In 2018 alone, Google made over 3,200 changes to Search, which was a dramatic increase from 400 in 2010. This year, Google publicly announced core and other search algorithm updates that shaped how we can optimize the use of the search engine.
Aside from the number of algorithm updates this year, Google also established a naming convention for its core updates and brought the natural language to understand search queries better.
In this guide, we’ve detailed what webmasters should know about the Google algorithm updates that shaped the search rankings this year.
BERT Natural Language Processing
On 25 October 2019, Google introduced the BERT algorithm to U.S. English search queries. According to Google Fellow and Vice President for Search Pandu Nayak, the introduction of BERT aims to make searches better by understanding language. “At its core, Search is about understanding language. It’s our job to figure out what you’re searching for and surface helpful information from the web, no matter how you spell or combine the words in your query,” he said on the blog post.
Last year, Google introduced and open-sourced a neural network-based technique for natural language processing pre-training called the Bidirectional Encoder Representations from Transformers or BERT for short. With BERT, anyone could train their own question answering system.
Google researched more on transformers, which are models that process words in relation to all other words in a sentence as opposed to one-by-one in order. BERT models can now consider the full context of a word by looking at the words that come before and after it. Using BERT in Searches means that your queries on the search bar are now better understood by considering your intent.
With the previous Google language algorithm, users would often use “keyword-ese” queries on the search bar to come up with relevant results. But with BERT, the Google Search is now learning to recognize the intent behind the natural-sounding way of asking questions. Using Search with BERT means that you can use correct grammar in your queries instead of making it sound like a string of keywords that don’t make sense.
The introduction of BERT to the Google algorithm also necessitated the improvement of hardware. The need for better hardware is because of the complexity of the models built with BERT would push the limits of traditional equipment. Google is now using the latest Cloud TPUs to serve search results to get users with more relevant information quickly.
The BERT models and learnings generated from U.S. English were taken and used to improve search queries for other languages.
BERT’s Impact on Searches
According to Nayak, BERT is one of the most significant milestones in the history of Search:
“With the latest advancements from our research team in the science of language understanding — made possible by machine learning — we’re making a significant improvement to how we understand queries, representing the biggest leap forward in the past five years, and one of the biggest leaps forward in the history of Search.”
How does the implementation of BERT exactly affect Google searches?
Because the language algorithm understands queries a bit better than the previous one, this has improved. When it comes to ranking results, BERT will help Search better understand one in ten searches.
Google did testing to ensure that the changes are actually more helpful for users. One of their examples was a query for “2019 brazil traveler to usa need a visa.” The prepositions “to” and its relationship to the other words in the query are essential to understanding the meaning. When we read it, this means that a Brazilian is traveling to the U.S. and not the other way. Google’s previous algorithm wouldn’t recognize this importance and would return results about U.S. citizens traveling to Brazil. But with BERT, Google Search can grasp the difference. It recognizes the importance of the word “to” in the context of the query. This means that users get search results that are more relevant to them.
Another example is the query “do estheticians stand a lot at work.” With the previous system, Google Search would match the term “stand-alone” with the word “stand” included in the query. But of course, that isn’t the correct definition of the word in this query. With the BERT model, Search now understands that the word “stand” is related to the physical demands of the job and gives users more relevant information.
After the implementation of BERT in U.S. English, Google noted that it affected 10% of search queries. We would probably see the same figure for search queries in other languages.
BERT Rolled Out for 70 Languages
Last 9 December 2019, Google announced that it rolled out BERT for over 70 languages around the world.
BERT is now available for the following languages: Afrikaans, Albanian, Amharic, Arabic, Armenian, Azeri, Basque, Belarusian, Bulgarian, Catalan, Chinese (Simplified and Taiwan), Croatian, Czech, Danish, Dutch, English, Estonian, Farsi, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Javanese, Kannada, Kazakh, Khmer, Korean, Kurdish, Kyrgyz, Lao, Latvian, Lithuanian, Macedonian, Malay (Brunei Darussalam and Malaysia), Maltese, Marathi, Mongolian, Nepali, Norwegian, Polish, Portuguese, Punjabi, Romanian, Russian, Serbian, Sinhalese, Slovak, Slovenian, Spanish, Swahili, Swedish, Tagalog, Tajik, Tamil, Telugu, Thai, Turkish, Ukranian, Urdu, Uzbek, and Vietnamese.
Google Core Algorithm Updates
Core updates are released by Google multiple times in a year. These updates don’t have a clear or specific focus on the particular search query or in the characteristics of a website. Google’s core algorithm updates simply make subtle changes. These updates aim to improve search results by giving relevant content to the users.
Because of this, webmasters cannot do anything to remedy errors to recover the rankings lost from core updates. The best way to approach this problem is to come up with better and relevant content for your users. Using resources like the Google Search Quality Rater Guideline will help you with content creation.
March 2019 Core Update
Google launched its first significant algorithm change back in March 2019. This update was also the first to follow Google’s new name convention for its updates.
According to a survey launched by Search Engine Land, the feedback did not reveal any apparent patterns as to the types of sites or content affected by the update. Over 58% of the respondents said that they experienced a negative impact; 33 % said they saw a positive change while the remaining 9% reported no change.
June 2019 Core Update
Released on 2 June 2019, this was also the first time that Google pre-announced a core update. It announced the rollout of the update a day ahead. This month’s update had a more positive effect in search rankings, unlike the March 2019 Core Algorithm Update.
According to reports, the June 2019 update affected websites that relied mostly on amassing search results. The Daily Mail complained of a 50% loss in its daily search traffic.
Sistrix detected more sites that have experienced massive drops in visibility after the implementation of this update. However, sites like Mirror, Sun, and HuffPost have enjoyed significant gains in search visibility.
September 2019 Core Update
Google announced the release of the update several hours ahead before the launch. Released on 24 September 2019,
While Google’s core updates are more general, sources say that the September 2019 update considered the quality of links. According to the Search Engine Journal, some sites already experienced a significant drop before the announcement. Drops were experienced by sites with irrelevant 301 domain redirects and links from expired URLs.
Google also announced a change in how they looked at no-follow links. Before this update, the search engine looks at no-follow links as uncounted or unincluded indirect statements. But with the September 2019 update, Google looks at the links as a hint which can either be followed or not, counted or not, depending on their discretion for ranking purposes.
According to SEO forums, the September update focused on EAT (Expertise, Authority, and Trustworthiness) and the YMYL (Your Money or Your Life) category. Reports show that the impact was about a 10-20% drop in page views, which is not as significant as the core update in June 2019.
Other Confirmed Updates
Aside from the major core and language algorithm updates, Google also launched other types of updates within the year.
Valentine’s Day Update
Launched on 13 February 2019, this update focused on identifying the best pages to address a user’s query based mostly on relevance and not quality. With this change, page titles became important for search queries.
Different Signals for YMYL Queries
Google published a paper in February 2019, explaining how it fights disinformation in Search, News, YouTube, and other products. In the paper, Google wrote that it gives more weight in its ranking algorithms to signals that indicate EAT for queries that might impact a user’s well being. These types of queries are YMYL queries.
Google writes on the document: “For these YMYL pages, we assume the users expect us to operate with our strictest standards of trustworthiness and safety…As such, where our algorithms detect that a user’s query relates to a YMYL topic, we will give more weight in our ranking systems to factors like our understanding of authoritativeness, expertise, or trustworthiness of the pages we present in the response.”
April 2019 Deindexing Bug
On 3 April 2019, Google confirmed issues with deindexing after users reported a significant drop in their rankings. About 4% of stable sites underwent deindexing, which meant a loss in traffic, sales and revenues, conversion, and overall site performance for the month. The bug fixed itself after a few days, but some sites still had to reindex their pages in the Google Search Console manually.
May 2019 Indexing Bug
Another indexing bug came out on 23 May 2019. Google confirmed that the issue was related to the inability to index new content. The bug was resolved two days after.
Pre-announced and released at the same time with the June 2019 core algorithm update, Google launched its Diversity Update but was not related to the former. June’s diversity update aimed to increase the domain diversity in search results by showing a maximum of two pages from the same domain in searches. The update counts the subdomain URLs as part of the root domain.
Timely Featured Snippets
An update in August was implemented to ensure that the featured snippets feature current information. Talking about the update, Google said: “For queries where fresh information is important, our systems will try to find the most useful and up-to-date featured snippets.”
For example, when users look for upcoming holiday dates, they will be presented with forthcoming holiday dates and not the previous year’s dates. As long as publishers keep their pages updated or regularly publish current information, they may see greater visibility in featured snippets.
Emphasis on Original Reporting
A week after Google updated its Search Quality Raters guidelines, it announced that it had been rolling out algorithm updates to give more preference to original reporting. This update was implemented in September 2019.
Within products like Google Search, News, and Discover, original news sources surface more prominently and stay at the top of the news cluster longer. This update rewards original coverage of a story over “value add” reporting or content syndication.
Local Search Algorithm Update
In November 2019, Google confirmed that the fluctuations in organic search reported throughout the month were the results of the rollout of an update. The update was officially named the Nov. 2019 Local Search Update. Google only released a comment about the update on 2 December 2019.
The local algorithm update will help users find the most relevant results for local search that matches their search intent. The new algorithm now understands the concept behind the search by understanding how words in the search query are closely related to each other. With the update, Google will use its AI-based neural matching system to rank businesses in Local Search results.
According to Google, webmasters don’t need to make any changes to the site because of the update. The company also said that this update impacts users worldwide across all languages.