BERT ALGORITHM UPDATE

The BERT algorithm, or “Bidirectional Encoder Representations from Transformers”, is Google’s latest update to its search system. Google has said it’s “the biggest update to the system since RankBrain launched five years ago”.

The BERT update was announced on October 24, 2019, but had started rolling out to parts of the search engine as early as October 18th. Due to the significance of Google’s statements attached to the announcement, Today’s Business has sought to inform clients on how BERT may impact performance as well as future strategy.

HOW MIGHT BERT DIRECTLY IMPACT
YOUR DIGITAL MARKETING PERFORMANCE

RANKINGS VOLATILITY:

  • Performance of many keywords may have greater variance in the second half of October. This should stabilize as the algorithm continues to mature.
  • Pages that were indirectly ranking for a query despite not being entirely accurate are less likely to hold their rankings. By and large, these should not impact traffic and lead generation significantly because the people searching were less likely to be finding what they want to begin with.

KEYWORD CHANGING:

  • While BERT is focused on improving its accuracy for search intent, it may also change the type of search results that Google provides. A search query that was previously displaying service pages may now be focused entirely on articles. This has been occurring with greater frequency over the past 18 months, with BERT being Google’s ultimate culmination of  experimentation. Because of this, paying attention to the types of results being displayed is more crucial than ever.

PAGE LENGTH:

  • Page length standards are likely to show greater variance depending on the needs of addressing a search, which means that once again some pages may be needing more content while others less.

STRUCTURED DATA:

  • Greater implementation of Q&A, FAQ Page, How-To, and Speakable schema types should be considered.

MORE LONG-TAIL KEYWORDS TO TARGET

  • As search behavior becomes more closely aligned to natural language, there is likely going to be more diversity in search queries by users, meaning even more opportunities to target and leverage for traffic.

WHAT IS THE BERT UPDATE?


BERT is a deep learning algorithm related to natural language processing – in other words, Google is seeking to better enable people to use natural language while searching instead of keyword-based queries. Google uses BERT to try to better understand the context of a search query and to more accurately interpret the meaning of the individual words. This breakthrough is built on mathematical models called Transformers: These analyze a word in relation to all the other words in the sentence – or in the case of Google search the search query – and do not simply look at the meaning of words in isolation. This is particularly useful when interpreting the meaning of prepositions and the position of individual words within a search query

BERT ALGORITHM UPDATE

EXAMPLE ONE:

  • A query for “Can you get medicine for someone pharmacy” would have previously returned a result simply for instructions on getting a prescription. Because of the BERT update, the result closely matches the intent of the search by picking up on the fact that the searcher is actually looking for information on retrieving a prescription for another person.
BERT ALGORITHM UPDATE

EXAMPLE TWO:

  • In this case, the query is “2019 brazil traveler to usa need a visa”. Previously, Google had difficulty understanding the significance of the word “to” in context as well as the function of “brazil” as an adjective towards the type of traveler in the subject. The user’s intent is to find out whether a visa is needed for a Brazilian to travel to the USA. But Google’s result was geared towards Americans traveling to Brazil. After the update, it is now displaying the US Embassy website for Brazilian tourists, a much more relevant search.

WHAT IS THE BERT UPDATE GOING TO IMPACT?


Overall, the BERT update is more subtle in the immediate future than many core algorithm updates in the past 18 months. Below, you can see that a check of how significant this shifted rankings indicated this as being moderate.

graphbert

With that being said, it appears the sites that were most commonly affected fall into the “Your Money Your Life” category
(Websites that involve your health or items you may buy):

  • Shopping
  • Apartments
  • Food
  • Healthcare

THE BERT UPDATE WILL IMPACT ONE OUT OF TEN QUERIES

  • Google described this update as a massive change impacting 10% of all queries and called it one of their biggest steps forward for search in the past 5 years (and one of the biggest steps forward in the entire history of search).
    Google is now trying to match results with the user’s search intent rather than the exact keywords they used. This also applies to featured snippets

BERT IS FOCUSED ON UNDERSTANDING LONGER-TAIL, NATURAL LANGUAGE QUERIES.

  • Google’s implication that BERT will impact 10% of searches may mean that long-tail searches are on the rise. It is important to note that long-tail keywords already represent nearly 70% of search volume and that 15% of all annual search queries have never been seen before by Google.

Example of Short Tail vs. Long Tail keyword:

“Broken leg treatment” Short Tail

“What are the treatment options for a broken leg” Long Tail

Long tail queries that are not always targeted by SEOs and businesses because they tend to send less traffic to sites than shorter-tail queries.

  • The intent of Google is to provide users with more detailed and helpful information:

Search will be able to understand the context of the words in the query (for conversational queries or searches where the words “for” and “to” are significant).

Users can search in a way that is more aligned with the way they
would speak.

BERT also aims to take learnings from one language and apply them to others, which helps Google to return relevant results in all the languages that search is offered in.

WHAT CAN BE LEVERAGED
FOR SEO PURPOSES?


Initial strategies are limited. The BERT Update is to enhance Google’s natural language processing, essentially creating a stepping stone for Google to research and understand the natural language used in context within search queries, not the content on a page. Google can already pick up on synonyms and misspellings, however, according to Search Engine Journal, with addition to BERT, Google will better interpret the queries.

  • Optimizing pages to incorporate more FAQ sections, focused not just on keyword density, but quality answers can pose a positive impact on BERT. To the extent of featured snippets, marking up pages with FAQ Page, How-to, and Q&A schema is recommended for increasing the page’s probability of being picked up by BERT. There is not an extensive amount of page optimizing to go along with BERT other than to continue to produce relevant content of excellent quality.
  • BERT will also have a big impact on voice search as well. If users are seeing more accurate results for natural language being used, searches are likely going to expand. Utilizing schema to help google identify more conversational text should be considered by more sites. This should not be considered for pages not yet ranking in the top 10.
  • Creating shorter FAQ-style individual posts that satisfy a query can enable ranking for longer tail keywords.
  • Reducing the use of Keyword heavy sections/language and focusing more on properly addressing each concept attached to a topic.

In the upcoming months, changes to the update may occur as Google often tweaks algorithm updates once they are rolled out. Today’s Business will be monitoring these changes moving forward and reporting on any significant updates made that could impact our clients’ performance.

Post Your Facebook Comments