From mobile to vocal first: is your site already optimised for voice search?

Voice search: we like it, Google likes it

Voice search is a voice recognition technology that allows one to search by giving a command with one’s voice. Launched by Google back in 2009, it is gaining more and more popularity: after all, with a life extremely geared towards multitasking, it is undoubtedly convenient to ask Google, Alexa or Amazon Echo – just to name a few voice assistants – without having to stop and type!

Some data from the US market:

• Almost 50 per cent of people use voice search when searching for products(Social Media Today, 2016).
  Amazon sold 4.4 million Echos in its first year(Geek Wire, 2016).
• 65% of those who own Amazon Echo or Google Home cannot imagine going back to when they did not own a smart speaker(GeoMarketing, 2017).
• Sales from voice search reached $1.8 billion in 2016 and are estimated to reach $40 billion by 2022 (OC&C Strategy Consultants, 2017).
  72% of people who own a smart speaker say they use it on a daily basis(Google, 2018).

In Italy, the figures are slightly lower but growing strongly: even Italian users are finding it increasingly convenient to carry out voice searches through a device such as a smartphone, which is always at hand.
This is why voice search today integrates numerous services and applications, and Google itself is shaping up to handle the various types of queries (the words we use to perform an online search) differently, and to respond to new user habits by interpreting voice queries in the best possible way.

How voice searches are changing

For quite some time now, it is no longer the users who imitate the language of the machine, but Google that refines its algorithms to respond to increasingly complex queries: first with Hummingbird, an algorithm capable of deeply understanding user intent, and then with RankBrain – a machine learning system based on AI – that understands the semantics of the entire sentence and, through the substitution of terms and synonyms, returns results based on the context inferable from the query itself.

Now, with voice searches, language becomes even more discursive and ‘natural’, and the use of words and syntax is markedly different from written language: voice queries are less mechanical, with less use of single words in favour of long-tail keywords and phrases that reflect everyday speech.

SEO strategies for a ‘voice friendly’ site

In light of what has been said so far, it is quite clear that the time to have a site optimised for voice search is getting closer and closer. But where does one start? What are the elements to work on for voice search optimisation? Spoiler: content strategy, structured data, optimisation for mobile and local search.

Write like they talk: conversational language and long tail keywords

41% of voice search users say they interface with devices as if talking to a friend(Google, 2018). These people talk to their smart speakers as if they were human beings, so it is no longer a question of keyword searches, but of real queries: the first step in optimising one’s site for voice search is therefore precisely to identify long-tail keywords that reflect natural (conversational) language, i.e. the language most likely to be used by the target audience.

Featured Snippet: the only answer to the voice query

Not only that: the snippets (title and description appearing on the search results page) will also have to be optimised accordingly, hoping that Google will deem them worthy of being elevated to Featured Snippet status, i.e. results that respond so well to a query that they are highlighted at the top of the SERP, between the paid ads and the first organic result.
Now, the most interesting aspect as far as voice search is concerned is that 40.7 per cent of voice search answers come from Featured Snippets(Backlinko, 2018): Google in fact prefers to respond to voice search with only one answer and – lo and behold – it is a Featured Snippet!

Answering content: FAQ

Another unequivocally strategic move is to create content that answers real questions: FAQ pages, in particular, are very effective for mobile, because they contain the search query in the form of a question and a short answer. Lacking inspiration? To find new ideas for frequently asked questions check the reviews: they are already written in natural language and are real requests from your target group.
Last but not least: update your content often; answer any existing questions in the ‘Questions and Answers’ section of Google My Business; create interesting questions about your products and services that promote conversion.

Give Google a hand: optimise structured data

With voice search you only get back (when possible) the first result in the SERP: the aforementioned Featured Snippet, the result that most accurately and comprehensively answers the user’s search query.
Therefore, in addition to making the content as relevant as possible for the specific query and making sure that the pages are easy to scan, it is useful to help the Google bot better understand the content of the site: we can do this by using structured data markup (schema.org).

Structured data markup – besides being a good SEO practice for highlighting certain page elements in SERPs – helps to create metadata for content, telling Google what it is about without affecting user viewing.

Give your content a voice with Google actions and Alexa skills

Have you ever heard of Google Actions or Alexa Skills? Creating a Google Action or Alexa skill is another way to make your site content more accessible from smart speakers, connect your app to Google Assistant (or Alexa) and offer a voice response to users’ voice searches.
Don’t have an app? Google automatically creates Actions based on your articles, generating a dedicated page directly in the Assistant directory and, when requests via Google Assistant are compatible with your content, Google will ask users if they want to read or listen to your news.

The real novelty launched by schema.org and Google, however, is Speakable, a markup that selects the most suitable parts of text to be read aloud by devices using Google Assistant: at present, Speakable is only widespread in the US and is mainly used to read news, but tomorrow… who knows!

Mobile Friendly is Voice Friendly

Fully operational since 2021, Mobile First Indexing is an algorithm that rewards mobile-optimised sites by indexing them first. Will there be an update to the Vocal First Index? We cannot know, but it would not be that surprising considering the benefits of voice searches and the growing trend.

Optimisation for the mobile experience and optimisation for voice search are not mutually exclusive, on the contrary: the majority of voice searches come from mobile devices and many of the activities that need to be done to optimise the website for vocal search are also useful for mobile-first optimisation.
Backlinko, in an analysis of potential ranking factors for voice search, found that voice search results load 52% faster than average pages: a plus point for ranking in the SERPs.

Never without it again: Google My Business and local search

Voice searches, especially from mobile, are three times more likely to refer to geographical location than text searches (Search Engine Watch). In fact, people usually use the browser or voice search on their smartphone to obtain information on local businesses they intend to visit (e.g. exact location, opening hours, products available…).
You know what that means? That the more you optimise your content to make it appear in local search results, the more you will have optimised your site for both types of queries, text or voice.
Start by creating or claiming your business’ Google My Business tab, then develop a content strategy designed to easily answer location-based queries.

With this overview of the strategies you can put on your to-do-list to optimise your site, you can jump straight into action: you won’t need to ask Google how to have a voice-friendly online presence – which would sound more or less like ‘Hey Google, how can I optimise my site for voice search?