What is Google BERT and what does it mean for the future of SEO?
BERT, which stands for Bidirectional Encoder Representations from Transformers, is a neural network-based technique for natural language processing pre-training.
What that means to you and I is that it can be used to help Google better discern the context of words in search queries, and to more accurately interpret the meaning of the individual words.
BERT analyzes a word (in this case, a Google Search Query) in relation to all the other words in the sentence and does not simply look at the meaning of words in isolation. This is particularly useful when:
- Interpreting the meaning of prepositions such as “with”, “to”, “in”, “from”
- The position of individual words within a search query.
According to Google, around 15% of all search queries are new – that means they are being searched for the very first time. Furthermore, the phrasing of search queries is growing closer and closer to real human communication – in part under the influence of technological advances like Voice Search, supported with the rapid growth of mobile devices and smart home device adoption.
As a result, search query length is also increasing where 70% of searches can be considered long-tail. People turn to Google with fully-formulated questions and expect precise answers in a fraction of a second – and BERT now makes up a significant part of the technology that makes this possible.
When was Google BERT released?
BERT was deployed the week of October 21, 2019 for English-language queries, including featured snippets (e.g. Quick Answers and People Also Ask).
The algorithm will eventually expand to all languages in which Google offers Search, but there is no concrete timeline just yet. A BERT model is also being used to improve featured snippets in two dozen countries.
What was the impact of Google BERT?
Google said BERT impacts roughly 1 in 10 queries (10%). Yet, many SEO professionals and many of the keyword tracking tools did not notice massive changes in the Google search results while this algorithm was released. This is largely due to the same individuals and tools tracking short-tail queries, e.g. “ceiling fan with lights” as opposed to a long-tail query like “how to buy a ceiling fan with lights with remote control”.
Google said BERT is not only one of the biggest steps forward for search in the past 5 years, but it’s one of the biggest steps forward in the history of search altogether.
Can a website optimize SEO for Google BERT?
Google's Danny Sullivan, who is the company's public Search Liaison stated the following:
There is no simple answer to optimize SEO for BERT. There aren’t any easy tactics that you can use to suddenly make your website rank better.
However, there are still possible tactics that can be taken to improve the content on your website, especially to target more natural-language, long-tail queries.
Generate more long-tail content
Consider structuring descriptions of products and services to be more conversational in nature with connecting words:
The [product/service] product with [unique differentiator] is the ideal [product/service category] for:
Use case 1
Use case 2
Use case 3
This enables the ability to cast a wider net for different use cases.
Other options to consider
- Generate more online reviews on product / service pages
Reviews make a great form of user-generated-content, coming directly from users – which is more influential for future customers. As reviewers are buyers of your ceiling fans, the natural language they use is similar to the one your target audience will use when searching for products. Thus, the consistent flow of reviews will help you target more traffic through the right keywords, coming directly from your own customers. Naturally, replying to reviews will also help generate more long-tail query content.
- Create more robust FAQs
What are the most common questions asked about your products or services? The more you can state and answer these questions, which are the lion's share of long-tail queries, the more you will show up in search, especially if you provide the best answers. If the answers are standardized across the product types, you can still state the question within the product/service display page and link the answer to the relevant FAQ page.
Ultimately, you want to focus on writing great content — for real people. If you are putting the customer’s needs first, chances are you are already “optimizing” for Google’s BERT algorithm.