Google’s new BERT algorithm means search engine marketers will need to change their approach to keyword research in 2020 and focus more on intent research. Adam Bunn, the Director of SEO, Content & Digital PR at Greenlight Digital, looks at what companies should expect from keyword searches in 2020.
The way people search online is changing. The introduction of Google’s ‘BERT’ algorithm (Bidirectional Encoder Representations from Transformers) is evidence of this and highlights the complexity with which people have begun to utilise search engines. BERT utilises Natural Language Processing (NLP) which helps analyse natural human language beyond just basic keywords and gather more information about how all the words in a query relate to each other.
In this way, BERT can look at the search query as a whole, rather than focusing on independent keywords in an attempt to reveal, and then prioritise, the intent of the search. This ensures the search results are the most relevant not only when it comes to the specific topic the user is researching, but also to the user’s intention behind the search.
As a result of this change in the way search engine queries are being performed, marketers must adapt to the way they tackle Search Engine Optimisation (SEO). Fréderic Dubut, Senior Programme Manager Lead at Bing, recently said that search engine marketers must change their approach to keyword research in the following year and focus more on intent research. But does this mean keywords are going to become redundant?
Voice search changing SEO terms
BERT is one of Google’s biggest updates to web search in recent years. It uses NLP to better understand the context of search queries by taking into account things such as linking words (and, the, from) or interrogative markers (how, what, which). While some users have learned to query Google using unconnected and grammatical keywords such as ‘best digital strategy SEO’, the popularisation of voice search demands that search engines understand the way people naturally speak and look beyond just keywords.
Voice search produces queries that use conversational language such as “what is the best digital strategy for SEOs” which means they require NLP in order to render the best results. BERT can also take into account previous recent searches to some extent in a similar way to how a regular conversation works. Asking “How long does driving there take?” after inquiring about the location of nearest supermarket will provide relevant results without the user having to specify the supermarket again.
In today’s fast-paced information age, users are no longer willing to spend time going through countless search results to find the page that delivers the information they are looking for. Many people don’t even go beyond the first page of Google’s search results nowadays. As such, search engines are looking to provide results which are relevant not only to the keywords a user puts into the search engine, but also to the ‘why’ behind a search query: the search intent. In other words, search engine page results (SERPs) are optimised to understand what direct action a user wants to undertake through their search result (learn, purchase, find a specific website etc.) and prioritise the specific websites that match that intent.
Shifting from keywords to intent
As search engines become more advanced, incorporating more intent-based models and practices into research should be a key focus for digital marketers in 2020. However, intent research models can be quite subjective from a classification perspective as they rely on one person’s perspective to decide the user intentions behind a list of keywords. Moreover, the main types of search intent – informational, navigational, transactional and commercial – are very broad and, realistically, not very actionable.
For intent researches to be most effective, marketers need to have reliable data metrics, such as click-through rates and conversion rates to support the agreed intention behind a keyword. This allows them to create relevant lists of purchase and research intention keywords whilst ensuring the keywords they use for a specific search intent are the most relevant to their users. Checking SERPs statistics for keyword reliability over time also provides insight on which keywords are best to target for the specific type of intent.
There is still room for keywords
Understanding search intent, and taking it into account when delivering the most relevant answer is ultimately Google’s priority. While keywords are still a big part of search queries, digital marketers must understand that relying solely on keywords is not enough for SEO anymore. Backlinks and other traditional Google rankings are still important, but if the page doesn’t meet the user’s search intent, it’s not going to rank highly on SERPs.
However, that doesn’t mean that keywords are going to become obsolete. John Mueller, Senior Webmaster Trend Analyst at Google, agreed that keywords are always going to be helpful, even if they are not the main focus. Showing specific words to users makes it easier for them to understand what the page is about which, in turn, provides a better user experience.
Ultimately, optimising for user experience should be key in 2020 and shifting an SEO strategy to prioritise search intent is part of that. Focusing more on intent research and enforcing more intent-based practices off the back of keyword research is definitely something we’ll see more of in 2020.
By Adam Bunn
Director of SEO, Content & Digital PR