What that looks like, and how that influences companies' content marketing strategies, is the challenge.
Free Guide: The Inbound Marketer’s Guide to Search Engine Optimization
Instead of using keyword matching — albeit on a more and more complex scale — BERT will seek to understand the meaning behind sentences, paragraphs, and the whole article of a page.
Content strategies that already focus on answering search queries and providing meaningful information have nothing to worry about. But companies that relied on keyword stuffing may see their rankings drop if they don't refocus on relevance and answering to searchers' real intent.
About Google's BERT
BERT, or Bidirectional Encoder Representations from Transformers, is Google's most recent algorithmic update to how it processes queries and curates page results. It uses machine learning to process natural language and interpret the meaning of queries and page content.
(BERT is also an academic, open-source project, but it's important to keep the two separate as your company researches SEO strategies.)
BERT will impact the following areas of Google search the most:
Google made this change to drive more attuned curation of page results.
Rather than taking the words and sentence format of a query and compiling a list of sites that have high domain authority and rank highly for the keywords (along with a host of other factors), BERT "understands" the question. Then it pulls up the pages that best answer the question.
This means Featured Snippets are going to be able to better address more granular questions.
It also means that when people use Google Home or say, "Okay, Google," the singular answer they get in response is more likely to be the answer they need.
Rather than picking about the sentence string for identifiable and popular words, it appears that every word is taken into account to understand why the user is requesting information, the information that actually addresses the question, and the context of the answer on the sites it pulls up in response.
This means forums, how-to pages, and hyper-specific explanations are rising in the ranks to meet users' needs. Popular articles and sites with general-topic information are less likely to be first.
What Should Businesses Be Doing About BERT?
In a tweet, Google Public Search Liaison Danny Sullivan said, "There's nothing to optimize for with BERT."
But that assumes businesses are already committed to the ideal of great, natural-sounding content. The ultimate focus must be on authentic content marketing, starting with easy-to-read web pages, regularly posted blogs, and video content (with closed captions and transcripts whenever possible).
Having strong answers and explanations for user queries, especially in a high-quality, natural-sounding voice, has been a content priority for years, and BERT capitalizes on this focus.
With its machine learning background in studying different voices, angles, and language expressions, BERT doesn't just recognize meaning. It may also identify content that sounds too pushy, too promotional, and too low-quality — and, most importantly, too keyword-stuffed.
The release of BERT is too new for marketers to have ironclad rules about what does and doesn't pass muster anymore. Some companies may see high-performing posts drop in the rankings, and others may see new traffic. The biggest impacts will be felt by brands that rely on organic traffic.
Businesses should focus on market research strategies to rank higher
The word "relevance" has come up in a lot of research and analysis surrounding BERT's release.
Addressing their questions, their problems, and how they think — with long-form posts for these granular topics — will make a site more useful in both BERT's eyes and a reader's. Rather than looking into keyword strategies, companies (especially local companies) should focus on:
Questions that their ideal customers ask, no matter how specific
What words and word order signify different types of user intent and need
Including forums, review sites, and social media threads that can informally answer a greater range of search queries
The advice of continuing to "write naturally" is more helpful than it seems at first glance. It's the SEO cheat code for this algorithm: write in ways that the algorithm understands to be helpful, natural, and authentic.
And answer the question the users are really asking.