Voice-enabled search is one of the biggest topics in SEO circles, but many content marketers haven’t embraced the implications of virtual assistants and other voice-powered technology.
As an SEO and content architect at Moz, I urge marketers to understand voice’s importance and the need to do so today to prepare for the voice revolution.
Voice search growing
Whether it’s Google Home, Siri by Apple, Cortana by Microsoft, or Amazon’s Alexa, voice represents an ever-growing share of the search pie. Statistics on usage are difficult to come by, but in mid-2016, voice search represented 20% of queries on Google’s mobile app, according to Google CEO Sundar Pichai. And Baidu reports that one of 10 searches on that platform are voice searches.
Mid-2016, voice search represented 20% of queries on @Google’s mobile app, says @sundarpichai. Click To Tweet
Usage rates are likely to grow as more consumers buy virtual assistants, and as the natural language processing and machine learning behind these engines become more accurate and useful. The more people use virtual assistants, the more training data machine-learning models will acquire, and the more powerful and accurate they will become.
Research from MindMeld’s 2016 Intelligent Voice Assistants Research Report breaks down the primary reasons consumers use voice:
- 61% – Useful when hands / vision occupied
- 30% – Faster results
- 24% – Difficulty typing on certain devices
- 22% – They’re fun / cool
- 12% – To avoid confusing menus
- 1% – Other
The survey also reveals the most frequent settings for use of voice-assisted search – home (43%), car (36%), on the go (19%). Work comprises the remaining usage.
Voice SEO basics every marketer should know
First, know there isn’t just one approach to SEO for voice. Google Home uses Google data, Siri pulls Bing data and Alexa pulls in Bing and Amazon data from your personalized shopping habits (unless customized to do otherwise).
No single approach exists for #SEO for voice as each search engine is different, says @BritneyMuller. Click To Tweet
It’s really interesting to see how different the results are for each voice technology. For example, I recently tested a question on three voice-search technologies. I asked Google Home, “What are the best laptops?” Google Home read a list from TechRadar.com. Then I asked Siri the same question, and she responded, “The Apple Macintosh is my favorite computer.” Alexa admitted, “Sorry, I don’t have the answer to that question.” Google Home is the most powerful in terms of the answers it currently provides.
Also, keyword research is more important now than ever. Specifically, in regards to capturing Google Home answers. Google provides most of its voice answers from the featured snippet (i.e., answer box) at the top of search results. Marketers can do strategic keyword research to discover which keywords in their space result in a featured snippet, then they can try to provide a better answer that’s marked up accordingly to grab that featured spot. “What” and “how” questions result in the most featured snippets.
.@Google provides most voice answers from the featured snippet at top of search results, says @BritneyMuller. Click To Tweet
HANDPICKED RELATED CONTENT:
Snippets, SERP, and other voice search-related terms
If you feel uncomfortably “noobish” when the conversation wanders into voice technology and machine learning, here are some keywords to master. (If you are comfortable with the terms, skip to the next section.)
Machine learning
Definitions vary, but the simplest is by Arthur Samuel in 1959: “Machine learning: The field of study that gives computers the ability to learn without being explicitly programmed.” The application of machine learning to marketing is vast – from online recommendation engines to social media sentiment analysis. In voice-search applications, machine learning will allow virtual assistants to understand customer intentions with greater accuracy and so improve search results and transaction fulfillment.
Featured snippet
When you type a question into Google, the first result is often a featured snippet – a short summary of the answer pulled from a third-party site. For SEO experts, becoming the featured snippet is highly desirable for visibility and traffic; layer in voice-search considerations, and winning the featured snippet is critical because Google-enabled voice search only delivers a single answer.
For #SEO experts, winning the featured snippet is highly desirable for visibility & traffic. @BritneyMuller Click To Tweet
SERP: Pronounced “surp,” it’s the first page of your search result or the “search engine results page.”
Entity authority
Google uses online and offline signals to determine how credible or valuable your company is. The higher your entity authority, the more highly your content will rank. According to local SEO expert David Mihm, entity authority boils down to three major signals: a business’s name recognition locally (determined by online mentions), its reputation (do consumers and the media have positive things to say about it), and the level of engagement that business earns online.
Your entity authority comes down to 3 things: online mentions, reputation, & engagement level. @davidmihm Click To Tweet
People Also Ask (PAA) boxes
You’ll see more of Google’s PAA boxes that continuously expand with new questions (based on what PAA you click). Google makes money from advertisers, so keeping us on Google with interactive options makes sense. Research from STAT Search Analytics shows a single PAA can show up in 21 unique SERPs. PAA opportunities can take over some serious SERP real estate.
A single PAA can show up in 21 unique SERPs via @getSTAT #research. #SEO Click To Tweet
Keyword research for voice SEO (Hint: Look at featured snippets)
There are two primary tools to help you figure out which queries result in featured snippets: SEMrush, where you can research keywords that result in a featured snippet. And Moz’s (the company I work for) Keyword Explorer, a tool that allows you to save lists and determine which keywords have a featured snippet.
Use @SEMrush or @Moz Keyword Explorer to figure out which queries result in featured snippets. @BritneyMuller Click To Tweet
Click to enlarge
Essentially you want to target any keywords that (a) already result in a featured snippet in your space and (b) you already rank for on the first page. This is because you are more apt to steal a featured snippet if you already rank on the first page for that query.
Once you’ve isolated which queries matter most, it’s a matter of detective work:
- Is the featured snippet in paragraph, list, bullet, or table format?
- Look at the web page that is currently the featured snippet you are aiming for. How is that page marked up?
- Is there a <ul><li> list format on the website containing the list content?
- Is there a title being pulled into the featured snippet?
- Is it marked up in an <h2> tag?
Does that page provide a better overall experience than yours? If so, why? What page authority and domain authority does the page have?
By evaluating these things, you can provide a stronger answer on your page while tweaking the format similar to the featured snippet being used … and, fingers crossed, get Google to replace it with yours.
Keep in mind, it’s a painstaking process.
I go into granular detail when evaluating a featured-snippet page – studying everything from how fast the page loads, to what photos they are using and how easy the content is to read according to the Flesch-Kincaid readability score. Overall page experience can also be a big factor.
Here’s an interesting anecdote about the type of testing involved in securing featured snippets. At Moz we rank for numerous featured snippets, but snagging them is a lot of work. It occurred to me that I could learn more about what it takes to get one if I tried to lose one. What would it take for me to get one of our featured snippets dropped, and what could it teach me? The process was fascinating.
For example, adding typos to meta descriptions lost me a lot of featured snippets. When you lose one, you can toggle back and forth to see the effect. It’s a great way to quickly determine important featured-snippet ranking factors in your space.
3 more insights on voice search vs. traditional search
Young people are much better at using voice search than we are. Kids use more natural phrasing when searching online, whereas adults are used to typing truncated queries into Google. It’s an interesting issue to think about and an important consideration for marketers.
Young people use more natural phrasing when searching online, says @BritneyMuller. #SEO Click To Tweet
Also, we’re seeing more and more voice-search products designed for use in the kitchen. Amazon noticed a surge in voice searches for recipes and other dietary-related information, and now sells displays that can provide more static answers in the kitchen. It’s possible that these types of fixed displays – which are like big, dumb phones – may finally be a way to monetize voice search. This is an interesting development because advertisers fear the disappearance of screens.
While Google will likely be the leader in the virtual assistant space, Alexa is positioned better for transactions. This may change. Before long Google will be providing unprompted suggestions, such as “Hey, you need to leave for the airport!” This type of interaction could ultimately tie over to the transaction space, such as “Can I order you a car?”
Voice-savvy marketers focus on FAQ pages
The more mature players tend to be all over the FAQ pages. These are jackpots for traffic, SEO, and especially for voice. That’s because most featured snippets and nearly all People Also Ask (PAA) boxes occur when a specific question is asked. Putting together a strong FAQ page can help your website take over more featured-snippet real estate quickly.
FAQ pages are jackpots for traffic, #SEO, & especially for #voice, says @BritneyMuller. Click To Tweet
There are a number of tools to help you figure out the top questions people are asking about your product and/or service. AnswerThePublic.com is a free tool that provides all sorts of question data related to a specific keyword – that’s an easy way to jump-start a FAQ page. I typically start with bottom-of-the-funnel queries. What are the questions customers ask before they make a purchase, (e.g., questions about product, price, and the all-important “versus” questions)? The bottom of the funnel is most advantageous because those people are ready to purchase.
Google Suggest is also a powerful, quick, and easy way to get question data. Google Suggest occurs when you start typing keywords into Google (don’t hit enter or search) and it provides a drop-down of popular/trending questions. Use these suggestions to fuel a FAQ page for your space.
In addition to sleuthing featured snippets and PAAs, marketers should be paying close attention to local search – and specifically, the way Google is evolving its ranking for local search results. Google is moving more toward entity authority, meaning it uses diverse data sources to determine a local business’s credibility.
.@Google is moving toward entity authority. Evolving its ranking to local search, says @BritneyMuller. #SEO Click To Tweet
Entity authority is a complex concept and the way it’s calculated isn’t fully known outside of Google, but it’s becoming clear Google is looking beyond online data – using real-world information about local businesses to determine authority such as credit card transactions, average time spent at a location, how often people come back, and even how clean a business’s bathrooms are. Google is now working with retailers like Mattress World (and many others) to provide in-store beacons that will assist with data collection. This type of data is harder to manipulate for SEO gain, and so experts believe Google will expand its real-world data collection — a trend marketers should pay close attention to.
I’m most fascinated by the brilliant machine learning that goes on behind the scenes of search. Machine-learning applications will provide everyday assistance to us in incredible ways – everything from driving us to work and replying to emails, to more accurately viewing/diagnosing our recent CT scan.
I’ve been opening talks lately with a silly example. You’re getting ready for work in the morning, smell smoke, and run to the kitchen. You see your coffee maker is in flames. Once you put it out, a calming voice greets you:
Good morning. Based on your recent coffee preferences, I’ve found a Keurig 2.0 to best suit your needs for $149. To have this ordered and delivered by tomorrow, please confirm by saying ‘yes.’
It’s interesting to think about whether you find that intrusive or helpful (or downright suspicious!). Local SEO expert David Mihm just wrote a great article theorizing that transaction fulfillment might be Google’s voice-monetization strategy.
To create an accurate (low-loss curve) machine-learning prediction model, however, huge amounts of training data are required. And, because Google has incredible access to the world’s data, it will accelerate its machine-learning capabilities to predict many things – even anticipating your needs.
This article originally appeared in the November issue of CCO magazine. Subscribe for your free print copy today.
Whether you want to address a micro or a macro issue of content marketing, make plans today to attend Content Marketing World 2018. Registration is open. Sign up today.
Please note: All tools included in our blog posts are suggested by authors or sources, not the CMI editorial team. No one post can provide all relevant tools in the space. Feel free to include additional tools in the comments (from your company or ones that you have used).
Cover image by Joseph Kalinowski/Content Marketing Institute