{"id":39716,"date":"2016-03-02T15:00:06","date_gmt":"2016-03-02T23:00:06","guid":{"rendered":"http:\/\/www.bruceclay.com\/blog\/?p=39716"},"modified":"2019-08-08T11:06:43","modified_gmt":"2019-08-08T18:06:43","slug":"rankbrain-google-machine-learning-system-smx","status":"publish","type":"post","link":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/","title":{"rendered":"RankBrain: What Do We Know About Google&#8217;s Machine-Learning System? #SMX"},"content":{"rendered":"<p>SEO is\u00a0very tactical and we always try to look behind the curtain of Google&#8217;s algorithms. So, it&#8217;s no surprise that we all want to know more about RankBrain. RankBrain is Google&#8217;s machine learning system that they <a href=\"https:\/\/www.bloomberg.com\/news\/articles\/2015-10-26\/google-turning-its-lucrative-web-search-over-to-ai-machines\" target=\"_blank\">confirmed<\/a> out of the blue in October 2015.<\/p>\n<figure id=\"attachment_39990\" aria-describedby=\"caption-attachment-39990\" style=\"width: 300px\" class=\"wp-caption alignright\"><img loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-39990\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-speakers.jpg\" alt=\"Marcus Tober, Eric Enge, Danny Sullivan\" width=\"300\" height=\"247\" \/><figcaption id=\"caption-attachment-39990\" class=\"wp-caption-text\">Speaking on RankBrain are Marcus Tober (left) and Eric Enge, pictured with moderator Danny Sullivan<\/figcaption><\/figure>\n<p>In this session, we&#8217;ll learn about RankBrain via the studies done by our presenters.<\/p>\n<p><strong>Moderator<\/strong>:<\/p>\n<p>Danny Sullivan,\u00a0Founding Editor, Search Engine Land (<a href=\"https:\/\/twitter.com\/dannysullivan\" target=\"_blank\">@dannysullivan<\/a>)<\/p>\n<p><strong>Speakers:<\/strong><\/p>\n<ul>\n<li>Eric Enge,\u00a0CEO, Stone Temple Consulting (<a href=\"https:\/\/twitter.com\/stonetemple\" target=\"_blank\">@stonetemple<\/a>)<\/li>\n<li>Marcus Tober,\u00a0Founder\/CTO, Searchmetrics Inc. (<a href=\"https:\/\/twitter.com\/marcustober\" target=\"_blank\">@marcustober<\/a>)<\/li>\n<\/ul>\n<p>Danny Sullivan says that a\u00a0few months ago, Google said, &#8220;Oh hey, by the way, we have this new thing called RankBrain and it&#8217;s the third most important signal that factors into ranking.&#8221; SEOs asked, &#8220;Can you tell us about that?&#8221; Google said no. We still want to understand it.<\/p>\n<h2>Marcus Tober: Machine Learning Ranks Relevance<\/h2>\n<p>Tober&#8217;s company Searchmetrics studied Google results to understand:<\/p>\n<ul>\n<li>What RankBrain does<\/li>\n<li>How RankBrain works<\/li>\n<li>What RankBrain means for websites and SEO<\/li>\n<\/ul>\n<p>He had two full-time people work for\u00a0three weeks to collect the RankBrain data he is presenting here today.<\/p>\n<p>Before we talk about details and key findings, it&#8217;s important to look at <strong>machine learning (ML)<\/strong> and <strong>artificial intelligence (AI)<\/strong>\u00a0first. So they dug through patents and papers.<\/p>\n<p><strong>ML is not AI<\/strong><\/p>\n<p>Machine learning is an algorithm that improves over time. Deep learning aims to bridge the gap between ML and AI and it solves more complex problems. Human-like intelligence (AI) is the end game.<\/p>\n<p>Examples of ML:<\/p>\n<ul>\n<li>The junk mail filter in your email inbox<\/li>\n<li>Photo recognition, like how Facebook recognizes you in a photo<\/li>\n<li>Recommendations in iTunes, Spotify, and Netflix<\/li>\n<\/ul>\n<p>The limits of machine learning are shown in the recent news that Google&#8217;s machine learning system beat a human in the game Go:<\/p>\n<ul>\n<li>More than 2,500 years old<\/li>\n<li>Played by 40 million people worldwide<\/li>\n<li>Is Go really that hard? Well, actually yes.<\/li>\n<li>There are 10^50 possible moves in a chess game.<\/li>\n<li>There are 10^171 possible moves in a game of Go.<\/li>\n<li>Google&#8217;s program AlphaGo beat the world champion 10\u20130.<\/li>\n<\/ul>\n<p>Deep learning and AlphaGo:<\/p>\n<ul>\n<li>Traditional A.I. methods, which analyze all possible positions, failed.<\/li>\n<li>AlphaGo uses deep neural networks across 12 different network layers.<\/li>\n<li>One neural network selects the next move to play. The other neural network predicts the winner of the game.<\/li>\n<li>AlphaGo has a lot to show us about RankBrain.<\/li>\n<\/ul>\n<p>There&#8217;s one project that sticks out in their research of Google patents, and that&#8217;s the work of Geoffrey Hinton:<\/p>\n<figure id=\"attachment_39991\" aria-describedby=\"caption-attachment-39991\" style=\"width: 580px\" class=\"wp-caption alignnone\"><a href=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton.jpg\" rel=\"attachment wp-att-39991\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-39991\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton.jpg\" alt=\"Slide picturing Geoffrey Hinton\" width=\"580\" height=\"359\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton.jpg 1510w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton-300x186.jpg 300w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton-768x476.jpg 768w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton-1024x634.jpg 1024w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-geoffrey-hinton-600x372.jpg 600w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/a><figcaption id=\"caption-attachment-39991\" class=\"wp-caption-text\">Slide about researcher Geoffrey Hinton (click to enlarge)<\/figcaption><\/figure>\n<p>Here&#8217;s Hinton&#8217;s\u00a0research in a simplified nutshell. Let&#8217;s talk about thought vectors. Imagine empty space. Every word in the world has a position and a proximity to other words. Google can map queries in search in that space so you get the proximity and distance of different queries.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-39992\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-visualizing.jpg\" alt=\"Visualizing RankBrain slide\" width=\"580\" height=\"441\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-visualizing.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-visualizing-300x228.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>When a query such as <em>&#8220;What&#8217;s the weather going to be like in California?&#8221;<\/em> is searched, RankBrain can interpret it as &#8220;Weather Forecast California.&#8221; Using training data, similar query sentences (with similar results) are closely positioned.<\/p>\n<p>Searches for &#8220;credit card&#8221; and &#8220;debit card&#8221; will be close to each other in space. Results scoring is where Google results rank better based on proximity. Good results, with higher relevance, have higher proximity to each other.<\/p>\n<h3>The Searchmetrics Study<\/h3>\n<p>Here was the Searchmetrics Hypothesis: Traditional ranking factors can no longer make sense for typical organic rankings. This isn&#8217;t for all queries. Google said that Google RankBrain is not used on every query. When it is used, it is the third most important ranking signal.<\/p>\n<ul>\n<li>RankBrain concentrates on relevant content.<\/li>\n<li>RankBrain uses thought vectors to map relevant results to queries.<\/li>\n<li>Relevance score, then order rankings in real time.<\/li>\n<\/ul>\n<p>Brief study background:<\/p>\n<ul>\n<li>Used top 30 results for Google U.S.<\/li>\n<li>Used 400K data points<\/li>\n<li>Used 3 different keyword sets with no overlap: loan, ecommerce, health<\/li>\n<li>The approach was to discover the ranking factors that were most important<\/li>\n<li>They tried to emulate RankBrain by adding scores to understand the content relevance<\/li>\n<\/ul>\n<p>In order to do this analysis, they had to remove all previous understanding of ranking factors. This includes backlinks and internal links, keyword in title, word count and interactive elements used. They cleared the slate of their presupposed ranking factors in order to get the new ranking signals.<\/p>\n<p><strong>Backlinks:<\/strong><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-39993\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-backlinks-slide.jpg\" alt=\"Backlinks chart for Tober's study\" width=\"580\" height=\"345\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-backlinks-slide.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-backlinks-slide-300x178.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>We see that for ecommerce keywords, there is a\u00a0positive correlation between rankings and backlinks. However, for health and loan queries, rankings and backlinks have a negative correlation.<\/p>\n<p><strong>Internal links:<\/strong><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-39994\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-inbound-links-slide.jpg\" alt=\"Internal links chart\" width=\"580\" height=\"346\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-inbound-links-slide.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-inbound-links-slide-300x179.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>For loan queries, the average number of internal links is less than 100. For ecommerce, the average number of internal links is much higher.<\/p>\n<p><strong>Keyword in Title:<\/strong><\/p>\n<p>Here&#8217;s a traditional <a href=\"https:\/\/www.bruceclay.com\/seo\/search-engine-optimization\/\">SEO<\/a> optimization tactic. In this analysis, they accounted for stemming and variations.<\/p>\n<p>In the loan category, only 10 percent of pages have the keyword in the title tag.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-39996\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-keyword-in-title.jpg\" alt=\"Keyword in title chart\" width=\"580\" height=\"324\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-keyword-in-title.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-keyword-in-title-300x168.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p><strong>Word count and interactive elements:<\/strong><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-39997\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-word-count.jpg\" alt=\"Word count chart\" width=\"580\" height=\"339\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-word-count.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-word-count-300x175.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>For e-commerce and health, pages that are more successful have more content and more interactive elements.<\/p>\n<p>Why do traditional ranking factors fail to explain these examples? If you look behind the curtains, you see that Google has changed a lot, with Hummingbird adding in machine learning.<\/p>\n<p><strong>Results of Searchmetric&#8217;s study<\/strong><\/p>\n<p>They emulated RankBrain and gave search results a relevance score. This score is based on how relevant a result is to a query. They found\u00a0around <strong>25 relevance ranking factors to assess relevance<\/strong>.<\/p>\n<p>Relevance factors are different for each keyword, so Tober says they can&#8217;t provide a table of all the relevance factors.<\/p>\n<p>For example, 9 out of 10 ecommerce websites have a keyword &#8220;add to cart&#8221; function above the fold. Rank #9 does not. However, it has the highest relevance score of the top 30 and that&#8217;s why it ranks.<\/p>\n<p>Another example: If you search for &#8220;best bluetooth headphones&#8221; then you&#8217;ll see:<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-40000\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-relevance-example.jpg\" alt=\"Relevance example in search results\" width=\"580\" height=\"333\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-relevance-example.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-relevance-example-300x172.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>See how many internal links for each result and compare their relevance scores.<\/p>\n<p>And another example, for the keyword &#8220;natural detox&#8221; you&#8217;ll see the word count comparison for two results, with the higher ranking result having fewer words, fewer internal links and fewer interactive elements.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-39999\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-relevance-example-2.jpg\" alt=\"Relevance example 2\" width=\"580\" height=\"341\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-relevance-example-2.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-relevance-example-2-300x176.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p><strong>Content with a high relevance score matches user intention, is\u00a0logically\u00a0structured and\u00a0comprehensive, offers\u00a0a good\u00a0user experience, and deals with topics holistically.<\/strong> Holistic means that other topics related to the topic are covered on this page.<\/p>\n<p><strong>Key Findings<\/strong><\/p>\n<ul>\n<li>Top ranking factors are different depending on keyword set.<\/li>\n<li>Relevance ranking factors dominate results across all keyword sets.<\/li>\n<li>All previous examples can be explained by having a higher relevance score.<\/li>\n<li>This score overpowered other ranking factors, meaning that these pages ranked highly.<\/li>\n<\/ul>\n<p><strong>Outlook for SEO<\/strong><\/p>\n<ul>\n<li>SEO is as important as ever, but it&#8217;s changing.<\/li>\n<li>RankBrain is not used on all queries. For example, short\/popular queries with well-known results aren&#8217;t filtered by RankBrain if your content matches user\/query intent.<\/li>\n<li>Relevance is crucial for good rankings, because RankBrain can detect how relevant your content is.<\/li>\n<li>Make sure your content matches user\/query intent.<\/li>\n<\/ul>\n<p>When Google already knows the best results, RankBrain is not used. He believes RankBrain is filtering long-tail queries. Relevance is crucial for good rankings and RankBrain can detect how relevant our content is.<\/p>\n<p><strong>The Future of Search<\/strong><\/p>\n<p>Tober says we&#8217;re in for an abundance of redundancy. He explains that machine learning and Searchmetrics share this philosophy, which he says applies for content creation, too: <\/p>\n<ul>\n<li>Incremental improvements through powerful data insights<\/li>\n<li>Using machine and deep learning to make sense of complex data<\/li>\n<li>A data-driven approach is necessary to make sense of all the redundancy online<\/li>\n<\/ul>\n<h2>Eric Enge: What Is RankBrain and How Does it Work<\/h2>\n<p>Eric had brain surgery in 2003. Does that make him a RankBrain expert? No, but his company Stone Temple Consulting did a study of Google results\u00a0before and after RankBrain to see\u00a0what&#8217;s changed.<\/p>\n<p>Notable quote from the Bloomberg article announcement:<\/p>\n<p><em>&#8220;RankBrain interprets language, interprets your queries, in a way that has some of the gut feeling and guessability of people.&#8221;<\/em><\/p>\n<p>Google is trying to understand the true meaning of queries and do a better job of providing results that are relevant.<\/p>\n<p>Some basic language analysis concepts figure in here. Take, for example, <strong>stop words<\/strong>. Google has traditionally stripped stop words out of a query or indexed page to simplify the language analysis.<\/p>\n<p>But there are places where this practice doesn&#8217;t work. Take the query <em>&#8220;the office&#8221;<\/em> as an example. Someone wants the TV show, but Google may give results for your office. A more sophisticated language analysis might look for those two words capitalized in the middle of a sentence as a cue that this is a specific kind of &#8220;the office.&#8221;<\/p>\n<p>Another example is <em>&#8220;coach,&#8221;<\/em> which sometimes refers to a brand. Google may have had to manually patch the results if they find poor engagement on SERPs for a query like this. Here&#8217;s a way as humans we might find that the Coach brand is being referenced, which Google ultimately wants to be able to recognize algorithmically:<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-40003\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-coach-query.jpg\" alt=\"Query example for &quot;coach&quot;\" width=\"580\" height=\"309\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-coach-query.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-coach-query-300x160.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>RankBrain is trying to understand different kinds of relationships in language. On the topic of RankBrain, Gary Illyes told Enge it had to do with &#8220;being able to represent strings of text in very high-dimensional space and &#8216;see&#8217; how they relate to one another.&#8221;<\/p>\n<p>RankBrain gets at the context, what other words are used in the same page, same paragraph, same sentence, etc. It can notice patterns in language usage:<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-40007\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-notices-patterns.jpg\" alt=\"RankBrain notices patterns in language usage slide\" width=\"580\" height=\"308\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-notices-patterns.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-notices-patterns-300x159.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>Along with the example of the stop word &#8220;the&#8221; affecting the query, Enge also shows how certain words were hard for Google before RankBrain:<\/p>\n<ul>\n<li>&#8220;The&#8221; \u2014 classic stop word, but in some cases critical to phrase meaning (e.g., &#8220;The Office&#8221;)<\/li>\n<li>&#8220;Graphic&#8221; \u2014 could relate to violence, language, or design (meaning determined by context)<\/li>\n<li>&#8220;Without&#8221; \u2014 negatives traditionally ignored by Google, but in some cases critical to meaning of page<\/li>\n<\/ul>\n<p>Example queries:<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-40009\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-example-query.jpg\" alt=\"Example search query\" width=\"580\" height=\"248\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-example-query.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-example-query-300x128.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>This is the example offered the the Bloomberg article.<\/p>\n<p>Another troublesome example query that Gary Illyes shared in an interview with Enge: <em>&#8220;Can you get 100% score on Super Mario without walkthrough?&#8221;<\/em><\/p>\n<p>The word &#8220;without&#8221; was traditionally taken out of a query, so now the user isn&#8217;t getting the answer they were looking for. This is an example of how Google was able to improve the results with RankBrain.<\/p>\n<h3>Stone Temple Consulting&#8217;s Study<\/h3>\n<p>From their database of 500K queries, they looked for examples of queries that in June\/July 205 Google didn&#8217;t understand and compared it to January 2016 results.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-40012\" src=\"https:\/\/www.bruceclay.com\/blog\/wp-content\/uploads\/2016\/03\/2-RankBrain-SERP-before-after.jpg\" alt=\"Search results before and after RankBrain\" width=\"580\" height=\"337\" srcset=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-SERP-before-after.jpg 580w, https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-SERP-before-after-300x174.jpg 300w\" sizes=\"(max-width: 580px) 100vw, 580px\" \/><\/p>\n<p>In the left example, you see that Google doesn&#8217;t answer the query about Ragnarok, and in the right we see that Google did answer the question.<\/p>\n<p>To find queries that Google misunderstood, they removed a number of queries that either were not clear what the searcher meant, or that didn&#8217;t make sense in other ways. Here&#8217;s their end aggregate analysis:<\/p>\n<ul>\n<li>Total Queries We Found: 163<\/li>\n<li>Results Improved: 89<\/li>\n<li>Results Improved (%): 54.6 percent<\/li>\n<\/ul>\n<p>How did the results improve? The 89 &#8220;better results&#8221; the study found were improved in three basic ways:<\/p>\n<ul>\n<li>39 results (43.8 percent) had an improved answer box.<\/li>\n<li>2 results (2.2 percent) added maps.<\/li>\n<li>48 results (53.9 percent) had improved search results pages.<\/li>\n<\/ul>\n<p>RankBrain did a good job of parsing the query and instructing Google&#8217;s back-end retrieval systems to get the right results. Enge believes that Google\u00a0can pass a better query not just to web results but also to featured snippets and other algorithms.<\/p>\n<p>Enge shows a January 2016 SERP result of Abz Love in Wikipedia for the query <em>&#8220;who is abs.&#8221;<\/em> He explains that Google now figures out to put weighted importance on the &#8220;who is&#8221; portion of the query, even though &#8220;abs&#8221; changed the way he spells his name.<\/p>\n<p>Types of improvements they found included:<\/p>\n<ul>\n<li>Phrase interpretation improved (e.g., &#8220;YouTube Plus&#8221;)<\/li>\n<li>Word interpretation improved (e.g., &#8220;weak&#8221;)<\/li>\n<li>Good user query (e.g., &#8220;convert glory into an adjective&#8221;)<\/li>\n<li>Poor user query (e.g., &#8220;convert into million dollars&#8221;)<\/li>\n<\/ul>\n<p>Phrase interpretation and word interpretation improved. About two-thirds of the time, the user query was pretty clear to a human and Google didn&#8217;t understand, while the other third of the time, the query wasn&#8217;t clear originally.<\/p>\n<p><strong>Impact on SEO<\/strong><\/p>\n<p>There&#8217;s a growing importance on relevance and context in ranking. We&#8217;re seeing a shift in that direction, although he&#8217;s not sure that all of that is due to RankBrain. He doesn&#8217;t see a lot of direct impact on SEO from RankBrain right now. The biggest change happening in SEO is with content and comprehensiveness and relevance.<\/p>\n<p>Summary of key SEO implications:<\/p>\n<ul>\n<li>More obscure phrases may bring up your page.<\/li>\n<li>Yes, you still need to do keyword research.<\/li>\n<li>You should\u00a0increase your emphasis on truly natural language.<\/li>\n<\/ul>\n<p>Danny Sullivan says his takeaway is, what are you really going to do any differently? Create great content for humans, which Google will reward like they&#8217;ve always said they do.<\/p>\n<p><em>For more information regarding Google&#8217;s RankBrain, read our post on <a href=\"https:\/\/www.bruceclay.com\/blog\/rankbrain-real-seo-impact\/\">The REAL Impact of RankBrain on Web Traffic<\/a>.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>SEO is very tactical. We always try to look behind the curtain of Google&#8217;s algorithms. So, it&#8217;s no surprise that we all want to know more about RankBrain, Google&#8217;s machine learning system that they confirmed out of the blue in October 2015 was &#8220;the third most important ranking signal&#8221; in their algorithm.<\/p>\n<p>In this SMX West session, we&#8217;ll learn about RankBrain based on studies done by the speakers: Marcus Tober of Searchmetrics and Eric Enge of Stone Temple Consulting. They&#8217;ll present findings as well as the implications for SEO.<\/p>\n<p><a href=\"http:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/\">Read the liveblog to find out about RankBrain<\/a>.<\/p>\n","protected":false},"author":38,"featured_media":39990,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"content-type":"","footnotes":""},"categories":[4],"tags":[18,39,1232,1313],"acf":[],"yoast_head":"<title>BruceClay - What We Know About Google&#039;s Machine-Learning System RankBrain - #SMX Liveblog<\/title>\n<meta name=\"description\" content=\"Speakers Eric Enge &amp; Marcus Tober sum up what they learned about RankBrain, Google&#039;s machine learning algorithm, and its implications for SEO.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"What We Know About Google&#039;s Machine-Learning System RankBrain - #SMX Liveblog\" \/>\n<meta property=\"og:description\" content=\"Speakers Eric Enge &amp; Marcus Tober sum up what they learned about RankBrain, Google&#039;s machine learning algorithm, and its implications for SEO.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/\" \/>\n<meta property=\"og:site_name\" content=\"Bruce Clay, Inc.\" \/>\n<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/virginianussey\" \/>\n<meta property=\"article:published_time\" content=\"2016-03-02T23:00:06+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2019-08-08T18:06:43+00:00\" \/>\n<meta name=\"author\" content=\"Virginia Nussey\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-speakers.jpg\" \/>\n<meta name=\"twitter:creator\" content=\"@virginianussey\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Virginia Nussey\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"11 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/\",\"url\":\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/\",\"name\":\"What We Know About Google's Machine-Learning System RankBrain - #SMX Liveblog\",\"isPartOf\":{\"@id\":\"https:\/\/www.bruceclay.com\/#website\"},\"datePublished\":\"2016-03-02T23:00:06+00:00\",\"dateModified\":\"2019-08-08T18:06:43+00:00\",\"author\":{\"@id\":\"https:\/\/www.bruceclay.com\/#\/schema\/person\/f7873389c9326ba658076ad70e14fda1\"},\"description\":\"Speakers Eric Enge & Marcus Tober sum up what they learned about RankBrain, Google's machine learning algorithm, and its implications for SEO.\",\"breadcrumb\":{\"@id\":\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.bruceclay.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"RankBrain: What Do We Know About Google&#8217;s Machine-Learning System? #SMX\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.bruceclay.com\/#website\",\"url\":\"https:\/\/www.bruceclay.com\/\",\"name\":\"Bruce Clay, Inc.\",\"description\":\"SEO and Internet Marketing\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.bruceclay.com\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.bruceclay.com\/#\/schema\/person\/f7873389c9326ba658076ad70e14fda1\",\"name\":\"Virginia Nussey\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.bruceclay.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/af79aee22cb093734144d2fea6700a9e?s=96&d=retro&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/af79aee22cb093734144d2fea6700a9e?s=96&d=retro&r=g\",\"caption\":\"Virginia Nussey\"},\"description\":\"Virginia Nussey is the director of content marketing at MobileMonkey. Prior to joining this startup in 2018, Virginia was the operations and content manager at Bruce Clay Inc., having joined the company in 2008 as a writer and blogger.\",\"sameAs\":[\"https:\/\/www.facebook.com\/virginianussey\",\"https:\/\/twitter.com\/virginianussey\"],\"url\":\"https:\/\/www.bruceclay.com\/blog\/author\/virginia-nussey\/\"}]}<\/script>","yoast_head_json":{"title":"BruceClay - What We Know About Google's Machine-Learning System RankBrain - #SMX Liveblog","description":"Speakers Eric Enge & Marcus Tober sum up what they learned about RankBrain, Google's machine learning algorithm, and its implications for SEO.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/","og_locale":"en_US","og_type":"article","og_title":"What We Know About Google's Machine-Learning System RankBrain - #SMX Liveblog","og_description":"Speakers Eric Enge & Marcus Tober sum up what they learned about RankBrain, Google's machine learning algorithm, and its implications for SEO.","og_url":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/","og_site_name":"Bruce Clay, Inc.","article_author":"https:\/\/www.facebook.com\/virginianussey","article_published_time":"2016-03-02T23:00:06+00:00","article_modified_time":"2019-08-08T18:06:43+00:00","author":"Virginia Nussey","twitter_card":"summary_large_image","twitter_image":"https:\/\/www.bruceclay.com\/wp-content\/uploads\/2016\/03\/2-RankBrain-speakers.jpg","twitter_creator":"@virginianussey","twitter_misc":{"Written by":"Virginia Nussey","Est. reading time":"11 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/","url":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/","name":"What We Know About Google's Machine-Learning System RankBrain - #SMX Liveblog","isPartOf":{"@id":"https:\/\/www.bruceclay.com\/#website"},"datePublished":"2016-03-02T23:00:06+00:00","dateModified":"2019-08-08T18:06:43+00:00","author":{"@id":"https:\/\/www.bruceclay.com\/#\/schema\/person\/f7873389c9326ba658076ad70e14fda1"},"description":"Speakers Eric Enge & Marcus Tober sum up what they learned about RankBrain, Google's machine learning algorithm, and its implications for SEO.","breadcrumb":{"@id":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.bruceclay.com\/blog\/rankbrain-google-machine-learning-system-smx\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.bruceclay.com\/"},{"@type":"ListItem","position":2,"name":"RankBrain: What Do We Know About Google&#8217;s Machine-Learning System? #SMX"}]},{"@type":"WebSite","@id":"https:\/\/www.bruceclay.com\/#website","url":"https:\/\/www.bruceclay.com\/","name":"Bruce Clay, Inc.","description":"SEO and Internet Marketing","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.bruceclay.com\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/www.bruceclay.com\/#\/schema\/person\/f7873389c9326ba658076ad70e14fda1","name":"Virginia Nussey","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.bruceclay.com\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/af79aee22cb093734144d2fea6700a9e?s=96&d=retro&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/af79aee22cb093734144d2fea6700a9e?s=96&d=retro&r=g","caption":"Virginia Nussey"},"description":"Virginia Nussey is the director of content marketing at MobileMonkey. Prior to joining this startup in 2018, Virginia was the operations and content manager at Bruce Clay Inc., having joined the company in 2008 as a writer and blogger.","sameAs":["https:\/\/www.facebook.com\/virginianussey","https:\/\/twitter.com\/virginianussey"],"url":"https:\/\/www.bruceclay.com\/blog\/author\/virginia-nussey\/"}]}},"_links":{"self":[{"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/posts\/39716"}],"collection":[{"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/users\/38"}],"replies":[{"embeddable":true,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/comments?post=39716"}],"version-history":[{"count":16,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/posts\/39716\/revisions"}],"predecessor-version":[{"id":71230,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/posts\/39716\/revisions\/71230"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/media\/39990"}],"wp:attachment":[{"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/media?parent=39716"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/categories?post=39716"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.bruceclay.com\/wp-json\/wp\/v2\/tags?post=39716"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}