Meet John Giannandrea New Head of Google Search Machine Learning
After 15 years as Google’s search chief, Amit Singhal announced yesterday on Google+ that he is moving on. After he stated, “Feb 26 will be my last day at Google”, it was quickly announced that John Giannandrea would take up the helm of Google Search.
We can all agree that being with a Google that long and contributing so much to search is a remarkable accomplishment and congratulate Singhal as he steps into a new time in life, focusing on philanthropy. As new leadership often means momentous refocusing, SEO professionals wonder how earned search may change as Giannandrea assumes this position, and if the change will generate ripples across the tech world as a whole. The future of how GoogleBot crawls and interprets web content looks promising under his leadership, as we observe how he impacts machine learning’s future on news and how the Metaweb is woven.
Amit went on to say that “search is stronger than ever, and will only get better in the hands of an outstanding set of senior leaders who are already running the show day-to-day. Our mission of empowering people with information and the impact it has had on this world cannot be overstated.” John Giannandrea, who has been the forerunner overseeing artificial intelligence, such as in Google Algorithm RankBrain, has been employed at Google for six years and is currently the VP of engineering.
As explained by Forbes in November, 2015 RankBrain’s role took “a very large fraction” of the millions of queries that went through the search engine. His team evaluates forms of machine learning inside Google Search. Formerly, Google depended heavily on algorithms that adhered to a strict set of search rules led by humans.
The new position as head of Google Search for Giannandrea denotes a further emphasis on the importance of artificial intelligence to Google, a unit of Alphabet Inc (AI). His foresight in technology are key to the latest efforts in search on mobile devices and searches that apply speech recognition. Machine learning is a different approach from writing a detailed computer program to take on an undertaking; rather you give a computer multiple examples of a job and attempt to have it learn automatically.
Search is the Foundation of Google Machine learning
Search is the foundation of Google. Google serves over 100 billion searches each month. Machine learning isn’t new or some far out algorithm attempting to band-aide short-comings in search; it certainly makes it easier to manage. Machine learning is proving to make digital marketing evolve faster and make the world of semantic search better. It requires a lot of thought, research, testing, and care in order to build something that can revolutionize the world of semantic search like this.
Search and machine learning is expected to handle search queries better than algorithmic rules hand-coded by human engineers. According to search experts at Moz, “machine learning is already being used by many major platforms.” So if artificial intelligence is the future of Google Search, it may be that we are on the tip of so much more in the world of SEO.
Google is merging its research efforts in all from of search to align with priorities of machine learning inside the company. Given his search experience, John Giannandrea, the current engineering VP who leads Google’s sprawling research and artificial intelligence wing, is the one most likely to succeed in the critical alphabet and the algorithm changes ahead. Google search improvements overseen by Singhal made the giant web-search tool faster, smarter and able to peer into the entrails of software running on mobile devices. Giannandreas is poised to be the best individual to carry this cornerstone of Google Search forward.
Machine learning is currently much more advanced within Google’s focus on search and user intent, which is undoubtedly set to continue. Due to the sophistication of search, SEO professionals don’t anticipate any overnight revolution when it comes to Google search. We are much embedded in the granular details and time of continuing our passions to win increased search visibility, which yields new web visits, better user experiences, and thereby, chances for higher revenue. Given all the significance Google has allotted to RankBrain, the details demonstrating how it really works in influencing Google ranking chances are still forthcoming. Google has currently offered two or three actual examples about the RankBrain changes in broad generalities.
Singhal: Overseer of Google’s Search Algorithm
Singhal has been hugely entwined with Google search in his duration at Google so far. He has been known to participate when guidance was the need for rewriting the algorithms created by Google co-founders Larry Page and Sergey Brin. Tasks involved in deciphering what Google would display in search engine result pages (SERPs) in response to searches are pivotal to the core of the search giant.
Machine learning is rapidly changing the landscape of what SEO is. John Giannandrea told reporters at Google headquarters last fall. “Increasingly, we’re discovering that if we can learn things rather than writing code, we can scale these things much better.” However, machine learning is not new. In 1957 American mathematician and researcher in artificial intelligence, Ray Solomon, published “An Inductive Inference Machine”, according to historyofinformation.com***. IRE Convention Record, Section on Information Theory, Part 2 (1957) 56-62. This was the first paper written on machine learning.
Singhal has never veered from this helm. His work continues to guiding Google search through transitions in the expansion of vertical search, seeking to integrate social signals, Google’s direct answers as search becomes more semantic, and perhaps most important, making Google successful in the mobile search world.
Machine learning is more than SEO focused on just getting the click. Today’s new Google search engines are tabulating how people are interacting with your website: Are they hitting the back button and clicking on other sites, or are they satisfied with the answers they’re searching for when they’re on your site? Machine learning takes search to a place that is more about the post-click activity. Not only do you have to win the click after showing up in SERPS, but you have to satisfy user with a great experience on your site.
Google Search Shifts to Machine Learning to Better Evaluate Customer Experiences
Is there a huge shift at Google in Search? Customer experience remains a top priority, and the change at Google’s helm may influence Marketers’ strategic priorities to a more centered approach on structured data and machine learning. Webmasters and SEOs can optimize web content that goes on each page of a website to facilitate user engagement, but it is complete up the site visitor how they consume and react to that content. Heat maps, like Crazy Egg are tremendously helpful to reveal the hot spots that users prefer on your site. While its seems that a searcher can ask Google anything and get an answer, knowing if they are satisfied with the answer when landing on your website is a tougher thing to determine.
We can expect to see deep learning or machine learning a bigger part of Google Search going forward. Given how complex algorithms are, the growth of machine learning will be just one component of how Google Search works. Basically, the work of John Giannandrea and his team have advanced Google’s machine learning engine to be more adept at analyzing the words and phrases that are part of a search query. By being more intuitive at deciding what related words and phrases carry much the same meaning, the new Google Search improves on the old rules-based system when handling brand new queries. Schema mark-up aids in the process of making it easier to understand web page content and match it to those queries. There are many new queries Google Searches every day – ones that it has never seen before.
It is 12:58 pm as I write and the number of queries Google Searches has searched for this far today is 2,464,000,000 and climbing fast. Google now processes over 42,000 search queries on average every second which ends up being over 3.5 billion searches per day and 1.2 trillion searches annually worldwide.
With his expertise in artificial intelligence, as John Giannandrea takes responsibility for Google’s search algorithms, it makes since that Google is acknowledging the future of Google Search is shifting to a great reliance on machine learning. Managing so many new searches is a huge task. “According to a 2013 comScore public release, as of December 2012, Google enjoyed a 65.2% share of web search volume worldwide, with 114.7 billion searches that month”, states www.internetlivestats.com*
Combine How Machines and Users Rate your Content
How John Giannandrea Influenced RankBrain for Google Search
Singha was named a Google Fellow, which is the highest honor Google bestows on its search engineers. During his duration at Google, he has masterminded the company’s search engine, and, as we all recognize, the Google search engine has fairly ruled the Internet
From title tags to description Meta tags and structured data, they each help give a brief overview of your website’s content. When optimized correctly, such data may help pages be introduced in SERRPs with a rich snippet. A user performs the query, and even a deep page, with its unique content may show up more often in machine learning.
John Giannandrea’s efforts at Google to date have done much to advance the search engines summary of what the page is about. Already in the first quarter of 2016, Google Webmaster Tools have broadened, providing a rich content analysis section to help SEO professionals to improve their HTML documents.
At Google, artificial intelligence sits at the far end of machine learning. Google has invested heavily in machine learning under John Giannandreas: this includes Google Search for videos, speech, translation, and, recently, search. Millions of search queries every second that individuals type into a Google search box are interpreted by the artificial intelligence system nicknamed RankBrain. Greg Corrado, a senior Google research scientist says, “RankBrain uses artificial intelligence to embed vast amounts of written language into mathematical entities — called vectors — that the computer can understand”. When RankBrain sees a new word or phrase, the machine guesses as to what words or phrases may have a similar meaning to answer the search query.
According to Bloomberg news, additional search work he has overseen includes advancements in image recognition and technologies that procure information based on what users are doing with their devices, rather than the explicit search terms used.
QUESTION: What is machine learning?
ANSWER: “Machine learning is the subfield of computer science that gives computers the ability to learn without being explicitly programmed and evolved from the study of pattern recognition and computational learning theory in artificial intelligence, machine learning explores the study and construction of algorithms that can learn from and make predictions on data – such algorithms overcome following strictly static program instructions by making data-driven predictions or decisions,through building a model from sample inputs”, according to Wikipedia the explicit search terms used.
SEO Experts Adhering to Machine Learning
Advanced SEO experts understand the linear algebra behind machine learning to some degree. The results it produces can retrieve very accurate results to meet user intent. It is much harder to explain what led them to those accurate results and where Google Search is headed. Alphabet initially stole the top market value position from Apple mid-Monday leading up to its fourth-quarter earnings report. Google Search is interpreted by some as gaining supremacy around Alphabet’s main Google unit.
It is assumed that John Giannandrea will continue to spearhead tweaking the behavior of neural nets to adjust the math through intuition, further research, trial, and error. You must retrain them on new data, with still more trial and error. As Google moves search to this complicated machine learning model, it’s unclear all that is ahead in search. Human brains remain the guides behind these machines, but just how has changed. Just yesterday, Google added more extensive JSON-LD support For structured data markup; it may be a given that coincides perfectly with the updates in Google search machine learning.
If your site has thousands of web pages, hand-crafting additional structured data probably isn’t feasible. We suggest starting with your core service pages and key pages that Google Analytics shows already are most preferred by viewers.
We are learning more as to how Google might use machine learning to answer questions that are served up in the Google Quick Answer Box. The artificial intelligence aspect offers intelligent machines to reduce the blurs some search engines struggle with when a puree is all that could otherwise be offered to match user intent. No matter how much we’d like everything to be standardized, in the search industry results are not based directly on exact rules. It relies heavily on the work of semantic language, speech recognition, translation and visual processing that continues to ignite deep scientific and engineering challenges.
Don’t Confuse User Engagement with User Experience
Michael Martinez, owner of seo.theory.com states it this way; “People continue to confuse ‘user engagement’ with ‘user experience’. User experience is what you put on the page. User engagement is what the visitor does on the page”. Google cannot track what these people do on the vast majority of Web pages when they exist without Google Analytics in place.
If you analyze you incoming traffic through Google Analytics, Google search is one of many ways visitors reach your website. When you track “average page views per visit” to learn how users travel through your digital conversion funnel from page to page, once that is over two pages, your content is engaging users not Google search. Predictive data that businesses leverage needs to be more than the fractional amount of data previously extrapolated by Google Bot.
Google, like other search engines, gauges user satisfaction within its own search results. Understanding the scope of user satisfaction with your Web messaging extends to the full digital arena, including all of your social networks, news publications, Buzzsumo and more. Knowing where search is headed is critical for content marketing and SEO campaign planning.
What Does Your Business Need As Google Search Shifts?
A high number of the request that come to us reveal how often businesses find it confusing where to go next regarding search and digital marketing. Because SEO is growing up so quickly, whether a business is just starting out or looking for what they need to do next to gain visibility online, they are seeking guidance from an SEO expert who stays current in earned and paid search.
Marketers in every digital space are embracing the evolving customer experience reality. Taking the new Google Search in stride inspires our approach to SEO strategies, creativity, enthusiasm, and ultimately, our client’s success.
The tactics for successful SEO have shifted seismically in recent months. Under the new leadership of John Giannandrea, the current state of SEO will live evolve. His deep experience in search and Google reads content in 2016 means search marketers must put new time and effort into learning the changing landscape of Google Search. Previous strategies of driving traffic to your site with optimized headlines and carefully chosen keywords now need assessing to evolve into new SEO strategies that include the Google mobile algorithm. As a new year, and new generation of search engine optimization leadership emerges at Google, we trust this report is useful to you, and perhaps even ignites a greater focus on marketing ideas that work for your company.
Many using the Internet are not involved in on-going aspects surround cloud training. It’s clearly a good thing to have heavy-weight influencers like John Giannandrea at the helm of Google Search to make algorithms that understand language and images in a way that is a bit less computer-like and more human-like in their interactions and interpretations Having the experience of leading many cloud-trained staff at Google, the future of search may become more influenced by machine learning than ever under this new leadership.
Combine How Machines and Users Rate your Content
It is important to understand both how advancements in Google search machine learning and how web page visitors view your content.
What machine learning will mean to SEO professionals includes identifying what’s a good match when humans write web copy and hopes that prospective buyers searching for their products can find their corresponding pages. They write code designed to help the machines themselves learn what that page is about so it can reward the one who searched for relevant content.
Previously, Google’s algorithm sought to figure out what consumers think would be good answers to search queries. Once a viewer has landed on your website, if they follow links from one web page to another page, that is a human, not a machine, who finds your pages of value. Structured data mark-up is code written to make your content easier for machines to decipher. The use of text on a page, which keyword is used, how often and where, compelling imagery, and the user’s benefits, are additional items that can reflect quality as perceived by humans. Danny Sullivan of Search Engine Land commented, “To greatly simplify, it’s (machine learning Search) like teaching the search engine to paint by numbers, rather than teaching it how to be a great artist on its own.”
The contribution of the RankBrain algorithm to search is part of a half-decade-long push by Google into machine intelligence, as the company seeks to embed the technology deeper into its search core. Machine learning is a core transformative way that Google is using to re-structure everything search related. John Giannandrea has earned respect among many for his extensive technical background in search for his extensive knowledge in across every aspect of search. Google’s search is central to SEO professionals as we anticipate in his new role, he will continue to develop his team and advance Google Search.
How Machine Learning Works
Fundamentally, machine learning works in a two-step process.
Step 1 is an assessment process that is carried out on a set of data using a Google algorithm to accomplish a task is the output. When RankBrain is used, the algorithm created is one that gets better natural language parsing proficiencies of Google.
Step 2 in machine learning takes the acquired algorithm and applies it to new data sets. Again when RankBrain is used, this algorithm processes the user search queries received to help better ascertain the content and objective of those queries, and then distinguish documents pertinent to that query.
Google Search machine learning relates to the Hummingbird algorithm for is just like a plane or tractor has an engine that powers it. The engine itself is a complex unit consisting of multiple various parts, each that relate to the other. Likewise, Hummingbird encompasses various parts, with RankBrain being one of the newest.
In particular, we know RankBrain’s machine learning is a component of the overall Hummingbird algorithm and is increasingly managing a higher percentage of all Google searches.
Hummingbird’s algorithm is frequently spoken of with related names familiar to those in the SEO industry. These include, such terms as Google Penguin, Panda, Pigeon, Payday, Pirate, and Top Heavy, which was intended to demote ad-heavy pages in SERPs. Machine learning is core to the Mobile-First algorithm, which brings nice rewards to mobile-friendly pages, especially Accelerated Mobile Pages (AMP).
What RankBrain Machine Learning Does and Doesn’t Do
Stone Temple conducted a deep dive study into RankBrain and concluded that it “does NOT change the way Google interprets links, content, other aspects of relevance, spam or any of the other algorithms”. It is helpful to decipher both what RankBrain does and doesn’t do. Foremost, their conclusion is that it doesn’t vary these algorithms at all. Based on information available today, our understanding is that RankBrain doesn’t classify web pages. Its classification is around query analysis.
What it does do is a better job on query interpretation, and comprehending page content, and because of that, RankBrain machine learning extensively improves relevance matching. The result of AI learning human behavior better is that the way it puzzles out user intent leads to many shifts in SEO results. We are all attempting to learn how machine learning affects the evolution of search and what that means to the future of SEO.
Modern search engineering is a complex semantic analysis of keywords and much more; including HTMM (Hidden Topic Markov Models) and PLSA. Marketers that strive to better understand what the user is looking will find their content more aligned with advances in how search works.
By understanding the main concern RankBrain is trying to solve, we can rule out the more outlandish speculations of what the AI system does. Any SEO professional who strives to understand semantic search, machine learning and case-based structures will find themselves in the midst of the exciting, high-demand field of semantic SEO. Combine that with the ability to read your AdWords and Analytics Reports, and your business can benefit from predictions drawn from your data, to create and adapt your web content to improve over time. We closely follow and work to implement new search strategies that incorporate what we know about the latest machine learning algorithms or applications that rely on Google’s Mobile Search Algorithm.
Why Was Machine Learning Created?
In order to further understand how Google Search is advancing, it helps to consider the fixes that the processes are designed to help remedy:
• Handling complex searches and the growing number of new searches
Google daily faces lengthy and multifaceted new searches (15% of all search queries) that never before were entered into the search box. Since there is no previous history to easily manage such searches, Google must determine how to relate them to previous search user interests in order to display the most relevant results.
• Managing vague search queries
Internet users often use ambiguous word formats or spoken language in voice search that can mean more than one thing based on varied context. Having a machine that knows how to process existing search histories empowers Google to better guesstimate better the user’s intent.
• Refining correctness and accuracy
Advancing from former keyword matching tactics to context understanding permits improved accuracy when dealing with variants of the same keyword (singular vs plural, synonyms, acronyms, abbreviations, misspellings etc)
Machine learning handles search queries in a more natural way, and voice search is on the rise. Voice search technologies like Siri, Google Now and Cortana are enjoying being more widely accepted. Wherever its intent was when it originally started RankBrain’s machine learning plays a central role in helping voice search assistants come up with improved accuracy when providing results for voice searches.
“That’s probably a fair assessment, too. Machine learning has already grown much within Google and is clearly set to continue. However, it’s not likely to be an overnight revolution, when it comes to Google search. Instead, it’s more likely to be the tired but true “evolution, not revolution” unfolding that we’ll see.” – Danny Sullivan
Google’s decision to deploy AI into search shows that companies are starting to entrust their most valuable businesses to systems controlled in part by machine intelligence. Facebook Inc. uses AI techniques to filter the newsfeed that comprises the personalized homepage of the social network and Microsoft Corp. is using artificial intelligence to increase the capabilities of its Bing search engine.” – Bloomberg News**
“The fact that they got the head of artificial intelligence to take over the role is very telling of where they expect to evolve going forward.” – Sameet Sinha, senior equity analyst at B. Riley & Co.
And as usual, nothing is certain especially in fast-evolving industries like SEO. For now machine learning is something we are all striving to grasp. In the future, many expect that RankBrain will be used in ever-widening applications – well beyond that of only interpreting search queries. Should it advance sufficiently to actually comprehend all forms of online contents, then it will be a giant innovation in the industry with ripples farther reaching than we can foresee today.
To benefit from guidance to formulate a custom marketing plan for your business based on our tested knowledge of the SEO best tactics and strategies, call us at 651-206-2410.
We offer services in both paid and earned search for businesses of all sizes and types. Gain the benefits of staying of top SEO strategies as Google search shifts to machine learning. Ready for fresh ideas on how to improve your website? Contact us today to leverage SEO Tactics that fit Google’s Algorithm Update