Optimising for Natural Language Search
If you cast your mind way back to the days before Google (yes, those days did exist!) and Ask Jeeves was the go-to search engine, think how we used to structure our search terms. We had to phrase our search in the style of a real question. If we want to know the third longest river in the world now, we'd just enter “third longest river” and Google would do the rest, filling in the intent of the search with AI, and returning the results we want. Back in the last millennium we had to type in “what is the third longest river in the world” if we wanted to get the results we needed.
Search is now coming full circle, with what we call “natural language search”, and it's happening because of voice search. The way we type a search query and the way we utter the same request out loud are different so we'd not say “Alexa, third longest river”, instead we would say “Alexa, what is the third longest river in the world?”.
This is what we mean by natural language, it is the phrasing, vocabulary and syntax we use in conversation. People who own smart, voice controlled devices use them daily to find out information, and because the devices “speak” back we feel like there is a real conversation happening, reinforcing the use of natural language. In SEO terms, we call these natural language phrases “long-tail keywords”
Search engines use the context of the way the query is phrased to understand more about search intent, which helps inform the results of traditional text based search queries too, so both types help Google improve its results. Google doesn't just use keywords to determine intent, however, so optimising your content for this isn't a matter of adding new keywords or cutting old ones.
Because Google is always learning and refining it is able to tell whether your content meets the requirements of the searcher with more accuracy than before. BERT, (Bidirectional Encoder Representations from Transformers) was a 2019 addition by Google to the way it understands searches conducted in a more natural, conversational fashion. Context-defining words like “with” or “not” are now given equal weight in the query as nouns, verbs or adjectives leading to more precise results from long-tail search queries.
Google is now able to return Featured Snippet results to long-tail search queries with a stark degree of accuracy – as browsers we may not have even noticed this development because it simply addresses our needs straight away; no need for us to refine our search term input to get the results we want. We're likely to notice if Google was getting worse, but not so readily when it gets more intelligent.
Ensuring you have these long-tail keywords and phrases in your content is more important than simply including the keyword – you need to match your content to the intent of the searcher, whether that's to buy, research or compare. You're probably already doing this, marketing to people at different stages of the buying journey and with different needs. Your customer personas are the thing which informs the content you publish to get featured for these long-tail keyword searches, so optimising for natural language search may be something you're already doing without even realising.
Parua can help you keep on top of the natural search progression, ensuring you're up to date with how Google (which really leads the way for all search engines) looks for answers to the questions it is asked, whether that's out loud as with smart, voice-activated devices or through typed queries that follow the same patterns. If only Jeeves could see how far we have come!
