Technique 4 minutes, 21 second read Mark Johnson, Editor, Just.Marketing
SEO experts are starting to get excited. There is an emerging technology called natural language processing (NLP) that could allow websites to be more creative in their use of language and use richer content to improve search traffic.
Although NLP is not entirely new, it is improving constantly. It is an artificial intelligence that extracts meaning from human language. Every time we shop online for a product, NLP gets better at understanding exactly what we are looking for – even if our search terms contain errors or phrases – by studying our natural use of language.
This has important implications for digital marketing and SEO.
Ben Garry, SEO Strategist at Impression, says that a combination of factors have created a major opportunity for marketers. These include: improvements in NLP, an academic paper written by Google that explains how Google ranks salience, and the release of a free, Cloud Natural Language API demo tool that could help marketers understand the effectiveness of their content.
Together, these developments mean marketers can now create content that is both helpful to the user and for Google search.
Ben says, “Google is trying to make its tools as close to humans as possible. Using their demo tool, you can get importance scores for nouns based on how Google sees or ranks them. This is Google telling us how it uses and views grammar and syntax. So you can create content that is now beneficial to readers but also SEO friendly”.
“It’s another tool to see how effective our keyword targeting actually is – you can use it as a sense tracker. It can tell us why something is not working as opposed to just guessing.”
The human touch
Nick Wilsdon, Search Product Owner, Vodafone, agrees and sees other benefits. He says, “The API is very useful because it allows us to understand how words and intent of words are understood by Google; whether it is a search for information or a purchasing enquiry and also the difference between ‘Apple’ the company and ‘apple’ a piece of fruit.”
NLP is also enabling Google to respond to search queries in a new way.
“Google is moving towards giving you the actual answer to your query rather than a web page. So to do it requires a better understanding of language and context,” says Nick.
He offers a word of advice on this point.
“Brands often use their own terms to describe their offer where online users will use different terms. A better understanding of language used in searches enables us to be closer to the searches carried out,” he says.
NLP is also improving entity recognition so that marketers need to consider the intent of the search more than ever, says JJ Grice, Organic Performance Director, Connective3.
“In theory, NLP should allow marketers to produce better and richer content without having to worry about the impact this has on SEO. One aspect of NLP is better entity recognition – which allows Google to understand whether an image or video supports the content surrounding it,” says JJ.
“If a user expects to see a collection of images or content in video format, marketers should think less about the keywords they want to optimise for, and instead optimise for the intent of the query.”
Deep enough understanding of sentiment?
Google is also getting better at understanding positive versus negative attitude in a text. Essentially, the tools enable it to see more context than previously. So it can better understand both the content of reviews, for example, and the context of links.
But this does come with a warning: these are early days.
Ben says, “On the evidence of the API, the sentiment analysis is nowhere near as good as human perception. We've done a bit of research into this ourselves and have found that Google struggles to categorise obviously opinionated news articles. Sentence-level sentiment analysis is okay, but as soon as it's stretched over a whole text it gets sketchy.”
Search in the sales funnel
With Google’s understanding of salience improving, search is increasingly being used at the top of the sales funnel. So consumers researching a wider topic are an opportunity for brands to appear if they are producing quality content with an authoritative voice combined with content that Google recognises is both broad and deep.
“Part of the score of entity salience using NLP algorithms will be if it appears alongside other language or phrases,” highlights Ben.
He uses the example of a senator being mentioned by name receiving a higher salience ranking if it appears alongside mentions of The White House.
He says, “The implication is that you need to know a lot about your topic and how Google will rank it.”
Strong content still rules
Although there are many reasons for being excited about NLP, some fundamental principles in SEO remain unchanged.
Ben says, “In the rank and search world, better and longer content will reward you with higher visibility – that’s still unchanged. And all content we use for search should have a benefit for users too.”
JJ agrees, saying, “The fundamentals remain the same – write for users, not for search engines. Don’t get hung up on optimising solely for that one keyword. Think about what other topics and keywords are used in conjunction with the subject matter being covered.”