Google’s BERT Update Will Impact 10% Of All Search Queries
Background
On October 25, 2019, Google rolled out its biggest search algorithm update in five years, called BERT (which stands for Bidirectional Encoder Representations from Transformers). The update is a huge advancement toward understanding natural language and the intent behind search queries, and it’s predicted to impact 10% of English searches in the United States. BERT will roll out to more languages and locales in the coming months.
How Does the BERT Algorithm Work?
BERT models use machine learning to evaluate a word’s meaning based on the words surrounding it, as opposed to only the word right before or after it. Which means that Google Search can distinguish how a word’s meaning can differ depending on its context. This will especially affect longer, more conversational queries, or searches where prepositions such as “for” and “to” impact the meaning. For instance, see Google’s example of BERT’s ability to understand the intent behind a search query:
“For the search query “2019 brazil traveler to usa need a visa”, the word “to” and its relationship to the other words in the query are especially important to understanding the meaning. It’s about a Brazilian traveling to the U.S., and not the other way around. Previously, our algorithms wouldn’t understand the importance of this connection, and we returned results about U.S. citizens traveling to Brazil. With BERT, Search is able to grasp this nuance and know that the very common word “to” actually matters a lot here, and we can provide a much more relevant result for this query.”
– Pandu Nayak from Google
How Will Google Search Be Impacted By BERT?
BERT models have been applied to both search rankings and featured snippets, and Google predicts the algorithm will affect one in 10 English searches in the United States. Instead of using strings of keywords and phrases that don’t match how people naturally speak, Google hopes to provide more relevant and useful information to searchers who use conversational language to search.
What This Means For SEO Performance
BERT is a significant improvement in how Google understands queries. With BERT models, Google can now grasp the subtle nuances of language, much like humans. And with a better understanding of sentences, Google Search results and rankings will change, impacting informational content that often ranks for long-tail queries and Quick Answers.
Before panic sets in, it’s important to know that most websites won’t see drastic changes in SEO key performance indicators (KPIs). Here’s why:
- Website traffic: BERT is focused on long-tail natural language queries that typically have less search volume than short-tail queries. For websites that do rank for long-tail queries, it’s likely that the traffic changes from an already low search volume will be minimal in comparison to short-tail queries that typically drive more website traffic.
- Keyword rankings: Many keyword tracking tools track shorter keywords, not long-tail queries, so it is likely that tools will not pick up on significant movement.
- Quick Answers: BERT will most likely impact owners of Quick Answer results, so it is important to monitor your Quick Answer results and associated rankings.
On the positive side, some websites may see improvements in performance. In particular, BERT could benefit websites that have existing user-friendly and well-optimized content that answers those long-tail, natural language queries.
Moving Forward With BERT
Google is evolving to become more of an answer engine than a search engine. This evolution will move it closer to providing very specific answers to detailed questions, rather than simply providing a more general set of results for users that are loosely related to a query.
From a marketing perspective, Collective Measures’ approach to organic search is not changing. In fact, the update validates our current approach – we will continue to focus on creating high-quality content that aligns to topics and intent. As with all algorithm changes, the full impact won’t be fully known for a few weeks. But if Google is telling us it will be big, we’ll take their word for it.
