Optimising for voice search & digital assistants

Posted on 08/03/2017 by Ben Wood

This post will look at how search marketers can adapt and optimise for the rise of voice search and digital assistants.

how to optimise your website for voice search

More searches are now conducted on phones and tablets than on desktops. In addition to mobiles, marketers now need to consider the emergence of wearable tech such as smart watches, the huge rise of connected home voice assistants, such as Amazon Echo and Google Home, and the ever increasing range of voice search programs such as Siri and Cortana.

It is no surprise then that voice search usage is on the up – 40% of adults now use voice search once per day,  Google voice search queries in 2016 were up 35x compared to 2008, and a quarter of all searches on Bing are now voice searches.

This trend is only likely to increase in the future as search engines such as Google are better able to interpret complex conversational queries with the help of AI software such as RankBrain.

What is voice search?

Voice search technology uses a combination of Natural Language Processing (NLP) and Text-to-Speech (TTS) software in order to understand a voice search request. The request is then processed in a database which aims to match the question with the most appropriate answer the software can find – cutting out the search results seen on Google and Bing when conducting a text based search.

In its initial stages, voice search was pretty poor. Most of us have had very frustrating exchanges with a phone that simply doesn’t understand us. But such exchanges may be a thing of the past. In fact, the speech recognition error rate has reduced dramatically to just 8%. This is precisely why voice search is growing: it’s no longer odd to see someone carry out a voice search on their phone, as many of us recognise that a lot of the results delivered are actually better than natural searches.

Voice vs. typed search queries

There are four fundamental differences between traditional (typed) search queries and those conducted via voice search:

1. Query length

When typing in a query we tend to just search for simplistic keywords, as they are much quicker to type. For example, “mobile phone”. However, when we’re “speaking” to our phone, we tend to use natural language which results in longer search queries. So we’d be more likely to search for “best mobile phone in 2017” when using voice search.

Text vs Voice Search Query Length
Source: Moz – How Voice Search Will Change Digital Marketing

2. Questions

Again, as we are speaking, we use more natural language. And as we’re usually requesting information from our phone, we tend to use questions when using voice search. So the previous query would actually be more likely to be “what are the best smartphones in 2017”.

3. Stronger intent

When a user carries out a voice search they’re more likely to use natural language. This might be the most important difference of all, since natural language shows intent more strongly.

When you type “mobile phone”, your intent could be anything: are you looking to research mobile phones in order to buy one? Or are you looking to see how much yours is worth as you are trying to sell it?

However, “what are the best smartphones in 2017” clearly indicates that you’d like to do your research in order to eventually buy. Unless you really like reading about great phones for no particular reason.

4. Local impact

Mobile voice search is reportedly three times more likely to be location specific than text based search. We have recently seen studies claiming that searcher proximity is now the #1 local SEO ranking factor, so location signals are clearly going to be important for businesses looking to optimise for voice search.

google my business

How to optimise for voice search

Optimise location data

When a user turns to voice search to get information about physical places they want to go, the voice search software will refer to the users physical location to retrieve results.

To optimise for these location specific searches, you should:

  • Make sure thatthird-partyy sites such as Google My Business, Yelp etc. have up to date information (working hours, address, contact details etc.)
  • Make your Google My Business (GMB) profile as complete as possible, filling in all the other data attributes specific to your locations
  • Ensure that schema markup is implemented on your site
  • Incorporate readable feeds into your site – xml sitemaps and micro data such as location and prices in order for search engines to easily find and serve information to potential customers
  • Keep building high quality local citations in order to rank locally

Digital assistants are only as good as the data they have. This means making sure your location data is always updated and correct, as well as consistent across the web.

Use structured data

If you want digital assistants to serve your content for voice search queries, it’s crucial that you make it easy for them to gather and understand your page content. You can do this by implementing structured data markup (using schema.org vocabulary) on your website.

Schema mark up is designed to help search engines understand the context of your content, which means you’ll appear more relevant in specific queries made using voice search.

Focus on mobile

People use voice search almost exclusively on mobile. So if you’re focusing on voice search optimisation, mobile user experience and mobile performance (including page speed) should be  prioritised.

chrome dev tools mobile test
Tip: use Chrome Developer Tools to quickly test the usability of your site across devices

Research and target conversational queries

The majority of voice searches take the form of questions. While some digital assistants such as Amazon Alexa now allow users to make direct purchases, transactional search queries are far less common on voice search than informational queries.

voice search informational queries

Informational queries typically begin it with “Who,” “What,” “Where,” “When,” “Why,” and “How,” with searchers looking for answers that fulfill an immediate need.

While informational content won’t necessarily drive traffic to a product/service page, it will help you become a useful font of knowledge in responding to these queries.

There are a range of tools you can use to generate lists of informational search queries related to a specific topic. My favourite (free) tool to quickly and easily generate a tonne of informational content ideas is Answer the Public:

voice search answer the public

In the example above, taken from Answer the Public, we have identified a huge range of questions around our chosen topic (voice search), which could be used to generate content to answer the following questions:

  • How to use voice search on iPhone
  • How does voice search work?
  • How to voice search on Google maps

It’s these kinds of conversational, long tail keywords that you should be focusing your attention on if you want your content to be served for informational voice search queries.


Voice search is clearly on the rise and it’d be foolish for SEOs to ignore this trend. A huge chunk of voice search will be conversational, so start thinking about how to optimise for these searches now in order to gain first mover advantage.

Re-evaluating your content by focusing on natural phrases and conversational sentence structures, rather than keywords, will help boost your visibility in voice search, as will focusing on optimising location signals and mobile performance.

If anything, the steps that need to be taken in order to optimise a site for voice search are not too dissimilar from the steps you should be taking in terms of having a local, mobile focused SEO strategy.

Further reading

Digital assistants, conversational search and the future of local

How Voice Search Will Change Digital Marketing — For the Better

Are You Optimized for Voice Search? Here’s How to Do It

Enjoy this article?

Subscribe for weekly insights

Optimising for voice search & digital assistants

Let’s thrive together

Get in touch to take the first step.

Contact us