Google displaying information on a page that could mean the user doesn’t have to bother clicking through to a third-party website.

    What do you think of this feature? Good or bad for users? For webmasters? Share your thoughts in the comments.

    The feature is called “structured snippets”.

    “Google Web Search has evolved in recent years with a host of features powered by the Knowledge Graph and other data sources to provide users with highly structured and relevant data. Structured Snippets is a new feature that incorporates facts into individual result snippets in Web Search, Google explains on its Research blog.

    The company shares this example for a result for the query “nikon d7100”.

    Google is displaying little “facts” that it deems interesting and relevant, and is doing so algorithmically. Here’s another one for the query “superman” as shown on a mobile device:

    “The WebTables research team has been working to extract and understand tabular data on the Web with the intent to surface particularly relevant data to users,” Google says. “Our data is already used in the Research Tool found in Google Docs and Slides; Structured Snippets is the latest collaboration between Google Research and the Web Search team employing that data to seamlessly provide the most relevant information to the user. We use machine learning techniques to distinguish data tables on the Web from uninteresting tables, e.g., tables used for formatting web pages. We also have additional algorithms to determine quality and relevance that we use to display up to four highly ranked facts from those data tables.”

    That bit about fact ranking is interesting. We recently heard about Google’s “Knowledge Vault,” which had already pulled in 1.6 billion “facts” with about 271 million of them ranked as “confident facts”. These are the ones, which Google believes there to be a 90% chance of being true. It would seem that there is room for error.

    In fact, Google implied that there will likely be some inaccuracies with the structured snippets.

    “Fact quality will vary across results based on page content, and we are continually enhancing the relevance and accuracy of the facts we identify and display,” the company said.

    In other words, not all of this stuff will necessarily be accurate, but hopefully more of it will be over time. How often will users know when they’re seeing inaccurate information? As we’ve seen with Google’s Knowledge Graph time and time again, this has been an area of concern. At times, it has even directly affected businesses with little apparent urgency on Google’s part for correcting errors.

    On how Google collects the information for structured snippets, Pierre Far, a webmaster trends analyst at the company, said in a Google+ post:

    It’s not structured data (schema.org) driven! Instead, it’s powered by algos that try to find interesting tables within webpages to extract the key facts related to the topic of the page. This deeper understanding of the contents, plus some quality checks, gives us a new kind of snippet.

    As Barry Schwartz at Search Engine Roundtable points out, these snippets mean users will have even less reason to click on third-party results, as it will mean more instant information right from the Google page.

    Unlike schema.org-driven data, Google is going out on its own and finding this algorithmically meaning that webmasters will have less control over when Google plucks such information from their pages to present “facts” and prevent clicks.

    Schwartz points to an interesting comment from a Webmaster, who said in a forum post: “Formerly, Google really distinguished itself from all other well-known properties on the web by being ‘the best place, bar none to find websites.’ If Google continues to transition from that paramount search engine to being merely ‘one of several places to find knowledge about nearly everything’ then it becomes more like Wikipedia, Freebase, Wolfram etc. And that means there will be less reason to visit Google, not more.”

    I’m not sure if I agree that people will use Google less when they’re getting information directly from it, but it could create some opportunities for other services that actually do want to help users find websites of interest.

    The fact is that we’re living in an increasingly mobile world, and along with that comes quick voice-activated searches, which is when Google’s quick answers come in most handy. When you’re on your phone, you don’t always want to have to navigate around the web. You want instant gratification.

    Google seems to value this concept greatly, even at the expense of the occasional inaccuracy, and certainly at the expense of sending traffic to other websites.

    Google has said it time and time again, but Executive Chairman Eric Schmidt reminded everybody once again a couple weeks ago: “We built Google for users, not websites.”

    He was speaking about Google’s dominant position in Europe, where the EU is awaiting the company’s latest proposal with new concessions to avoid formal antitrust charges and fines.

    Schmidt said Google is not the “gateway to the Internet as the publishers suggest,” and that “to get news, you’ll probably go direct to your favorite news site. It’s why newspapers like Bild, Le Monde and the Financial Times get most of their online traffic directly (less than 15% comes from Google). Or you might follow what other people are reading on Twitter. To book a flight or buy a camera for your next holiday, you’re as likely go to a site like Expedia or Amazon as you are Google. If you’re after reviews for restaurants or local services, chances are you’ll check out Yelp or TripAdvisor. And if you are on a mobile phone — which most people increasingly are — you’ll go straight to a dedicated app to check the sports scores, share your photos or look for recommendations. The most downloaded app in Europe is not Google, it is Facebook Messenger.”

    “Nor is it true to say that we are promoting our own products at the expense of the competition. We show the results at the top that answer the user’s queries directly (after all we built Google for users, not websites),” he added.

    He went on to say that he thinks it’s okay to rank weather sites lower and give them less traffic when Google shows local weather at the top of the page because it’s “good for users.” He said it’s the same if you want to buy something, and that if you want directions somewhere, a Google Maps result is a “great result for users.”

    Increasingly – as the new structured snippets show – Google thinks the Google results page itself is a “great result for users.”

    What do you think about structured snippets? Good or bad? Let us know in the comments.

    Images via Google

  • Google Introduces Structured Snippets

    Google announced the launch of “structured snippets,” a new feature that puts “facts” in the snippets of web results. As with Google’s Knowledge Graph, these facts may or may not be accurate.

    Here’s what they look like:

    The company says, “The WebTables research team has been working to extract and understand tabular data on the Web with the intent to surface particularly relevant data to users. Our data is already used in the Research Tool found in Google Docs and Slides; Structured Snippets is the latest collaboration between Google Research and the Web Search team employing that data to seamlessly provide the most relevant information to the user. We use machine learning techniques to distinguish data tables on the Web from uninteresting tables, e.g., tables used for formatting web pages. We also have additional algorithms to determine quality and relevance that we use to display up to four highly ranked facts from those data tables.”

    “Fact quality will vary across results based on page content, and we are continually enhancing the relevance and accuracy of the facts we identify and display,” Google adds.

    Well, that’s encouraging. Not all of this stuff will necessarily be true, but hopefully more of it will be over time.

    Image via Google

  • Google To Keep Even More Traffic For Itself

    Google announced the launch of Knowledge Graph a little over two years ago. It’s already become inadequate for Google’s needs. The search engine wants to be able to quickly answer all of your questions without having to send you to those pesky other websites, and while Knowledge Graph is helpful in many cases, it just won’t do for all queries.

    You may be thinking that at as long as that’s the case, Google will still have reason to send people to other sites, but increasingly, that is becoming less the case. Welcome to the era of the “Knowledge Vault.”

    Do like Google’s strategy of increasingly attempting to provide content for users on its own, or should they be sending traffic to third-party sites more often like they used to? Share your thoughts in the comments.

    New Scientist recently published an article about Google’s “Knowledge Vault,” (via Search Engine Land) which we’ll likely be hearing a lot more about as time goes on.

    Here’s a snippet of explanation:

    Knowledge Vault is a type of “knowledge base” – a system that stores information so that machines as well as people can read it. Where a database deals with numbers, a knowledge base deals with facts. When you type “Where was Madonna born” into Google, for example, the place given is pulled from Google’s existing knowledge base.

    This existing base, called Knowledge Graph, relies on crowdsourcing to expand its information. But the firm noticed that growth was stalling; humans could only take it so far.

    So Google decided it needed to automate the process. It started building the Vault by using an algorithm to automatically pull in information from all over the web, using machine learning to turn the raw data into usable pieces of knowledge.

    Based on how the Knowledge Vault is described in the article, it sounds like the basis for those quick answers you often see at the top of search results that provide information or “answers” from the text of third-party webpages. Sometimes these answers are questionable.

    Fun Fact: Knowldege Vault is already something that exists outside of Google.

    We learn from the New Scientist article that the Knowledge Vault has already pulled in 1.6 billion “facts” and that 271 million of them are rated as “confident facts”. That means Google believes there’s more than a 90% chance they’re true.

    You have to wonder if that level of confidence should be considered to be high enough to be presented as knowledge. It is, after all, essentially displacing the traditional search results, which have historically given users an opportunity to see for themselves what the best answer to a query was, based on the their own judgment. Frankly, it’s a little scary as Google has shown time and time again that it can present erroneous information as “knowledge”.

    Of course the whole thing also serves to drive less clickthrough to third-party sites, which isn’t bound to sit well with webmasters. It’s only a continuation of a trend that has been steadily building over recent years.

    As we live in an increasingly mobile search-driven world, where much searching is done by voice, it’s understandable that Google would want to give these quick answers. From the user perspective, it’s often helpful, and Google is all about the user experience, as the PR team will often tell you.

    Whether or not this is actually a good direction for search to be moving in is debatable. Google has long talked about wanting to be a Star Trek-like computer, and it’s certainly getting closer to that. Star Trek, however, was fiction. I’m not a big Trekkie, so I might be wrong, but I’m going to assume that the accuracy of the information given by the computer was often taken for granted. Were there any episodes where the crew visited various third-party websites to fact-check the information they were presented with? For that matter, was the computer run by a giant company motivated by advertising revenue? Did the computer push aside information from competing sources, and sometimes accidentally include branded info? I’m asking. I honestly don’t know.

    Do you trust the Knowledge Vault to deliver the right information? Is Google’s evolution problematic for websites? We want to know what you think. Tell us in the comments.

    Image via YouTube

  • Google’s Knowledge Graph May Get A Timeline Feature

    Google is testing a new feature for its Knowledge Graph, which lets users explore timelines for select queries.

    Florian Kiersch, who frequently identifies Google features in testing, posted a video showing the functionality on Google+ (via Search Engine Roundtable):


    It’s unclear what all types of queries might display this type of result, but the video shows queries for Google, Andy Rubin, and “world war”.

    This could turn out to be a pretty helpful tool, particular for studying, though until Google’s Knowledge Graph becomes a more reliable source of information, I wouldn’t put too much stock into the accuracy of the information presented, at least without fact checking.

    Google was even recently spotted showing brand results in the Knowledge Graph, though this appears to have been a mistake. But that’s part of the problem. Google seems to make a lot of mistakes in its “Knowledge” Graph.

    Image via Google+

  • Google Has a Jesus-Shaped Hole in Its Graph

    We already know that Google hates America, so it should be no surprise that the largest search engine in the world also hates Jesus.

    Search Engine Roundtable’s Barry Schwartz posted about a thread on the Google Web Search Help forum, in which one Google user says that he’s noticed something amiss with Google’s knowledge graph results. Apparently, Jesus doesn’t get his own knowledge graph result, but other prominent religious figures like Muhammad, Moses, and Buddha do.

    “Why the absence of Jesus?” asks Steve.

    I don’t know, Steve. You might be asking a bigger question than you think. But before tackling anything like that, let’s make sure Google is really throwing Jesus this major diss.

    Nope, no knowledge graph there.

    You won’t find a knowledge graph in searches for Jesus Christ, Christ, or Jesus of Nazareth either. In fact, for that latter search, Google suggests you see results about the 1977 television miniseries, Jesus of Nazareth, which I’m positive is what most people who search that phrase are looking for in the first place. Good on you, Google.

    Ok, so maybe Google just stays away from religious figures in knowledge graph results – just to play it safe.

    Nope.

    You’ll also find knowledge graphs for Moses, Buddha, and biblical figures like Judas Iscariot, King David, Solomon, and many more – even Adam and Eve.

    So, once again, where’s Jesus?

    Much of Google’s knowledge graph pulls from Wikipedia, and Jesus has quite the thorough Wikipedia page. The snippet from Wikipedia that would appear in a hypothetical knowledge graph box isn’t controversial, really:

    Jesus, also referred to as Jesus of Nazareth, is the central figure of Christianity, whom the teachings of most Christian denominations hold to be the Son of God. Christianity regards Jesus as the awaited Messiah of the Old Testament and refers to him as Jesus Christ, a name that is also used in non-Christian contexts.

    Yeah, that about sums it up. So why has Google purposefully removed Jesus’ knowledge graph?

    It turns out, Google is pretty inconsistent when it comes to religious figures and the knowledge graph. While you find graphs for Buddha, Muhammad, and Moses, you won’t find a graph for Vishnu, the Supreme God of Vaishnavism, one of the three main sects of Hinduism or Shiva, another popular Hindu deity. You won’t find a graph for Kirshna either.

    There’s no knowledge graph for God, Allah, or Yahweh either. What do Jesus, Vishnu, Allah, and these left-out entities have in common?

    They’re all God, or gods. Sure, Jesus Christ is also believed to be a human, the son of God. But in Christian teachings, Jesus is also God himself. It looks like Google is simply shying away from assigning any sort of god their own knowledge graph.

    “Possibly because of the ‘controversy’ surrounding Jesus. Everyone accepts who Muhammad, Moses and Buddha are, but not every-one accepts Jesus is the Son of God. If they define him as a prophet, Christians could take offense, if they define him as the Son of God, Muslims could take offense…. that is one argument it might be better to stay out off (if you are Google),” one Google user replies to Steve’s original question.

    Jesus doesn’t have a graph because Google isn’t touching that with a 39 and a half-foot pole. Google’s just staying away from deities. Even do a search for ancient deities – Ra, Zeus – no knowledge graph.

    Or, Google hates Jesus. You’ll most definitely find a graph result for The Big Bang.

    Image via Wikimedia Commons

  • Google Puts Knowledge Graph Info In Ads

    Google is testing a new feature for search ads, which includes links to popups that contain Knowledge Graph information. It appears to replicate a feature that Google already utilizes in organic results.

    This was launched in January:

    Typically, like most Knowledge Graph descriptions, info is taken from Wikipedia. Recently, it was discovered that Google is actually tapping DMOZ for some of it.

    Arianna Wolf got some screenshots of the feature being used in ads, which were picked up by Search Engine Land.

    Here’s what it looks like:

    When clicked, it produces essentially the same effect as what you see in the first image from the organic results.

    Apparently the feature only appears on the ads that appear above the organic search results (as opposed to those on the side of the page).

    We’ve been unable to reproduce the test. It’s unclear how big it actually is.

    As Ginny Marvin at SEL points out, this feature could potentially drive people away from the actual ad click, similar to how the Knowledge Graph can already take traffic away from organic results. On the other hands, if it has a Wikipedia page or something else to point to, it could lend a bit of credibility to an advertiser the user isn’t familiar with.

    Image via Google

  • Google Extends Knowledge Graph To Google Maps

    Google continues to expand its Knowledge Graph in different ways. It’s now being extended to Google Maps and getting video content.

    James Gibbons spotted a new Google Maps feature called Quick Facts, which appears when you search for some locations. Search Engine Land confirmed that the content in these cards comes from the Knowledge Graph. Here are a couple of examples.

    As far as the video content goes, Google has been showing video thumbnails on searches for basketball teams. Here’s what it looks like:

    If you click the video, it opens up right on the search results page. In this example, it’s an official NBA YouTube video of highlights from the latest game.

    It’s unclear if Google is doing this with other types of searches.

    Images via Google

  • Google Drops Some ‘Upper Decker’ Knowledge (Courtesy Of Another Parked Domain)

    Remember that story from the other day about Google’s questionable “answers” as it relies on websites to fill in the gaps in “knowledge” that its proper Knowledge Graph can’t answer?

    Well this one’s just funny.

    Just think of the traffic Urban Dictionary is missing out on. Oh, and the source is a parked domain again. Seriously, read this.

    Thanks, phillytown.com/glossaryhtm!

    Via Gizmodo

    Image via Google

  • Are You Buying The Answers Google Is Selling You?

    Google is once again displaying some questionable content in its Knowledge Graph-style results, and displaying it as the answer to your question.

    Have you personally come across questionable content in the Knowledge Graph? Let us know in the comments.

    “We try to take parked domains out of our results…”

    That’s a quote from Matt Cutts in Google’s latest Webmaster Help video. These videos are sometimes uploaded months after being recorded, but we’ll assume that this is still Google’s policy.

    Google doesn’t have to look much further than its own Knowledge Graph-style results to find parked domains, apparently. Check out what comes up when you search for “What is guest blogging?” (a highly relevant query these days):

    This was first spotted by Andrew Steel (via Search Engine Roundtable).

    Google’s answer is: “(guest bloggers) Someone who posts an article on a blog that is not their own. Their incentive for doing so is getting backlinks to increase their own site’s search engine ranking.”

    OK. A couple things about this.

    Parked Domains

    The source of Google’s answer here is moneyonlinemaking.com/learn-terms. The link takes you to…you guessed it…a parked domain. I don’t know why I’m surprised given how often we’re seeing questionable content come up in Google’s Knowledge Graph and similar-style results.

    We’ve seen Google display inaccurate business information in the Knowledge Graph. We’ve seen inaccurate marital status information. We’ve seen it show a man’s death as occurring before his birth. We’ve seen it confuse Brandy the spirit with Brandy the entertainer. We’ve seen it accidentally display nudity. We know it can struggle with real time. During the World Series last year, some Wikipedia vandalism led to Google displaying information for the St. Louis Cardinals, calling them a “gay butt sex team”.

    Since early this year, Google has been turning to websites to fill in its gaps in “knowledge,” when providing the quick answer-style results. That appears to be what we’re seeing in the guest blogging example. As I said last month, when we reported on this, you have to wonder what the potential for error in these types of answers is, considering how often we’ve seen errors in the actual Knowledge Graph. This particular example isn’t so much an error as a biased perspective, but that’s not what these “answers” are for (as far as I know).

    How can users expect Google to provide relevantly ranked search results when it has so much trouble getting its alleged “answers” right? These are supposed to be the absolutely most relevant results for queries where they appear. If Google’s unsure, it’s supposed to offer you an alternative. For instance, if you search for “orange”, there are several things you could mean, so Google shows you this (curiously there’s no fruit option for an “apple” search):

    Having an absolute answer (as in the guest blogging example) is telling users that Google is pretty sure this is the answer to your query, which brings us to the next point.

    Questionable Answer

    Google has been sending a message to people engaging in guest blogging for SEO purposes. You know the story. If not, read this. But is guest blogging, in general, the same thing as guest blogging for SEO? No. There are other reasons to write guest blog posts (believe it or not). Not everything is about Google. Matt Cutts even acknowledged as much when he had to clarify his post about it earlier this year – the one he pointed to when announcing the penalization of a guest blog network. That happens to be the same post that Google displays as the top organic result for the query in question.

    So why would Google display this SEO-related “definition” from a parked domain in the big box at the top? Is it trying to further its own message by finding a source that matches it? Are there no other definitions out there?

    This one at About.com seems reasonable: “Guest blogging is used by bloggers as a way to network with other people within the blogosphere, grow relationships with other blog readers, and increase traffic for their own blogs.”

    That happens to be the top result, though not in direct answer style, on Bing. Bing doesn’t have a matter-of-fact definition, so it just gives you the old fashioned organic search results (god forbid), and you get a legitimate definition on the first result. Isn’t that basically how Google used to work? Isn’t the Knowledge Graph supposed to improve search? In fact, I would go so far as to say Bing wins the “Bing it On challenge” hands down on this one, displaying some nice Quora content in the side column to supplement the organic results.

    Yes, this is just on example, but as discussed above, it’s far from the only example of questionable content being thrown in our faces as “the answers” to our questions. Also, it could be some coincidence that the answer Google pulled here just happened to fit directly with its own messaging related to guest blogging, but it looks biased at the very least.

    Not all guest blogging is about “getting backlinks to increase their own site’s search engine ranking.” Sometimes people want to make a name for themselves. Sometimes people simply want to increase their exposure.

    Even Cutts said on his personal blog, “There are still many good reasons to do some guest blogging (exposure, branding, increased reach, community, etc.). Those reasons existed way before Google and they’ll continue into the future. And there are absolutely some fantastic, high-quality guest bloggers out there.”

    But if you ask Google as of the time of this writing simply what guest blogging is, you’re going to be told that its strictly for search engine ranking, and told by a search engine that tries to keep parked domains out of its search results but is failing to keep them out of their own “answer” results.

    Could it be that it’s simply easier for Google to not have to determine the motives behind your guest blogging, and just wants you to not do it?

    I’ll be surprised if the answer doesn’t change soon as somebody gets wind of the articles being written about it, but either way, like the St. Louis Cardinals example, this is Google giving questionable “knowledge” at a highly relevant time.

    Update: Google has updated the the search results page that is the subject of this article. Rather than drawing from a different source, they’re no longer displaying an answer box. Now the Cutts post is the top result. As predicted, they probably saw this article or another talking about it, and pulled it. That doesn’t mean there aren’t other examples out there in the wild, in which Google is giving people questionable answers.

    Generally speaking, do you trust the answers Google gives you with its Knowledge Graph results? Let us know in the comments.

  • Yahoo Appears To Be Testing Its Own Version Of Knowledge Graph

    Yahoo appears to be testing a new type of search results that resemble Google’s Knowledge Graph and Bing’s version of that.

    The test was spotted by All Google Testing (via Search Engine Land), which shared this look:

    Here’s what it looks like for a search for “Apple,” followed by Google’s and Bing’s results for the same query for comparison:

    As you can see, the look is quite similar, though Yahoo’s results are a bit different. They have, at least for these examples, a “Top Products” section for those looking to do some shopping. There’s also a “You may also like” section, which puts competitors right in front of users. It’s also missing some of the info provided by the other search engines like stock price, CEO, founded date, etc.

    It’s worth noting that a search for Apple first took me to a page showcasing a local search result box on the right side pointing me to the local Apple store. I had to click another box to take me to the Apple results pictured.

    All Google also has some steps for enabling the test yourself, though some have had trouble getting it to work.

    While it’s just a test, Yahoo would do well to have this kind of thing if it’s going to make a legitimate case for replacing Google on iOS. It also recently replaced its local business reviews with those from Yelp, which could also help (though some businesses are pretty frustrated with the change).

    Images via YouTube, Google, Bing