Google has done it again. The Internet giant has introduced a new algorithm known as RankBrain that determines search results. RankBrain is an artificial technology that enables Google to process and sort through its search results in unprecedented ways. Anyone who's serious about their search results will need to know the following.
RankBrain – The New Machine Learning Artificial Intelligence Technology
Machine learning enables computers to learn to perform functions on their own without having to follow programming codes or human orders. On the other hand, artificial intelligence enables electronics and computers to become as smart as humans in the acquisition of self-taught and programmed knowledge. Google's RankBrain harnesses both of the above terms almost synonymously—its technology is frequently referred to as machine-learning artificial intelligence.
RankBrain is the program that forms part of the algorithm Google uses to sort through billions of search result pages in order to determine which search results are most relevant to specific requests. It forms the latest component of Google's Hummingbird search algorithm, which drives the overall system.
The Hummingbird algorithm is composed of numerous components, including Penguin, Payday and Panda, which fights spam. Pigeon is a component which enhances local results, while Top Heavy downgrades web pages stuffed with ads.
Another crucial component to pay attention to is Mobile Friendly, which rewards pages for publishing mobile-friendly content. Pirate battles copyright infringements, while PageRank plays an important role in a strategy that dishes out credit to pages based on the in-bound links they receive from other websites.
RankBrain harnesses artificial intelligence to implant words into numerical entities in order to make them more understandable to computers. Any words or phrases unfamiliar to RankBrain will be converted into words and phrases with similar meanings for filtered results. This increases Google's accuracy and efficiency when tackling new search queries.
Signals For Search Rankings
Google determines web page rankings with the help of elements called 'Signals'. Content read on a web page is regarded as a signal. Keywords highlighted in bold font will be treated as another signal for the purposes of determining web page rankings.
These calculations contribute to the overall PageRank score, which is then interpreted as a signal. In the same manner, if a web page has been evaluated as mobile-friendly, this is then registered as a signal with Google. Every signal is processed by the various components within the Hummingbird algorithm, which then determines which pages Google highlights in response to particular queries.
Google has been quite open about acknowledging the presence of over 200 major signals for ranking. These signals themselves may have up to 10,000 smaller sub-signals or variations. Despite the short time it's been around, RankBrain is now considered Google's third most important factor in the search rankings.
RankBrain is just one of numerous signals in Google's search page algorithm. The question on everyone's lips is which considerations are first and second in importance when it comes to determining search rankings. Thus far, Google has been reticent about them, so experts are forced to rely on intelligent guesswork.
While knowing which factors are most important is certainly useful, how they work is a mystery until Google actually explains them. For now, links remain a firm favorite, and that doesn't look set to change until Google confirms their most important signals, if ever. The second most important signal is widely considered to be 'words', although this term can be unhelpfully vague as it actually includes everything from Google's textual interpretation to the actual words that appear on a page.
How Does RankBrain Work?
RankBrain's primary function is to analyze search queries submitted via Google in order to identify specific pages that may not necessarily contain the exact words contained in the query. RankBrain's objective is to enable users to find pages that may not necessarily match their specific queries but are nonetheless relevant.
Google has already enabled users to locate pages that don't match their particular search terms for a long time, but that wasn't always the case. Years ago, a search query containing the keyword “jacket” may not have tossed up pages containing the plural “jackets”.
Google has corrected this over the years, and these days’ plurals are simply a variant of the singular, just as the word “walking” is a variant of “walk”.
Google has also ramped up its intelligence when it comes to the use of synonyms. Users searching for the word “jackets” are now more likely than ever to find different variants of winter and summer coats for both sexes.
Concepts can now also be told apart to a certain degree, which means that Google knows the difference between Apple's technology and the fruit, and is able to determine which of the two users are searching for based on its own analytics.
In 2012, Google's Knowledge Graph was launched, and boosted Google's capabilities in making connections between different words. Knowledge Graph consists of an enormous database of information on a staggering range of topics, including the relationships between them.
Try searching for a phrase like “When did the first man reach the moon?” and you're likely to receive information on Neil Armstrong's lunar touchdown on 20 July 1969, even without ever mentioning his name in your query.
RankBrain Smartening Searches And Refining Queries
In order to enable more refined searches, Google has traditionally had people create stemming or synonym lists or establishing different connections, which requires a gargantuan amount of human effort. Even more daunting is the fact that almost 3 billion searches take place on Google every day. In 2007, Google revealed that nearly 20-25 percent of the queries had never been run before. In 2013, that figure fell to 15%. Still, 15% out of 3 billion is an intimidatingly high number of 450 million.
Users these days are using longer and more complex queries, often inserting multiple words or long-tail keywords. It is virtually impossible for human beings to handle such data on their own, hence RankBrain, which was designed to interpret such requests in ways that would enable the user to obtain the most relevant search results.
RankBrain has the ability to identify patterns between words and understand how they are related to each other. This also helps it to refine future searches related to existing topics, and produce increasingly suitable results for users.
RankBrain does this by connecting original and complicated queries into shorter answers that are relevant to the user's query. Google often manages to produce answers even to the most unusual queries.
There are high hopes for RankBrain, which is expected to dramatically improve the user experience. Google has traditionally been quite conservative with updates to its ranking algorithm, preferring to undertake small tests bit by bit. Big changes are usually only launched when complete confidence in their delivery has been achieved.
So a huge move like introducing RankBrain as Google's third most important signal is something to really sit up and take notice of.
Starting in early 2015, RankBrain was gradually rolled out in stages. It has since gone fully live and has been available globally for a few months. Google claims RankBrain is processing a large portion of queries, although a specific figure has not been supplied.
According to Google, all of RankBrain's learning takes place offline using the following mechanisms:
- Historical search data is processed by RankBrain, which then establishes how to make predictions based on that data.
- These predictions are tested and proven before they can go live.
- The cycle is repeated each time a new search is entered into the system.
This is reminiscent of Microsoft's RankNet machine-learning technology, introduced in 2005. RankNet played a big role in what became Bing's search engine, but is barely mentioned nowadays. Now that RankBrain has appeared on the scene, experts are expecting Microsoft to step up their game.
RankBrain Appears To Be Here To Stay
Part of the reason RankBrain is considered so important by Google is because it directly contributes to page rank. While signals are usually factors related to keywords, links, server security and other web content, they can also be tied to specific users, thanks to details like their geographic locations and browsing history.
While the actual mechanics of RankBrain are unclear, it's likely some sort of RankBrain score is generated in order to indicate the quality of different webpages. This might be done in a way that enables RankBrain to help Google classify pages in a more organized manner based on their content. In all likelihood, RankBrain is equipped with the capability to summarize the contents of a web page better than any of Google's prior systems, making it one of the most impactful changes introduced by the company in a long time.
While Google is still remaining tight-lipped about RankBrain's technology, it is known that a ranking component is definitely involved. It is possible that RankBrain is actually a combination of a ranking signal and search processing tool. But one thing remains clear—RankBrain's artificial intelligence application has broken new ground and looks set to remain relevant for a long time.
Courtney Capellan is a “technically creative” freelance writer and digital media analyst for Hotel Marketing Works. Her specialties include digital marketing, hospitality management. Brie Moreau is a digital Marketing analyst for DigitaladvertisingWorks. DigitaladvertisingWorks specializes in AdWords & PPC management and SEO.