How to optimise for voice search – SEO best practice can prepare you for the future of search

Jan 25th, 2017

There is little doubting that voice search will one day overtake that entered in to a search bar, but even at this stage – with Google reporting that one fifth of searches are currently spoken – it is well worth implementing the necessary changes now


Voice search has been a feature of popular culture since at least the 1950s science-fiction golden age – with authors such as Isaac Asimov, P.K. Dick, Robert Heinlein and many others all featuring spoken interaction with (admittedly wildly inaccurate depictions of) computers.

Though many aspects of popular culture’s depiction of the future have changed since the fifties (Isaac Asimov was the first to employ the Slavic word ‘robot’ as it is currently used, but failed to predict miniaturisation, therefore describing planet sized super computers) one thing has remained constant in popular culture and that is the future prevalence of spoken interaction with AIs.


While the 20% figure may not seem that large – considering that mobile search was at a similar level as recently as 2011, it doesn’t require much of a leap to see a similar trend occurring over the next five years with voice search versus text search.

In fact, the reason for believing in voice search’s future prevalence is identical to the reason mobile overtook desktop – the growing utility of technology. 2010, for example, was the year of the original Samsung Galaxy S and the iPhone 4 – the first phones that (for me, at least) really promised the possibility of a mobile UX on a par with that of desktop.

Similarly, the last year has seen a similar level of technological advancement, with accuracy of voice search exceeding 90%, meaning that the ‘I didn’t catch that’ response is almost a thing of the past. The speed and functionality of voice search has also greatly improved (as you can see by comparing these two videos here and here) meaning that it is easier and more accurate than ever to search on mobile. Add to this the self-perpetuating nature of tech development (the better something performs, the more is invested in refining it, meaning it performs better, meaning… and so on), the more spoken language search seems a certain bet for the future.

In terms of search marketing, this has a number of possible implications – not only does it further reinforce the position of mobile as the go-to platform for search, it also adds weight to the importance of latent semantic indexing (LSI) and natural language search as well as that of local and ‘near me’ searches (which, Google states, have increased 3400% since 2011).

So, if we accept even the chance of voice search following a similar trajectory, then it makes sense to begin optimising now – especially your evergreen content.

There are two ways which are likely to make the most impact short to mid-term for your search marketing strategies:

1. Natural language

While this is, most likely, the second in order of added value of the two must have changes to optimisation, it is also the simplest and least time consuming of the two to implement. The key to this is, really, to attempt to forget a lot of the lessons learned about content optimisation the previous decade has taught us.

While, at the time, the keyword focusing of content made sense from a utilitarian perspective – insofar as it catered to the peculiarities of search engines – even the best writers were forced to work hard to produce content that was both useful, well-written and sufficiently focused on the specific keyword or term (while some other writers ended up producing stilted and repetitive nonsense).

The advancement of voice search and the tremendous improvements in the vocabulary of search engines can (and should) be incredibly liberating to content producers across all industries. While there may be industry-specific terms which have no useable synonym, for the most part the content produced across all sectors is freer to be itself than ever before.

Much of the advice remains the same as for traditional SEO – though the approach differs. The requirement for content providers to produce useful content which seeks to answer consumer questions, or is in some other way relevant and engaging, remains, but the ways in which such content are being found has subtly changed (there is a handy list of all OK Google commands here).


Listening is precisely what you should be doing – the way people phrase questions for voice search is all around you (unless you happen to work from home), so keep an ear out for how the people around you are asking each other (or their phones and digital assistants) and look to replicate the formulation of these questions as jumping off point for your content.

It doesn’t end there, however – as stated above, content providers are now uniquely free in the search marketing industry as:

  • LSI infers meaning from surrounding words allowing it to derive the obvious (to humans) synonymous nature of ‘business’ and ‘corporate’, for example, due to context in the many occasions the words are encountered in the same company.
  • RankBrain functions to predict and evaluate content for previously unsearched for terms.
  • The overall increase of vocabulary provided by both.

This means that writers can approach content without having to consciously optimise for specific key terms, allowing them to write with the sole aims of reader enjoyment and the best way to communicate the information.

2. Markup

Markup is the key to succeeding not only in voice search, but also in local, Google’s Knowledge Graph and in mobile search as a whole. The main drawback, however, is that if it has not been implemented early on, it can be difficult and time consuming to do at later stages. However, with markup serving as essential sign-posts for search engines, it is more than worth the time it may take.

There are a few tools online that can help, however – with JSON-LD and microdata generators available here, here, and here, while Google provides its own assistance here.


In essence, what these tools seek to do is to annotate various types of ‘thing’ (for want of a better term), so that its nature is clearly communicated to search engines, reducing the work it needs to do to categorise the content and also the chance of it misidentifying it or missing it entirely.

A number of examples can be found schema.org, for example – a track listing for Radiohead’s ‘King of Limbs’ album is depicted like this:


Markup works in the same way in all its variations – so think of its use for all areas of your site and all content in much the same way. Just as you want your brand’s phone number and address to be easily accessible to search engines for inclusion in local search, so too do you want it to be easily able to recognise when you are answering a question or communicating a particular piece of salient information for your consumers.

While it can take a long time to implement this on an item by item basis, one of the best tools to take some of the sting out of this process is again provided by Google. The Google Tag Manager is able to dynamically generate schema/JSON-LD tags and, while it is less than perfect, there was a fantastic post on the Moz blog which goes through the process of overcoming some of the teething problems with its use.


code-image


It is now apparent to the point of incontrovertible that, barring a sea-change in the tech world, that mobile is the future of the internet and, therefore, of search marketing. While best practice may change over time to reflect this (as coding languages adapt to mobile-first priority and the speed and utility of devices improve), it is absolutely vital to ensure that your brand is riding the crest of the mobile wave.


Click Consult is an award winning search marketing agency committed to leading and communicating industry best practice, to find out what Click can do for your brand – contact us today.

Facebook Twitter Instagram Linkedin Youtube