Just a few years ago, voice-activated technology sounded like science fiction, it reminded us of Star Trek‘s iconic characters commanding their computers with just their voices. Now it’s 2017 and voice search is real and going to stay: in 2016 Google voice search has been used 35 times more than in 2008, according to Google Trends. What is even more impressive is that 60% of people who frequently use voice search have started only last year – the data was gathered by MindMeld, who ran a survey on the adoption of voice assistants in year 2016. We are facing the adoption of a fast-growing technology which will reshape the way we experience the web.
Voice search became more and more sophisticated in the past few years, and its latest developments have brought it in our pockets in the form of virtual assistants like Google Assistant, Siri and Cortana, and in our homes in the form of home assistants such as Amazon Echo, which was unpacked by a lot of American families last Christmas, and the newborn Google Home.
Why are Internet users adopting voice search?
- It finally works! The error rate in speech recognition is getting close to zero. Andrew Ng has long predicted that as speech recognition goes from 95% accurate to 99% accurate, it will become a primary way that we interact with computers. This 4% accuracy gap is the difference between annoyingly unreliable and incredibly useful. Thanks to deep learning, we’re finally cresting that peak.
- It’s fast and simple. Most people speak faster than they type. That’s a fact. So, when you are in a hurry, it would be much easier to speak to a virtual assistant and get the answer you were looking for straight away. Also, when searching the web by typing a query you have to evaluate the results and pick the one that comes from the most relevant source (probably between those in first SERP), if you are not satisfied with the results, you will have to narrow down your research by adding more keywords to the query. It requires some time and experience. Since virtual assistants are becoming more and more conversational, the voice search process is now simpler and more user-friendly: it resembles a natural conversation where the virtual assistant keeps track of previous questions and replies accordingly.
- It’s multitasking-friendly. Say you are at home cooking an apple pie and you just need to know how many eggs you need or you are driving to the venue of some event which is about to begin and you are not sure about the fastest route. In both cases, voice search can help you giving the information you need while you are busy doing something else and without using your hands!
- It’s mobile-friendly. Anywhere, anytime, our virtual assistants are in our pockets, just a voice command away. Most times, thanks to instant answers, you don’t even need to read something on your device, you just get the right answer… in zero time!
A 5 steps checklist to write effective content for voice search
1. Answer questions (and do it accurately)
Many voice searches are questions which start with the 5W + H (in case you don’t remember from your school days they are Who, What, When, Where, Why and How). Usually, the answers to questions which start with Who, What, When and Where are a single piece of information that can be isolated: that is why Google and Intelligent Agents tend to answer this kind of questions with instant answers. Owning or being the result picked as instant answer is a strategic asset (the so-called Rank Zero case).
WordLift offers you the chance to structure your content assigning a role to each piece of information using the Who, What, When and Where categories.
2. Make your content easy to read
Organize the content of your page with subtitles, bullet points, tables, and some highlights: it will help machines understand the content on your page. Also, remember that many people tend to quickly scan a page before actually reading it: a neat content organization of your page will help your hasty readers and give them a reason to thoroughly read your article.
3. Optimize for mobile and local search
Most times, voice search comes from mobile devices. Keep it in mind and consider that a mobile friendly website is crucial for your SEO. Let’s take a step forward and think about how proximity could eventually change the relevance of your content. If you own a local activity, it probably does: think about how your products or services may help people around you and structure your content accordingly.
Think about the needs of people walking nearby your office and feed them with useful information that relates to your business.
4. Add schema.org markup
Schema.org markup helps you adding a context to your content: you can use it to explicit the relations between information in your content, to disambiguate the words you are using and to give structure to your pages. When using schema.org markup you are communicating with machines through semantics instead of keywords: you are helping machines to define the context around an information so that it will turn into useful results for humans.
5. Embrace everyday language
Do you think that average users talk to their virtual assistants with complicated, hyper-specialized words? Wrong! They will most likely use everyday language or even slang to ask questions. So, more than ever, remember the KISS principle: “keep it simple, stupid!“. Ask yourself what would users who know nothing or nearly nothing about the topic you are writing about would search, and then write for them.
Einstein used to say: “If you can’t explain it simply you don’t understand it well enough“.
Moving from keywords to semantics
The rise of voice search trend is just the tip of the iceberg: search engines are slightly moving from keywords to semantics and natural language processing is heading towards human-centric results. It’s not just happening on voice search, it’s true for any kind if search, especially on Google.
Since fall 2013, thanks to the algorithm known as Google Hummingbird, SERPs have been more and more populated with results which take into account context information such as the kind of device used, geolocation, time of the day, previous searches and more. To give each user the best results, Google is also using a machine learning technology called RankBrain, which is the third signal in order of importance contributing to the results of each search. RankBrain is used mainly to interpret the searches in order to find pages that offer a good answer for users, but might not have the exact words that were searched for. It clearly relies on semantics since it goes beyond keywords and gets deeper into the contextual meaning of the search.
When using context information, search engines serves more relevant results to users, picking the answers which seem to be more accurate and case-specific. Considering the biggest picture, voice search is one more reason for content creators to focus on content structure and rethink each page or article as an answer for specific personas.
WordLift gives content marketers the opportunity to focus on writing while it does the techy job and gives structure to your article with metadata.