BERT: What to do to Remain Ahead in the Next Generation of SEO
What is BERT?
BERT is the abbreviation for Bidirectional Encoder Representations from Transformers. It is a Google algorithm update that aims to understand the intent of search queries better based on the context of words, in order to return the most relevant results.
It was rolled out in 2015 and Google referred to it as “the biggest leap forward in the past five years, and one of the biggest leaps forward in the history of Search.”
BERT is Google’s latest AI introduction to search engine optimization. Like RainBrain, its aim is also to improve the bot’s understanding of search intent. It achieves this through Natural Language Processing, a technology that enables computers to process and understand natural human language.
BERT and RankBrain
It is important to understand another important update similar to BERT and released some years ago: RankBrain.
In October 2015, Google took the first major step in introducing artificial intelligence to its Search algorithm by introducing the RankBrain update. RankBrain is the machine learning system used by the Google search bot to judge the search intent of queries.
It immediately transformed how SEO was done as 1) it up ranked UX signals such as CTR, dwell time, and bounce rate and 2) it proved how important content relevance was to rank higher.
The BERT model is bi-directional because it tries to understand the context of a word by considering the preceding and proceeding words. It pays particular attention to subtle nuances in search queries, as well as context signals such as prepositions.
This change does not only affect SERPs but also featured snippets. The end of BERT, and for the algorithm as a whole is for Search to improve at returning the most relevant results for a user’s queries.
What BERT is NOT!
BERT is a huge update and it would set the pace for similar AI introductions to the search algorithm. It has, however, become somewhat overrated and that has given birth to a whole lot of misconceptions.
- BERT does not replace previous updates, and particularly not RankBrain. Instead, it runs parallel to other ranking factors. And it is not the biggest Search update either.
- It does not require optimization for long-tail keywords. Indeed, BERT has effects on long-tail keywords, as Google itself demonstrates. However, it helps Google understand user intent based on context rather than stretch the importance of long-tail keywords.
- BERT has a small impact. Just as there are misconceptions overstating BERT’s impacts, there are myths doing otherwise. According to Google, BERT affects 10% of English searches in the US. Likewise, many SEO experts expect that the algorithm would get better and gradually extend to the majority of queries.
In all, keep in mind that there is nothing to ‘optimize’ per se when it comes to the BERT update. Making unnecessary tweaks would do damage to your ranking. According to Google’s Danny Sullivan, “there’s nothing to optimize for with BERT, nor anything for anyone to be rethinking. The fundamentals of us seeking to reward great content remain unchanged.”
Instead of looking for ways to ‘optimize’ your website content for BERT, focus on improving the quality and relevance instead.
What to Do?
- Have a single, clear focus. It is of no use to try to answer ten different questions in one post. Your keyword research should aim to find, not just the most popular words and phrases, but what internet users are searching for the most. Each post should then be tailored to answering each question, particularly for informational search queries. Having a single focus not only brings clarity to your content but also the title, metatags, meta description, etc.
- Provide direct answers. Long-form content may not be as important as you think. Your content can rank tops even if it has only 300 words, as far as it answers the query directly. Focus on the quality of your content, rather than quantity.
- Be relevant to your target audience. The biggest takeaway from the BERT update is that relevance is much more important than ever. You should ask yourself, ‘if my question is this, does the post answer it satisfactorily?’ As far as you create content that is genuinely useful to your audience, you do not have much to worry about.
BERT and Voice Search
BERT is thought, and rightly so, to be the starting point for a shift in focus to voice search. The increase in AI assistant devices, home devices, smartwatches, and all other IoT devices has led to a corresponding increase in the number of voice searches.
A unique characteristic of voice searches is that they are more conversational in structure. Therefore, search algorithms need to pay more attention to intent and context to understand the users and return relevant results. BERT, being an NLP update, bridges the gap between computer processing and human language comprehension. In short, voice search has the capacity to disrupt Search in a major way.
Optimizing for voice search involves:
- Local Search Engine Optimisation. Keep your online business profiles up to date and tailor your content and keywords to your local audience.
- Post informational content. Aim to answer specific questions.
- Use accurate descriptive alt tags for your images.
Since algorithms are now able to ‘understand’ human language and search intent better, the most important SEO strategy would be to optimize your content for quality and relevance. BERT is an important reminder of this. Going forward, artificial intelligence has come to stay.