*To demonstrate the power of artificial intelligence and voice search technology, the entirety of this blog was dictated, not typed. Minor revisions and peer review were done by computer.
One of the biggest areas of technology growth as we head into the new decade is voice search, which is the use of one’s voice to dictate or ask questions to smart devices such as smartphones and smart speakers. This blog will look into voice search and determine what this technology looks like, as well as how your business or organization can take advantage of and optimize for it in your marketing.
The term voice search, also known as VEO (Voice Engine Optimization) is a broad and, at times, vague category of technology. It can actually be broken into two different areas.
The first category “voice search” can be broken into is smart speakers, including Amazon Alexa and Google Home. These devices can perform tasks such as calling and texting, setting alarms or reminders, jotting down notes for things to remember, or ordering things online. You can also enable new “skills,” which are things like news, games, meditation apps, kids & learning skills (like spelling quizzes), white noise, and many others.
The other area of the voice search category is smartphones, including software like virtual assistants (Siri on iPhones and Google Assistant on Android devices). These assistants have been around for years and have helped turn smartphones into hand-free devices. Common things to tell Siri or Google Assistant to do on your phone include composing text messages, dialing numbers, getting map directions, and setting alarms or reminders.
While all of these technologies use artificial intelligence and have similar natural language detection capabilities, we’ll mostly be focusing more on Google’s Technologies in this blog, because they are the most transparent and advanced in Voice Search, compared to their main competitors Apple and Amazon.
The idea of voice search and its technology can trace its way all the way back to the late 1880s, when Thomas Edison created his dictation machine. This device essentially took Edison’s voice and recorded it onto a record which could be played back using a phonograph at a later time. While it’s not exactly what we consider today is voice search, it was the first concept of using one’s voice and recording it to reproduce it at a later time.
Fast forward to the 1960s through 1980s, IBM began investing significant amounts of research into voice command and search technologies. They were on the cutting edge of many modern inventions at the time and investigated ways in which voice could be used to control or command devices like computers.
In 1997 a computer dictation software called Dragon Naturally Speaking was developed. Some of you may remember trying this out in the late 1990s with WinWord. It wasn’t nearly as advanced as today’s technology, but it was an interesting start which launched several other initiatives.
Artificial intelligence in voice technologies advanced again in 2006, when the NSA started using voice speech recognition to isolate keywords in recorded speech. In 2008, Google launched its voice search app. Three years later, Siri was introduced on Apple devices, followed by Alexa, and finally Google Home in 2016.
While the technology has come far in the last 20 years, the 2020s will show a significant increase in mainstream consumption and use of voice search. In fact, over 81% of Americans now own smartphones and over 25% of American households have a smart speaker such as Google Home or Amazon Alexa.
According to Nielsen, at least two out of five adults use some form of voice search every single day. Think about your habits; that could be something as trivial as asking Siri to get directions, dial the phone, or compose a text. You probably do these things with your voice more often than you realize. Google’s accuracy in detecting what people were saying in 2019 was up to 95% and that only continues to improve.
For smart speakers, the most common voice command is to listen to music. Nearly 90% of people who own smart speakers have reported that they use their speakers to listen to music on something like Spotify or Pandora. Approximately 81% searched for real-time information such as the weather and traffic approximately 81% of them, while roughly 75% of people search for factual information such as history or trivia.
It is predicted that by the end of 2020, over half of all searches will have shifted from the keyboard to the microphone. That includes website searches, questions and Google searches, real-time lookups, etc. Google has also announced that by the end of 2020, the accuracy of their devices will be up to 99%.
Another technology in the future that’s already beginning to roll out in major cities across the U.S. is the concept of the Google duplex. This technology allows you to command your smartphone (Google devices) to make appointments or set reservations at restaurants, taking those menial tasks off your “to do” list. The technology uses AI and mimics a human actually making the request, so the business owner may not even know they aren’t speaking with a person. Listen to one of the recordings of Google’s AI making a reservation at a Thai restaurant:
From the New York Times
A major component to knowing how to answer a question is understanding the intent of the question. One of Google’s main focuses in the last 5 years has been not to simply look for keywords on websites and rank them, or to present the websites with the most keywords to users, but rather to understand the intent behind the searcher’s query and to give the best result for that specific person’s question.
For voice search, this is especially important because it only offers the opportunity to present one singular answer. If somebody asks their Google Home device a question, it is going to return only one answer, as opposed to the ten organic results that a normal search engine results page will display. Therefore it’s increasingly important for search engines to be able to understand the intent of the question being asked in order to ensure the one answer given has the best chance of answering the question.
According to Google, the top six ranking factors when it comes to voice search are:
In order for Google to present an answer as quickly as possible, the correct answer must be clear and concise. It has to be long enough to accurately answer the question, but short enough for a smart device to deliver.
The average word count of an answer from voice search is 41.4 words.
That doesn’t give the opportunity to have a long and extensive answer. So, when you’re looking to rank for voice search queries, take a look at the questions being asked and answer them as clearly and quickly as possible. Here’s an example:
“Ok Google, what is voice search?”
“According to Wikipedia, ‘voice search,’ also called ‘voice enabled,’ allows a user to use a voice command to search the internet, a website or an app.”
Word count of answer: 26 words
Another important fact in being one of the top results for voice search queries, a website must rank well organically as well.
The top three organic results total 78% of all given answers, and 98% of all answers given on the first page, or the first 10 organic spots. Therefore, it is exceedingly important that you still focus on ranking your website organically for queries in traditional ways on Google as well. If you’re unclear how to rank organically on your website, read up into some SEO tips from Webspec.
98% of all answers to voice search come from the first page results on Google…
… organic ranking is extremely important.
What are SERP Features? SERP stands for the Search Engine Results Page, and there are several features that expand the results to have more than just links to websites. These are relatively new features for search engines, and Google and Bing continually roll out new features that display on their SERPs. Google’s mission is to give you the best answer to your question, and they’re always improving user experience. Search features include things like:
Of all voice search results, 68.5% came from a search feature; typically either a featured snippet, an answer, or a related question.
When Google has an answer to a specific question, an answer will be delivered. This happens when there are frequently asked questions using voice, but only if there’s one definitive answer to them. It’s often encyclopedic information that was relatively simple for Google to program into their system, such as:
Featured snippets are different from answers on Google SERPs, primarily because they are sourced and cited from websites. In contrast to answers, your business or organization can be used by Google as a featured snippet. The spot that voice search pulls featured snippets from is above all organic results, in a rank often referred to as “rank zero” because they appear above all ads and above all organic results.
In 2019, 23% of all search result pages included a featured snippet, which is pulled directly from the content on your website. This specifically answers a question being asked and can be formatted in five different ways:
However, the most common types of featured snippets are paragraphs (50%) and lists (37%).
As you can see, however, the answer is not always correct. In the example above, you can see that the question was about how to contact a state representative, but the answer given was from the federal house of representatives.
Over 98% of all featured snippets are featured under an area labeled “People Also Ask.” These are similar questions to the one that the featured snippet answered, but are taken from Google’s big data database to show what else people are asking. It also can serve you well in case the answer given for the featured snippet was incorrect.
When it comes to readability, search engines look for well-structured and well-written content that ultimately matches the intent of the query. The results chosen for voice search answers are simple to understand, with the average answers being understood by a typical 14-year-old. That’s an 8th grade reading level. The reason that an eighth grade level is important is that according to Google, 80% of Americans can easily read and understand content at that reading level. This goes back to Google’s mission to deliver answers quickly and concisely.
The speed of a webpage is another ranking factor to consider when optimizing a website for voice search. According to Google, faster web pages ranked higher organically on search engines than slower web pages do. In fact, when it comes to voice search results, the answers given were often over 10 times faster than the average non-chosen answers on the web.
We recommend that your site takes no longer than 2 seconds to load, otherwise you will see some negative impacts from search engine organic rankings, as well as any chance in ranking for voice search.
The number and quality of backlinks also plays a significant role in the answer selected for voice search queries. This follows Google’s famous and long-standing PageRank algorithm, which dictates that the higher the quantity and quality of websites linking to yours, the higher your search engine rankings could be. Sites ranking lower organically (spots 4 through 10) may still be selected as a voice search result if they have quality backlinks from more relevant sources than the organic results in spaces 1 through 3.
Three other factors, which do not have direct influence over the sites chosen as results for voice search are Schema, HTTPS, and URL depth: schema, HTTPS, and URL depth. Schema is additional bits of code inserted on web pages the tell search engines to extract more specific pieces of information such as reviews, events, recipes, etc. Hypertext Transfer Protocol Secure (HTTPS) is the term given to sites that are secured with a SSL certificate. URL depth is the number of clicks it takes to get to a page from the homepage, meaning that the closer the answer for voice search is to the homepage, the more weight it carries.
All three of these topics are not specifically correlated to the answer chosen in voice search, according to Google, but they all play a significant factor in organic rankings, and therefore indirectly affect the rankings for voice search through that.
We’ll be writing a follow-up blog about how to review your website’s performance for all of the items above. Check back later to read more, or contact our digital marketing team if you’re interested in an audit.
Using all the information above, you’ll notice how important it is to rank well organically. A lot of these things go hand-in-hand, so when we’re talking about good organic rankings, that also has to do with pagespeed, SERP features, and average word count. Ultimately, the best way to focus and begin creating content that’s going to rank well for voice search is to continue to focus on your organic SEO rankings and creating content that’s going to be pulled into a featured snippet.
There are several ways to find out what search term inquiries are used in featured snippets, but often the easiest way (without having to pay for an expensive tool) is to simply perform relevant searches yourself about your industry. Search for your own site on Google and see what some of the common questions are, relating to your business or industry. Then use the Google Search Console query filter tool to find the most common question words: who, what, how, is, and can.
The next step is to make sure that you have content on your website that clearly and correctly answers many of those questions. Put a question in a heading, such as an H2, and answer it directly below with paragraph text. Remember, the average results for voice search queries is 41.4 words, so keep your answers as short as possible. Keeping answers short will also help keep the content at an 8th grade reading level. Additionally, you should be answering your each question specifically and separately for each query.
More ways to help you rank organically and for featured snippets include creating bulleted lists or numbered lists, as well as using long-tail keywords to rank for very specific search queries rather than focusing on very bright and high-volume searches.
As voice search becomes more common, it will continue to become increasingly important to think about ranking for voice. While the results for voice search aren’t always pulled from the number one organic spot, 90% of all answers come from at least the first page of organic results on Google. Therefore, the big takeaway from voice search is that it is important to rank organically and as high as you can for very specific queries and questions. If your business or organization can answer the questions that people are asking clearly and concisely, you have a pretty good shot of ranking for voice search.
If you have any questions, or want to see how you’re doing on your search engine in voice search rankings, we’ve got a fully-staffed digital marketing team who are experts at on and off site SEO. Leave us a comment below to get in touch.