Binary and beyond - Using natural language processing in a successful SEO strategy

Digital Strategy

Language, content, search marketing – the three go hand in hand and without one you can’t have the others. The search marketing aspect is the user generated query and in order to generate an answer search engines require content to link to

This content, in turn, relies on language. Sounds simple right? In truth it’s pretty complicated and with natural language processing sure to be one of the most talked about items in the industry we thought we’d look a little deeper.

What is natural language processing (NLP)?

Natural language processing (NLP) is an area of computer science and artificial intelligence concerned with the interactions between computers and human (natural) languages, in particular how to program computers to process and analyse large amounts of natural language data.

Challenges in natural language processing frequently involve speech recognition, natural language understanding, and natural language generation.

In its simplest form it is the act of asking a search engine a question in the manner you would ask a friend. It is conversational and a good example comes from the very header of this section. When trying to find out ‘what is natural language processing’ you would almost certainly type in that sentence not ‘natural language search explanation’.

The evolution of NLP

Whilst NLP is featuring more and more in search marketing and SEO, it isn’t exactly new. The way that computers process natural language has been around for a long time. I think the reason that it’s picking up speed is as a direct result of the advancement in technology and the way in which people search, with voice search in particular picking up traction.

Other factors that make a difference as to the rate of progression in terms of NLP are:

Increased interest in human-to-machine communications Availability of big data More powerful computing Enhanced algorithms.

Search algorithms change greatly, they adapt and evolve in order to make sure that the user query is met with the best possible answer. Early adopters of the internet will have used program such as Ask Jeeves. They will have entered questions in the hope of gaining an answer which best suited their needs. Whilst this worked for a while, updates meant that Ask Jeeves (or similar tools) could no longer compete with the keyword based search engines such as Google.

Natural language search does still have a place especially with the rise of voice search, this we know and it is one of the most hotly talked about topics in the industry, but why has there been a shift towards keyword searches?

Well the answer is simple, as Google tinkered with their algorithm and brought out many updates, especially the hummingbird update; they did so knowing that there will be a focus on user intent and contextual relevance.

They wanted to provide the user with an algorithm that tried to show results based on exactly what you mean not just by picking out the keywords within a query. The main reason that they did this so that they were able to understand longer more complex queries and so they can identify other possible searches which might be relevant for the user.

Understand your audience

Without doubt this is the single most important thing for search engines today if they are to service a query with an answer. They must understand exactly what the audience are looking for in the same way that business must understand how to position their content for that users search.

Businesses must conduct keyword research and competitor analysis to ensure that their content is relevant in the correct format and most importantly that it resonates with the end or target user.

Businesses should be aware of all users no matter how well they are grouped in terms of their demographics. They are all different and for that reason it is essential that you take into consideration that they might search in a different way despite the fact that they are in your core user group.

One example could be that a user types in ‘Buy shoes near me” and another could look for “local shop shops”. Natural language processing will help a business to display for both of these despite the difference.

It is highly unlikely that two users will use exactly the same keywords or search in exactly the same order when they type a query out. This is especially true when it comes to voice search.

Now one of the big point here is that it doesn't mean keywords aren't important but if you want to have the best chance of getting found by voice you have to make sure that you understand what your audience actually means not just what they are saying.

One of the best ways to do this is to get a collective group of terms which a person may ask about your business if they were to meet you face-to-face. I.e. this is the natural language style of search and offers you a chance to get conversation with the audience. One of the best tools that we have seen is, this is an interactive tool that lets the user or the person using the program should I say, the chance to create a word wheel.

This wheel, at its centre has the key words with which you want to associate with your business. Branching out from this are numerous number of possible questions the majority of which will be conversational. If we are to do a quick search for one of our key terms (search marketing) we get the following word wheel.

You will notice that when you look from the centre of the wheel outwards the first set of words that you get to let you know the type of question with which the user may ask, these can be things like where, which, who, when, why, how etc. Once you have all of these questions it is worthless unless you actually implement the strategy with which they can be used.

Take your keywords and those associated with the word wheel and see how you can optimise pages on your website, blog, and on page content to include terms that will be picked up by these queries.

Make sure that you incorporate all of your answers onto your webpage. These can be broken out across the site map making sure the correct information is on the correct pages.

You can also group together questions and many businesses use a FAQ page to do this. This plays right into the hands of voice search and can give the user information they need quickly and efficiently.

The beauty of FAQ pages is that there is little or no cost other than the initial investment in terms of research and producing the content. These pages answer the most commonly asked questions of your business and increasingly we're seeing Google account things such as opening times all the address of the business within these brackets.

For those businesses that want to improve their brand image and go into more detail about the products services they offer they can do this through means of a blog or through more long-form content. If this content is keyword rich that has a greater chance of being picked up. Keyword stuffing is something we've looked in the past and whilst many wouldn't advise using the same keyword over and over again, this isn't something to worry about so long as you are answering the question asked correctly.

It is important that you create all of your content conversationally and a long-tail keywords are taken into account.

Businesses should take an SEO style long-tail keyword and turn it into a natural sounding question. Previously businesses might've targeted the term ‘shoe shop Chester’. This will have given the user a list of businesses that sell shoes in Chester and therefore fits the brief.

The problem with this is that as we move towards natural language searching this type of query could differ depending on if it is spoken or written.

The spoken search could be something like, ‘where can I buy shoes in Chester?’. Again this solves the query but actually opens up your business as the response it tells the user your location the products you sell and what you do. Typically as can be seen by the word wheel, the best sort of questions and long tail keywords of those that start with the who, what, where, when or why prefixes.

One of the best ways to use natural language search is when it comes to the use of chat bots and follow on questions. Google recently announced that they were implementing a system by which conversations could be struck off with AI through the process of machine learning.

This is such a breakthrough for the industry as it saves time and means that users can get right down to the heart of the problem or query by asking very few questions. If you think of the game 21 questions - where is possible to get to a solution through a series of yes and no questions in just 21 turns, then you're on the right track. Again if we take the shoe shop analogy it would be possible for a user to ask a virtual, voice automated assistant the following question. ‘Where is my nearest shoe shop?’.

Google will then answer the query for the user and the user can then ask a secondary question in a conversational tone such as, ‘how do I get there?’. These questions will not only pull from resources such as Google, but will also use other tools such as maps and other visuals if the user is looking on a device.

Something that we often look at here is the use of schema mark-up, this is where you allow Google to pull other information from your site telling the user what sort of products or services you may offer and additional business-related questions.

It may be as a business that you don't accept credit cards if you are a cash only or you can include opening times if you're a restaurant as well as the type of cuisine or perhaps whether you offer delivery or not.

Adding schema is a must as this mark-up could be the difference between winning or losing custom online and appearing for a query.

Initial reports suggested that only 20% of all websites were using schema mark-up but this has picked up rapidly over the last few years. If you're failing to use schema mark-up then you're more likely to miss out on conversions later down the line.

In essence this is probably the quickest way in to any business can have and is one of the easiest tasks to complete as it is essentially profile building.

The vast majority of mobile and voice searches are likely to be location based so it is imperative the businesses optimise for Google my Business. Adding images, questions and reviews can also boost your position in SERPS, and is also worth listing your business on other directories so that Google can find it in as many places as possible.

Implementing the points in this knowledge base will allow you to become fully optimised for voice ad natural language and can propel your search performance.

For the latest in search marketing news and views, sign up to our blog – or for actionable insights, see our ever growing collection of resources. Alternatively, contact us to see what we can do for your brand.