Google BERT has been dubbed, “Google’s biggest change in 5 years,” and “One that will impact one in ten searches.” But what exactly is it?
Rolled out on October 25, Google BERT is a deep learning-powered Natural Language Processing (NLP) model designed to help Google better understand language in order to serve more relevant results. Rather than returning results based on search strings, it looks at the relationship between the words in a sentence and at the wider context of a search.
Why Does This Matter?
BERT signals a shift from keyword search to semantic search. It will have a major impact on countless websites and how they optimize for SEO. For example, a trucking employment website would traditionally have contained separate pages variously optimized for the terms “trucking jobs,””jobs in trucking,” “trucking employment,” and so on in order to capture different keyword searches.
It’s All in the Context
However, BERT understands that these terms are functionally equivalent, rendering those extra pages extraneous. That’s because BERT understands the context (and content) of a search query at the sentence level – and even at the document level. Rather than focusing on keywords, BERT can determine whether a piece of content makes sense, and how that content scores against Google’s key EAT (expertise, authority, and trust) features. It’s all part of Google’s ongoing efforts to highlight quality content while penalizing its poor-quality counterpart.
How do we Optimize for BERT?
Since BERT’s launch, traditional SEO approaches involving keyword lists and keyword optimization are effectively dead. Instead, content marketers need to start building their content at the topic level. To do this, they should work with their internal subject matter experts to identify relevant, high-interest topics for discussion. Subject matter experts not only have the knowledge to be able to cover topical themes, but they can write about them in a way that will naturally capture the language and context BERT will be looking for.
Hold a Conversation
However, this doesn’t mean that they should get academic about their writing – or try to write to match the typed search strings we’re used to. Google has made it clear that conversational search is now the default. People don’t say “best artisanal ice-cream, DC” when using voice search. Instead, they ask, “Hey Siri, what’s the best artisanal ice-cream shop in Washington, DC?” Consider what kinds of questions people will be asking around your topic, and answer them in a natural, conversational manner.
Quality Always Wins
One thing hasn’t changed, however. When rolling out algorithm updates, Google always emphasizes producing quality content that fits your users’ goals – and with BERT this is even more crucial. If you’ve been focused on keywords or if you’ve been cheaply outsourcing your content, now’s the time to shift your approach.
StellaPop has been doing a deep dive into all things Google BERT, so for more insight into what BERT means for your content strategy, or for help developing a content strategy aligned with the BERT rollout, get in touch!