atom thought
atom thought

A Complete Guide to Voice Search in 2018

02 Jan 2018
By atom42
an image of a home assistant

The data is undeniable: voice search is becoming more popular, with mass adoption of ‘home assistants’ continuing to increase through 2017. Below we explore how the trend for voice search looks moving into 2018 and beyond, how voice search impacts the way people search and what you can do to ensure you’re well placed to be front and centre for voice search results moving forwards.


The rise and rise of voice search
Different types of voice search
Understanding how people use voice search
Optimising for organic voice search
Optimising for paid voice search

The rise and rise of voice search

If data from Comscore is to be believed, 50% of all searches will be done without a screen by 2020, whilst latest data included in Mary Meeker’s annual Internet Trends reports show just how rapidly the install base is growing for voice activated home assistant devices like the Amazon Echo over the past 18 months:Chart showing Amazon Echo install base Whilst trend data shows how the term ‘Hey Google’ has increased across the past few years (including annual spikes as people unwrap new Google Homes for Christmas):

Google Trends graph showing increase in hey google queries

All of this points towards what we and others in the industry have been saying for a long time, that voice search is on the rise and to stay visible and relevant for the increases in searches, you need to make sure your site is well optimised for voice in 2018 and beyond.

Below we’ve outlined some ways to ensure your digital marketing efforts across both organic and paid search are geared towards voice search, and that your site appears when users search.

Understanding the different types of voice search by device

It’s important to define what we mean when we’re talking about voice search, as different use cases will impact your optimisation in different ways. Broadly speaking, we can currently break up voice searches into the following 3 categories:

  1. Voice searches performed within a search engine themselves
  2. Voice commands on a mobile device that trigger a knowledge result
  3. Hands-free voice commands on a connected home device

Why is this distinction important? Firstly, the types of searches people make on each device will likely vary and it’s important to think about what users might want from voice search to ensure your optimisation is geared towards this (more on this below).

Secondly, only option one currently delivers a paid search result. When using an ‘Okay, Google’ command for instance, you aren’t served any paid ads. For the connected home device, there’s no screen to display ads on (more on that later), so whilst these can still be optimised for with organic work, paid visibility is much more limited (for now).

Search is increasingly monetised over time (as we saw with Shopping ads moving from free to paid some years ago) so it’s perfectly plausible that other forms of voice search or device specific searches  will eventually start showing ads too. In the short term, being optimised for voice searches within the search engine is achievable, and we should be ready for when voice search ads show more widely too.

As technology changes, so too does search behaviour. Google has previously released data that showed how ‘micro-moments’ have been born out of people understanding how their technology can help their search experience, in turn introducing a whole new breed of search terms that focus on locality and relevancy to their immediate need and location:
Image showing near me growth in 2014

This trend in 2014-2015 was a direct result of mobile devices that made searching on the go and ‘in the moment’ possible, and in turn, changed how people search for things. Interestingly, the latest data from Google shows that the above trend is slowing, not because people don’t have the same intent, but because they’ve adjusted their search behaviour as a result of understanding they don’t need to include ‘near me’ to still get localised results.

When typing out a search query, people generally search in a way that they believe a search engine will understand, to find a particular answer. With voice search, people interact more naturally, and in a conversational way.

Generally questions are asked, rather than making statements. For example:

  • Typed search: book hotel london bridge
  • Voice search: What hotels are available in London Bridge this weekend?

There is often a very strong local element to searches performed via voice as well. For example, someone might ask “Okay, Google. Which vegan cafés are open near me right now?”

The combination of question + local is the crux of many voice searches.

Unsurprisingly then, voice search is already having the same effect on larger shifts in the way people search, and we’re beginning to see an increase in a whole new breed of longer-tail and conversational style terms that otherwise wouldn’t have been used on a desktop or mobile device:

Image showing rise in voice based queries in 2017

Many of these are immediate information like the weather or traffic which Google will handle, but there’s a need for marketers to think about what kinds of voice queries might work for their brands and to ensure they have content optimise for this purpose, whether that’s on-site content, or ensuring that all local listings for branches have the correct store hours and address details.

Finding out what questions people are asking

Understanding how people are using voice search is only one part of the puzzle, and you need an effective way of finding out what questions people are asking to ensure you can be front and centre when they do with the right content, and with targeted keywords in paid campaigns.

You can do this in a number ways, including:

  • Analysing Google Analytics site search queries
  • Analysing Google Search Console analytics
  • Reviewing your AdWords SQRs for any existing question words triggering keywords
    • These can be useful for PPC but also for organic optimisation and content creation
  • Looking at social media and forums to see what people are asking
  • Surveying your customers

You should also consider which question words are being used, and prioritise them in order of commercial value, to determine the bids you want to set. For example, the question word “where” often implies that someone is closer to converting than “what”, and as such can be bid on more aggressively in PPC.

Optimising for Organic Voice Search

Optimise for featured snippets

When Google answers a query on Google Home it’ll state where the website or source of the information and can provide a push notification on your phone. However, unlike a text query on desktop or mobile that returns results for a user to select, voice search doesn’t give users an option, instead using the result located in position zero as a featured snippet.

Therefore, optimising your content to appear in featured snippets has a two-fold effect. You’ll appear top ahead of competitors on desktop and mobile, even if you don’t rank in #1; and in terms of voice, you’ll be the source of information used by Google and the result pushed through to a users phone.

The result of a #1 or #0 ranking has always been powerful, but with no screen (and no choice) being in the featured snippet #0 for voice offers dominance over competitors like never before.

The good news is that optimising for a featured snippet needn’t be too much extra work, and if you already rank on page one for a term or topic, you’ve already done a lot of work. The following steps can help get you where you need to be:

  • Establish which of your current rankings have a featured snippet on their SERP using a tool like Ahrefs site explorer
  • Use the methods mentioned above to find more examples and ideas for relevant questions and terms that you could be appearing for in a featured snippet
  • Optimise the on-page content so that you’re clearly answering questions in a concise manner. Your content should still read naturally, but an optimised question header followed by a concise answer paragraph will make it easier for Google to pull into a featured snippet

Integrating question terms into your account

You’ll want to set up specific question-focused ad groups in your account, so that question queries trigger ad text specific to the question, leading to the most relevant landing page.

If question keywords set on exact match have low search volume you’ll need to use broad match modified terms to capture these queries. Group together similar queries in the same ad groups, with bespoke ad text, and frequently review the SQRs to see what questions are triggering your ads. Add negative keywords to stop irrelevant queries from triggering your ads.

Monitor the conversion funnel and assisted conversions

While bidding on question terms in your PPC account is likely to bring in more volume, you’re also likely to be drawing in interest during the research phase. This may mean that CPA for these question keywords appears high (however, their CPCs are likely to be lower), but it’ll be important to monitor their impact further down the funnel. By bringing in more people at the research phase, they have earlier exposure to your brand, and you have the chance to remarket to them to get them to convert with you later on.


Voice search is still a relatively new area for digital marketers, but it’s never too soon to be optimising your organic and paid marketing to ensure you’re ready for what comes next. The most important step is to take the time to really discover and understand what it is your audience is looking for, and to ensure your content and campaigns are created and optimised in a manner that makes discovery by search engines as easy and digestible as possible.

If you’re looking to capitalise on voice search and deliver results in 2018, get in touch with us to discuss your needs and how we can help.

an image of a home assistant

Related Articles

Contact Us
Contact Us

Get in touch

Visit us at:

10 Queen Street Place
Thames Exchange