Voice search and personal voice assistants are changing the way we search for information online. Nearly 60% of searches are performed on a mobile device and 20%  of mobile search queries are voice queries. According to Google, by 2020 more than half of all queries will be done by voice search.

Optimizing for voice search is something that we can’t ignore anymore because mobile searches and the use of voice assistants keep increasing.

In order to optimize for voice search, we need to understand what makes voice search different from text search, the context in which someone might use voice search, the user’s intent, and how smart speakers work.

Then, we are ready to prepare our optimation strategy, which should include technical and content preparation.

Let’s start from the beginning.

What is voice search

Voice Search is a speech recognition technology that allows users to search using voice commands instead of typing them. It can be used on both desktop and mobile searches.

Top devices using voice search are:   Google Home, Google Assistant, Siri, Amazon Echo- Alexa, Microsoft Cortana.

Google Voice Search and similar technologies learn to recognize voice commands and keywords through natural language processing. Google and Siri can learn over time the patterns unique to the way you speak.

When Google Home answers a question with information from the web, it cites the source of the information by saying the website’s name and sending a link to the searcher’s Google Home app.

Google Home and Google Assistant read snippets from sites that are ranked in the “position zero” and have been granted a featured snippet. In a way, if we can optimize for featured snippets, we can optimize for voice search. This is the reason why we can find so many articles about “How to optimize your content for featured snippets”. But, of course, this is only a small part.

What makes voice searches different from text searches

  • The query length. Text is shorter and the most direct route to express our intent. Text searches are concentrated around one to three words. Voice search has longer queries (often containing 10 or more words) and tends to use a more natural and conversational language.
  • The use of question phrases. Voice search differs from text search in the usage of question phrases. Natural language means more question phrases. “Who”, “What”, “When”,”Where”, “How”, and “Why” are the most popular voice queries.
  • The user’s intent. Voice search shows intent more strongly than text search. When you add natural language in conversational search, the type of question asked can reveal the degree of intent. Questions beginning with “What” or “Who” show a need for information, and questions beginning with “Where” are used when someone is ready to take action.
  • The local value. Voice search has high local value (22% of voice queries are for local content).

How to optimize your site for voice search?

Voice search means that the queries are delivered in a natural and conversational language, and this changes the way we search. It also means that we need to consider the context in which someone might use voice search. Things like user’s location, time of the day, the device used, search history, similar searches by other users, user’s preferences, and user’s intent, give context to a search query.

Google has been making changes for a long time to better understand the users’ intent, context search, and conversational search.

Back in 2013, Google’s update called Hummingbird was made to determine the intent behind searches. Semantic search (the implied meaning of search queries) focuses on determining what the user really means and providing relevant results.

Semantic search tries to match appropriate SERP results to the language of users’ queries. It takes a broad context into account that goes beyond the meaning of individual keywords.

With Hummingbird and RankBrain, Google has improved contextual search and conversational search.

  • Contextual search. Google returns results that match the intent behind the query
  • Conversational search. Searches become more long-tailed and involve often whole sentences.

Now, years later and with the rise of voice search, Google is prepared to operate in a world of natural language thanks to semantic search and the Hummingbird update.

In order to optimize our site for voice search, we have to consider technical aspects and content preparation.

Preparing your content for voice search

  • Optimize for featured snippets. Featured snippets are the results that are read aloud in response to voice searches. They are often referred to as “position zero” and are the direct answer results.
    A featured snippet is the descriptive box at the top of Google’s results. The page description (the “snippet”) comes first, and then you have the URL of the page.
    According to Google, featured snippets are especially useful for mobile users or users searching by voice.

Some resources to learn how to optimize for featured snippets:

A reintroduction to Google’s featured snippets
How to Optimize for Google’s Featured Snippets to Build More Traffic
9 Tips for Ranking in Google’s Featured Snippets
How to Optimize Your Site for Google’s Featured Snippets
How to optimize for Google’s featured snippets
How to Optimize Your Content for Google’s Featured Snippet Box

  • Adopt a long-tail/conversational phrase approach when producing content
  • Start your keyword research searching for common and uncommon questions and create content that answers those questions.
  • Content must be easy to read aloud and make sense. The content display in a big table, for example, or a list of links may be difficult to read aloud and make sense to the listener.
  • Understand the context of the search query (location, time of the day, devise used, user’s intent,…)
  • Interpret user’s intent based on the kind of question phrase used. “What/who” are research phrases and “where” indicates that someone is ready to take action.
  • Whenever possible give a direct answer to the question asked and keep it short
  • Use structured data when possible because it makes content voice assistant friendly. It helps Google to understand your page and figure out what voice queries match that page. It is also important for local queries because it helps search engines to classify information related to your business (operational hours, contact info, address, directions), and have a strategy in place to gain positive reviews.
  • Claim your Google My Business listing if you don’t have it. For search queries like “near me”, Google relies on the user’s location and turns toward Google My Business listings. Make sure that NAP (Name Address Phone) is accurate.

Technical preparation for voice search

  • Be mobile friendly
  • Implement structured data
  • Improve the speed of your site since speed is one of the ranking factors.
  • Security. Search engines and digital assistants will prioritize a secure site over one that is not secure. Being secure requires technical work to go from HTTP to HTTPS
  • Develop Actions (Google) and Skills (Alexa). Actions and Skills are programs you build that enable you to create a curated space where your data is available for customers. The customer accesses your Action or Skill and asks questions which the system answers with your data. An example:
    • Action “Ok Google, talk to the Wall Street Journal” which reads news snippets from the Wall Street Journal.
    • Skill “Hey Alexa open Ambient Sounds: Thunderstorms Sounds”

Interesting articles about voice search

How to optimize for voice search
Type No More: How Voice Search is Going to Impact the SEO Landscape
How Voice Search Will Change Digital Marketing — For the Better
How to optimize your content for Q&As
The Evolution of Voice Search (and What it means for PPC)
Does Voice Search and/or Conversational Search Change SEO Tactics or Strategy? – Whiteboard Friday

Articles about structured data

The Beginner’s Guide to Structured Data for SEO: A Two-Part Series
The Beginner’s Guide to Structured Data for SEO: How to Implement It
The complete beginner’s guide to Schema.org markup
Introduction to Structured Data