Voice searches are growing in popularity but they are still used for basic tasks
The amount of investment from tech giants shows they are planning further growth
Most publisher content is not suited to how people use voice searches at the moment
Learn to optimize for voice now to be ahead of the game later
Voice search and read-aloud content formats will become popular
For everything that gets accepted by the odd machine of public opinion, there are New Cokes, Betamaxes, and a Google Glass wondering where exactly they went wrong. And, for a while, people weren't sure which side they were going to fall on voice search.
Speaking out loud to an AI voice and inviting privacy-casual data miners inside your house felt ... odd. It seemed voice search might sink before it had a chance to be perfected.
But, for major tech companies, voice search was too valuable to let fail. Despite user hesitations about privacy, how well it understood our speech, and how strange it felt to drip orders into an anthropomorphized Bluetooth speaker, Alexa, Siri, Google assistant, have all seemed to have pushed through the failure barrier and into the golden circle of accepted household tech.
According to Google, 20% of all searches are now done by voice. Thanks to constant updates, low initial price points, and advances in AI and natural language processing, voice search went from a novelty with hit and miss results, to being a real part of people's online routines.
Could voice search overtake typing?
Around 2018, there was a statistic making waves in SEO circles that said 50% of all searches will be conducted via voice search in 2020. If you're thinking that seems a lot, you'd be right. This statistic was debunked.
But, what was revealed to be true was that 48% of users, surveyed by Adobe, will use voice search for general search queries, with the majority of respondents saying they use voice commands at least once per week. By 2019, there were more than 100 million virtual assistants in people's homes, alongside a billion voice-enabled smartphones.
While voice search is not overtaking regular search results yet, it's also clear that it's not going away. The earlier publishers can connect themselves to this new routine, the more likely they will be to benefit as it continues to grow.
What do people use voice search for?
If your content is well optimized for SERPs, you might be thinking that you're already prepared for voice search, after all it just a different way of asking the same thing. Right?
The way people use voice searches is different from typed searches. The words they choose and the results they need have a very marked difference. To get any value from voice searches, publishers need to know what users want from the voice search experience.
At the moment, it's fast answers to short questions. Looking at user behavior, you can see that voice search is used to save a few clicks when people are occupied.
They don't use voice search as a replacement for typing, they use it for different tasks.
Directions while driving - 52%
Making a phone call - 51%
Sending a text - 50%
Checking the weather - 49%
Playing music - 49%
If we look at the breakdown of actual search queries you can see that people are searching for facts and easy answers — not opinion or entertainment.
Looking up measurement conversions.
Asking for a sports score.
Checking traffic alerts.
Reviewing news headlines.
Performing a calculation.
At the moment, people use voice search for the things they know it will get right. Without a screen or interface to select choices from, people only want the binary options and clear unmistakable information delivered back to them.
A Statista report shows that accuracy and recognition are still the biggest barriers to the adoption of voice tech.
People will ask Siri what 30 centimeters is in inches, they won't ask Siri how the agricultural policy of China has progressed over the last decade, yet.
How do I optimize my content for voice search?
While voice search may not deliver your most engaged visitors, a successful search for basic information can be the start of a deeper rabbit hole of secondary content.
And, as voice search and voice results become more sophisticated, it will benefit publishers to know how to optimize their content for voice search and build established channels with providers of Voice User Interfaces such as Siri, Alexa, and Google Assistant.
Here are the ways you can ensure that your content stands the best chance of being featured in voice search results.
Have fast page speeds
Google never wants users to wait. If your Time to First Byte is slow, you are less likely to feature in voice search results.
A Backlinio study into 10,000 Google Home searches showed that voice search result pages load in an average time of 4.6 seconds. This is 52% faster than the average page, a clear sign that Google wants to prioritize faster results across every search, every platform.
Answer users questions
20% of all featured snippets are triggered by a set of just 25 different words. How, What, and Best all top the listings for content that appears in featured snippets. The type of searches that can be answered in a featured snippet are exactly the searches users will use voice for.
If you want your content to reach more voice searches, think of how users ask questions when speaking compared to how they type. People generally speak in more complete sentences so this is a good chance to capture more long-tail search traffic.
Add structured data
Over 40% of all voice results are pulled directly from featured snippets. So, if you're optimizing your structured data to appear in featured snippets, you're also optimizing for voice search.
No SERP wants to send a user to a potentially unsafe page. Over 70% of all voice search results use HTTPS.
Build a strong domain rating
Voice searches are still in the phase of proving to users that they are accurate. If you're going to let a robot dish out confident answers, you need to be confident that they are pulling the information from a good source.
This means a high Domain Rating. The average Ahrefs Domain Rating for voice search results is 76.8.
Simple, easy to read content
Even now Siri and Alexa don't sound entirely natural. To keep answers smooth and easy to understand, average voice search results score between 60 and 70 on Flesch – Kincaid readability tests, which is about a 9th-grade level.
Strong social signals
If a lot of people share a page or a post, you can be pretty confident about the quality of the information (in most cases). Google is using social signals to identify strong information. The average voice search result has 1,199 Facebook shares and has been tweeted 44 times.
Good desktop = Good vocal search results
75% of all vocal search results feature in the top 3 desktop search results for the same query. If Google knows your content is likely to satisfy a regular search, the answer from a voice search is also likely to satisfy the search intent.
Write short answers but long pages
Search engines know that in-depth, researched content is often the most authoritative source of information, yet it needs short answers to read aloud.
The typical voice search result is only 29 words in length but the average word count of a voice search result page is 2,312 words. Good use of H2 titles and structured data markups can help Google pick out isolated answers from longer content.
With these techniques, you can help ensure that more of your content is served from voice searches. Also, all of these tactics are also general SEO best practices, so even if voice search results are not your main goal, they are worthwhile.
Is voice search relevant for digital publishers?
Facts, numbers, binary statements all translate well from text to voice. And, at the moment, voice search is still 3X times more likely to be local in nature. People want to know if their local restaurant is open as they are driving. As publishers tend to produce, longer more nuanced content, it can be hard to see how they will truly take advantage of more voice searches.
But, this doesn't mean that users don't want long-form news or entertainment content through voice search. It just means the product needs to be tailored to the format. Publishers can create products designed to be listened to rather than read.
Google and Amazon have invested heavily in the technology to understand speech and deliver accurate results, based on what users really mean when they talk. It's a long term strategy. Amazon frequently sells its Echo Dot virtual assistants at a loss. They're not selling the device, they're buying an audience.
By committing to an Amazon voice assistant, people are tying themselves to Amazon's network. When these users ask a voice assistant for the news By offering a premium voice search experience, publishers can make it more likely that big platforms choose their content to be used to answer their users' questions.
Content designed to be heard not read
As well as optimizing to appear in search results, publishers' relationships with voice search will also be shaped with direct, exclusive partnerships with the tech giants. We're likely to see publishers creating more voice-only products with the aim to get exclusive relationships with Voice User Interfaces, such as Amazon's Alexa.
Here you can see the 'Flash Briefings' created by newsgroups that are designed to be requested and played via voice searches.
Many media groups have already branched out into podcasts and live events to the transition to making audio-only content may not be a major transition. Micro-podcasts, daily bulletins, headlines, and story precises will all be valuable products when more users start to use voice search for more complicated queries.