BERT and Google put customer intent first
There has been some exciting news over the past 24 hours, with Google advancing its search capabilities via an update called BERT (Bidirectional Encoder Representations from Transformers, to be exact).
While we find it incredibly exciting, we acknowledge everyone doesn’t. So here’s the TLTR version:
What does BERT do in a nutshell?
BERT’s focus is interpreting the intent of search queries better.
Instead of assessing your query on a word-by-word basis, BERT views the search term in its entirety – i.e. it analyses how prepositions like ‘for’ and ‘to’ impact the meaning.
This enables Google to better understand the true intent of the search – and then provide a more tailored answer.
What’s it mean for me/my content?
This will affect one in 10 web searches and drastically improve Google Assistant’s functionality to answer your long-tail questions.
This isn’t one of those algorithm changes built to penalise, but to improve the results people searching get served based on the intent of the query.
This change enables – and will encourage – people to search in a more conversational way.
It reaffirms the importance of having engaging, conversational content that answers the true intent of the search query explicitly. (And obviously, it needs to be properly optimised.)
The content’s holistic value is far more critical now – underpinned by keywords.
What is customer intent research, and will it help?
In a nutshell, yes it will unequivocally help.
Too many brands try to talk to people without their permission, about stuff they don’t care about.
Customer intent research lets you understand the actual outcome a person - or tribe - is seeking, and to craft your marketing around their objective, interests or needs.
This insight should then be used to power everything from your performance marketing, to content, to TV creative.
Here’s an excerpt from a recent article I wrote about how to understand customer intent:
“Search and social media are tremendous data sources to build tribes. Overlay this with your first-party data, and you have an unadulterated view of customer intent.
Search is my personal favourite as it is a data-driven truth serum – a look into what consumers really want at any given time.
You should define each keyword into ‘category clusters’ that allow you to pinpoint your marketing bullseye: a tribe of like-minded people, not a persona.”
You can read the full article detailing how to undertake this data-led diagnosis here.